Search Engine News

A Search Engine Marketing News Site

A Machine Learning Guide for Average Humans

Posted by alexis-sanders

https://ajax.googleapis.com/ajax/libs/jquery/1.9.1/jquery.min.js
(function($) {
// code using $ as alias to jQuery
$(function() {
// Hide the hypotext content.
$(‘.hypotext-content’).hide();
// When a hypotext link is clicked.
$(‘a.hypotext.closed’).click(function (e) {
// custom handling here
e.preventDefault();
// Create the class reference from the rel value.
var id = ‘.’ + $(this).attr(‘rel’);
// If the content is hidden, show it now.
if ( $(id).css(‘display’) == ‘none’ ) {
$(id).show(‘slow’);
if (jQuery.ui) {
// UI loaded
$(id).effect(“highlight”, {}, 1000);
}
}
// If the content is shown, hide it now.
else {
$(id).hide(‘slow’);
}
});
// If we have a hash value in the url.
if (window.location.hash) {
// If the anchor is within a hypotext block, expand it, by clicking the
// relevant link.
console.log(window.location.hash);
var anchor = $(window.location.hash);
var hypotextLink = $(‘#’ + anchor.parents(‘.hypotext-content’).attr(‘rel’));
console.log(hypotextLink);
hypotextLink.click();
// Wait until the content has expanded before jumping to anchor.
//$.delay(1000);
setTimeout(function(){
scrollToAnchor(window.location.hash);
}, 1000);
}
});
function scrollToAnchor(id) {
var anchor = $(id);
$(‘html,body’).animate({scrollTop: anchor.offset().top},’slow’);
}
})(jQuery);

.hypotext-content {
position: relative;
padding: 10px;
margin: 10px 0;
border-right: 5px solid;
}
a.hypotext {
border-bottom: 1px solid;
}
.hypotext-content .close:before {
content: “close”;
font-size: 0.7em;
margin-right: 5px;
border-bottom: 1px solid;
}
a.hypotext.close {
display: block;
position: absolute;
right: 0;
top: 0;
line-height: 1em;
border: none;
}

Machine learning (ML) has grown consistently in worldwide prevalence. Its implications have stretched from small, seemingly inconsequential victories to groundbreaking discoveries. The SEO community is no exception. An understanding and intuition of machine learning can support our understanding of the challenges and solutions Google’s engineers are facing, while also opening our minds to ML’s broader implications.

The advantages of gaining an general understanding of machine learning include:

  • Gaining empathy for engineers, who are ultimately trying to establish the best results for users
  • Understanding what problems machines are solving for, their current capabilities and scientists’ goals
  • Understanding the competitive ecosystem and how companies are using machine learning to drive results
  • Preparing oneself for for what many industry leaders call a major shift in our society (Andrew Ng refers to AI as a “new electricity”)
  • Understanding basic concepts that often appear within research (it’s helped me with understanding certain concepts that appear within Google Brain’s Research)
  • Growing as an individual and expanding your horizons (you might really enjoy machine learning!)
  • When code works and data is produced, it’s a very fulfilling, empowering feeling (even if it’s a very humble result)

I spent a year taking online courses, reading books, and learning about learning (…as a machine). This post is the fruit borne of that labor — it covers 17 machine learning resources (including online courses, books, guides, conference presentations, etc.) comprising the most affordable and popular machine learning resources on the web (through the lens of a complete beginner). I’ve also added a summary of “If I were to start over again, how I would approach it.”

This article isn’t about credit or degrees. It’s about regular Joes and Joannas with an interest in machine learning, and who want to spend their learning time efficiently. Most of these resources will consume over 50 hours of commitment. Ain’t nobody got time for a painful waste of a work week (especially when this is probably completed during your personal time). The goal here is for you to find the resource that best suits your learning style. I genuinely hope you find this research useful, and I encourage comments on which materials prove most helpful (especially ones not included)! #HumanLearningMachineLearning


Executive summary:

Here’s everything you need to know in a chart:

Machine Learning Resource

Time (hours)

Cost ($)

Year

Credibility

Code

Math

Enjoyability

Jason Maye’s Machine Learning 101 slidedeck: 2 years of headbanging, so you don’t have to

2

$0

’17

Credibility level 3

Code level 1

Math level 1

Enjoyability level 5

{ML} Recipes with Josh Gordon Playlist

2

$0

’16

Credibility level 3

Code level 3

Math level 1

Enjoyability level 4

Machine Learning Crash Course

15

$0

’18

Credibility level 4

Code level 4

Math level 2

Enjoyability level 4

OCDevel Machine Learning Guide Podcast

30

$0

’17-

Credibility level 1

Code level 1

Math level 1

Enjoyability level 5

Kaggle’s Machine Learning Track (part 1)

6

$0

’17

Credibility level 3

Code level 5

Math level 1

Enjoyability level 4

Fast.ai (part 1)

70

$70*

’16

Credibility level 4

Code level 5

Math level 3

Enjoyability level 5

Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems

20

$25

’17

Credibility level 4

Code level 4

Math level 2

Enjoyability level 3

Udacity’s Intro to Machine Learning (Kate/Sebastian)

60

$0

’15

Credibility level 4

Code level 4

Math level 3

Enjoyability level 3

Andrew Ng’s Coursera Machine Learning

55

$0

’11

Credibility level 5

Code level 2

Math level 4

Enjoyability level 1

iPullRank Machine Learning Guide

3

$0

’17

Credibility level 1

Code level 1

Math level 1

Enjoyability level 3

Review Google PhD

2

$0

’17

Credibility level 5

Code level 4

Math level 2

Enjoyability level 2

Caltech Machine Learning on iTunes

27

$0

’12

Credibility level 5

Code level 2

Math level 5

Enjoyability level 2

Pattern Recognition & Machine Learning by Christopher Bishop

150

$75

’06

Credibility level 5

Code level 2

Math level 5

N/A

Machine Learning: Hands-on for Developers and Technical Professionals

15

$50

’15

Credibility level 2

Code level 3

Math level 2

Enjoyability level 3

Introduction to Machine Learning with Python: A Guide for Data Scientists

15

$25

’16

Credibility level 3

Code level 3

Math level 3

Enjoyability level 2

Udacity’s Machine Learning by Georgia Tech

96

$0

’15

Credibility level 5

Code level 1

Math level 5

Enjoyability level 1

Machine Learning Stanford iTunes by Andrew Ng

25

$0

’08

Credibility level 5

Code level 1

Math level 5

N/A

*Free, but there is the cost of running an AWS EC2 instance (~$70 when I finished, but I did tinker a ton and made a Rick and Morty script generator, which I ran many epochs [rounds] of…)


Here’s my suggested program:

1. Starting out (estimated 60 hours)

Start with shorter content targeting beginners. This will allow you to get the gist of what’s going on with minimal time commitment.

2. Ready to commit (estimated 80 hours)

By this point, learners would understand their interest levels. Continue with content focused on applying relevant knowledge as fast as possible.

3. Broadening your horizons (estimated 115 hours)

If you’ve made it through the last section and are still hungry for more knowledge, move on to broadening your horizons. Read content focused on teaching the breadth of machine learning — building an intuition for what the algorithms are trying to accomplish (whether visual or mathematically).

Your next steps

By this point, you will already have AWS running instances, a mathematical foundation, and an overarching view of machine learning. This is your jumping-off point to determine what you want to do.

You should be able to determine your next step based on your interest, whether it’s entering Kaggle competitions; doing Fast.ai part two; diving deep into the mathematics with Pattern Recognition & Machine Learning by Christopher Bishop; giving Andrew Ng’s newer Deeplearning.ai course on Coursera; learning more about specific tech stacks (TensorFlow, Scikit-Learn, Keras, Pandas, Numpy, etc.); or applying machine learning to your own problems.


Why am I recommending these steps and resources?

I am not qualified to write an article on machine learning. I don’t have a PhD. I took one statistics class in college, which marked the first moment I truly understood “fight or flight” reactions. And to top it off, my coding skills are lackluster (at their best, they’re chunks of reverse-engineered code from Stack Overflow). Despite my many shortcomings, this piece had to be written by someone like me, an average person.

Statistically speaking, most of us are average (ah, the bell curve/Gaussian distribution always catches up to us). Since I’m not tied to any elitist sentiments, I can be real with you. Below contains a high-level summary of my reviews on all of the classes I took, along with a plan for how I would approach learning machine learning if I could start over. Click to expand each course for the full version with notes.


In-depth reviews of machine learning courses:

Starting out

Jason Maye’s Machine Learning 101 slidedeck: 2 years of head-banging, so you don’t have to ↓

{ML} Recipes with Josh Gordon ↓

Need to Know: This mini-series YouTube-hosted playlist covers the very fundamentals of machine learning with opportunities to complete exercises.

Loved:

  • It is genuinely beginner-focused.
    • They make no assumption of any prior knowledge.
    • Gloss over potentially complex topics that may serve as noise.
    • Playlist ~2 hours
  • Very high-quality filming, audio, and presentation, almost to the point where it had its own aesthetic.
  • Covers some examples in scikit-learn and TensorFlow, which felt modern and practical.
  • Josh Gordon was an engaging speaker.

Disliked:

  • I could not get Dockers on Windows (suggested package manager). This wasn’t a huge deal, since I already had my AWS setup by this point; however, a bit of a bummer since it made it impossible to follow certain steps exactly.
    • Issue: Every time I tried to download (over the course of two weeks), the .exe file would recursively start and keep spinning until either my memory ran out, computer crashed, or I shut my computer down. I sent this to Docker’s Twitter account to no avail.

Lecturer:

Josh Gordon:

  • Developer Advocate for at TensorFlow at Google
  • Leads Machine Learning advocacy at Google
  • Member of the Udacity AI & Data Industry Advisory Board
  • Masters in Computer Science from Columbia University

Links:

Tips on Watching:

  • The playlist is short (only ~1.5 hours screen time). However, it can be a bit fast-paced at times (especially if you like mimicking the examples), so set aside 3-4 hours to play around with examples and allow time for installation, pausing, and following along.
  • Take time to explore code labs.

Google’s Machine Learning Crash Course with TensorFlow APIs ↓

Need to Know: A Google researcher-made crash course on machine learning that is interactive and offers its own built-in coding system!

Loved:

  • Different formats of learning: high-quality video (with ability to adjust speed, closed captioning), readings, quizzes (with explanations), visuals (including whiteboarding), interactive components/ playgrounds, code lab exercises (run directly in your browser (no setup required!))
  • Non-intimidating
    • One of my favorite quotes: “You don’t need to understand the math to be able to take a look at the graphical interpretation.”
    • Broken down into digestible sections
    • Introduces key terms

Disliked:

  • N/A

Lecturers:

Multiple Google researchers participated in this course, including:

  • Peter Norvig
    • Director of Research at Google Inc.
    • Previously he directed Google’s core search algorithms group.
    • He is co-author of Artificial Intelligence: A Modern Approach
  • D. Sculley
    • Senior Staff Software Engineer at Google
    • KDD award-winning papers
    • Works on massive-scale ML systems for online advertising
    • Was part of a research ML paper on optimizing chocolate chip cookies
    • According to his personal website, he prefers to go by “D.”
  • Cassandra Xia
  • Maya Gupta
    • Leads Glassbox Machine Learning R&D team at Google
    • Associate Professor of Electrical Engineering at the University of Washington (2003-2012)
    • In 2007, Gupta received the PECASE award from President George Bush for her work in classifying uncertain (e.g. random) signals
    • Gupta also runs Artifact Puzzles, the second-largest US maker of wooden jigsaw puzzles
  • Sally Goldman
    • Research Scientist at Google
    • Co-author of A Practical Guide to Data Structures and Algorithms Using Java
    • Numerous journals, classes taught at Washington University, and contributions to the ML community

Links:

Tips on Doing:

  • Actively work through playground and coding exercises

OCDevel’s Machine Learning Guide Podcast ↓

Need to Know: This podcast focuses on the high-level fundamentals of machine learning, including basic intuition, algorithms, math, languages, and frameworks. It also includes references to learn more on each episode’s topic.

Loved:

  • Great for trips (when traveling a ton, it was an easy listen).
  • The podcast makes machine learning fun with interesting and compelling analogies.
  • Tyler is a big fan of Andrew Ng’s Coursera course and reviews concepts in Coursera course very well, such that both pair together nicely.
  • Covers the canonical resources for learning more on a particular topic.

Disliked:

  • Certain courses were more theory-based; all are interesting, yet impractical.
  • Due to limited funding the project is a bit slow to update and has less than 30 episodes.

Podcaster:

Tyler Renelle:

  • Machine learning engineer focused on time series and reinforcement
  • Background in full-stack JavaScript, 10 years web and mobile
  • Creator of HabitRPG, an app that treats habits as an RPG game

Links:

Tips on Listening:

  • Listen along your journey to help solidify understanding of topics.
  • Skip episodes 1, 3, 16, 21, and 26 (unless their topics interest and inspire you!).

Kaggle Machine Learning Track (Lesson 1) ↓

Need to Know: A simple code lab that covers the very basics of machine learning with scikit-learn and Panda through the application of the examples onto another set of data.

Loved:

  • A more active form of learning.
  • An engaging code lab that encourages participants to apply knowledge.
  • This track offers has a built-in Python notebook on Kaggle with all input files included. This removed any and all setup/installation issues.
    • Side note: It’s a bit different than Jupyter notebook (e.g., have to click into a cell to add another cell).
  • Each lesson is short, which made the entire lesson go by very fast.

Disliked:

  • The writing in the first lesson didn’t initially make it clear that one would need to apply the knowledge in the lesson to their workbook.
    • It wasn’t a big deal, but when I started referencing files in the lesson, I had to dive into the files in my workbook to find they didn’t exist, only to realize that the knowledge was supposed to be applied and not transcribed.

Lecturer:

Dan Becker:

  • Data Scientist at Kaggle
  • Undergrad in Computer Science, PhD in Econometrics
  • Supervised data science consultant for six Fortune 100 companies
  • Contributed to the Keras and Tensorflow libraries
  • Finished 2nd (out of 1353 teams) in $3 million Heritage Health Prize data mining competition
  • Speaks at deep learning workshops at events and conferences

Links:

Tips on Doing:

  • Read the exercises and apply to your dataset as you go.
  • Try lesson 2, which covers more complex/abstract topics (note: this second took a bit longer to work through).

Ready to commit

Fast.ai (part 1 of 2) ↓

Need to Know: Hands-down the most engaging and active form of learning ML. The source I would most recommend for anyone (although the training plan does help to build up to this course). This course is about learning through coding. This is the only course that I started to truly see the practical mechanics start to come together. It involves applying the most practical solutions to the most common problems (while also building an intuition for those solutions).

Loved:

  • Course Philosophy:
    • Active learning approach
      • “Go out into the world and understand underlying mechanics (of machine learning by doing).”
    • Counter-culture to the exclusivity of the machine learning field, focusing on inclusion.
      • “Let’s do shit that matters to people as quickly as possible.”
  • Highly pragmatic approach with tools that are currently being used (Jupyter Notebooks, scikit-learn, Keras, AWS, etc.).
  • Show an end-to-end process that you get to complete and play with in a development environment.
  • Math is involved, but is not prohibitive. Excel files helped to consolidate information/interact with information in a different way, and Jeremy spends a lot of time recapping confusing concepts.
  • Amazing set of learning resources that allow for all different styles of learning, including:
    • Video Lessons
    • Notes
    • Jupyter Notebooks
    • Assignments
    • Highly active forums
    • Resources on Stackoverflow
    • Readings/resources
      • Jeremy often references popular academic texts
    • Jeremy’s TEDx talk in Brussels
  • Jeremy really pushes one to do extra and put in the effort by teaching interesting problems and engaging one in solving them.
  • It’s a huge time commitment; however, it’s worth it.
  • All of the course’s profits are donated.

Disliked:

  • Overview covers their approach to learning (obviously I’m a fan!). If you’re already drinking the Kool-aid, skip past.
  • I struggled through the AWS setup (13-minute video) for about five hours (however, it felt so good when it was up and running!).
  • Because of its practicality and concentration on solutions used today to solve popular problem types (image recognition, text generation, etc.), it lacks breadth of machine learning topics.

Lecturers:

Jeremy Howard:

  • Distinguished Research Scientist at the University of San Francisco
  • Faculty member at Singularity University
  • Young Global Leader with the World Economic Forum
  • Founder of Enlitic (the first company to apply deep learning to medicine)
  • Former President and Chief Scientist of the data science platform Kaggle

Rachel Thomas:

Links:

Tips on Doing:

  • Set expectations with yourself that installation is going to probably take a few hours.
  • Prepare to spend about ~70 hours for this course (it’s worth it).
  • Don’t forget to shut off your AWS instance.
  • Balance out machine learning knowledge with a course with more breadth.
  • Consider giving part two of the Fast.ai program a shot!

Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems ↓

Need to Know: This book is an Amazon best seller for a reason. It covers a lot of ground quickly, empowers readers to walk through a machine learning problem by chapter two, and contains practical up-to-date machine learning skills.

Loved:

  • Book contains an amazing introduction to machine learning that briskly provides an overarching quick view of the machine learning ecosystem.
  • Chapter 2 immediately walks the reader through an end-to-end machine learning problem.
  • Immediately afterwards, AurĂ©lien pushes a user to attempt to apply this solution to another problem, which was very empowering.
  • There are review questions at the end of each chapter to ensure on has grasped the content within the chapter and to push the reader to explore more.
  • Once installation was completed, it was easy to follow and all code is available on GitHub.
  • Chapters 11-14 were very tough reading; however, they were a great reference when working through Fast.ai.
  • Contains some powerful analogies.
  • Each chapter’s introductions were very useful and put everything into context. This general-to-specifics learning was very useful.

Disliked:

  • Installation was a common source of issues during the beginning of my journey; the text glided over this. I felt the frustration that most people experience from installation should have been addressed with more resources.

Writer:

Aurélien Géron:

  • Led the YouTube video classification team from 2013 to 2016
  • Currently a machine Learning consultant
  • Founder and CTO of Wifirst and Polyconseil
  • Published technical books (on C++, Wi-Fi, and Internet architectures)

Links:

Tips on Using:

  • Get a friend with Python experience to help with installation.
  • Read the introductions to each chapter thoroughly, read the chapter (pay careful attention to code), review the questions at the end (highlight any in-text answer), make a copy of AurĂ©lien’s GitHub and make sure everything works on your setup, re-type the notebooks, go to Kaggle and try on other datasets.

Broadening your horizons

Udacity: Intro to Machine Learning (Kate/Sebastian) ↓

Need to Know: A course that covers a range of machine learning topics, supports building of intuition via visualization and simple examples, offers coding challenges, and a certificate (upon completion of a final project). The biggest challenge with this course is bridging the gap between the hand-holding lectures and the coding exercises.

Loved:

  • Focus on developing a visual intuition on what each model is trying to accomplish.
  • This visual learning mathematics approach is very useful.
  • Cover a vast variety and breadth of models and machine learning basics.
  • In terms of presenting the concept, there was a lot of hand-holding (which I completely appreciated!).
  • Many people have done this training, so their GitHub accounts can be used as reference for the mini-projects.
  • Katie actively notes documentation and suggests where viewers can learn more/reference material.

Disliked:

  • All of the conceptual hand-holding in the lessons is a stark contrast to the challenges of installation, coding exercises, and mini-projects.
    • This is the first course started and the limited instructions on setting up the environment and many failed attempts caused me to break down crying at least a handful of times.
  • The mini-projects are intimidating.
  • There is extra code added to support the viewers; however, it’s done so with little acknowledgement as to what it’s actually doing. This made learning a bit harder.

Lecturer:

Caitlin (Katie) Malone:

  • Director of Data Science Research and Development at Civis Analytics
  • Stanford PhD in Experimental Particle Physics
  • Intern at Udacity in summer 2014
  • Graduate Researcher at the SLAC National Accelerator Laboratory
  • Podcaster with Ben Jaffe (currently Facebook UI Engineer and a music aficionado) on a machine learning podcast Linear Digressions (100+ episodes)

Sebastian Thrun:

  • CEO of the Kitty Hawk Corporation
  • Chairman and co-founder of Udacity
    • One of my favorite Sebastian quotes: “It occurred to me, I could be at Google and build a self-driving car, or I can teach 10,000 students how to build self-driving cars.”
  • Former Google VP
    • Founded Google X
    • Led development of the robotic vehicle Stanley
  • Professor of Computer Science at Stanford University
  • Formerly a professor at Carnegie Mellon University.

Links:

Tips on Watching:

  • Get a friend to help you set up your environment.
  • Print mini-project instructions to check off each step.

Andrew Ng’s Coursera Machine Learning Course ↓

Need to Know: The Andrew Ng Coursera course is the most referenced online machine learning course. It covers a broad set of fundamental, evergreen topics with a strong focus in building mathematical intuition behind machine learning models. Also, one can submit assignments and earn a grade for free. If you want to earn a certificate, one can subscribe or apply for financial aid.

Loved:

  • This course has a high level of credibility.
  • Introduces all necessary machine learning terminology and jargon.
  • Contains a very classic machine learning education approach with a high level of math focus.
  • Quizzes interspersed in courses and after each lesson support understanding and overall learning.
  • The sessions for the course are flexible, the option to switch into a different section is always available.

Disliked:

  • The mathematic notation was hard to process at times.
  • The content felt a bit dated and non-pragmatic. For example, the main concentration was MATLAB and Octave versus more modern languages and resources.
  • Video quality was less than average and could use a refresh.

Lecturer:

Andrew Ng:

  • Adjunct Professor, Stanford University (focusing on AI, Machine Learning, and Deep Learning)
  • Co-founder of Coursera
  • Former head of Baidu AI Group
  • Founder and previous head of Google Brain (deep learning) project
  • Former Director of the Stanford AI Lab
  • Chairman of the board of Woebot (a machine learning bot that focuses on Cognitive Behavior Therapy)

Links:

Tips on Watching:

  • Be disciplined with setting aside timing (even if it’s only 15 minutes a day) to help power through some of the more boring concepts.
  • Don’t do this course first, because it’s intimidating, requires a large time commitment, and isn’t a very energizing experience.

Additional machine learning opportunities

iPullRank Machine Learning Guide ↓

Need to Know: A machine learning e-book targeted at marketers.

Loved:

  • Targeted at marketers and applied to organic search.
  • Covers a variety of machine learning topics.
  • Some good examples, including real-world blunders.
  • Gives some practical tools for non-data scientists (including: MonkeyLearn and Orange)
    • I found Orange to be a lot of fun. It struggled with larger datasets; however, it has a very visual interface that was more user-friendly and offers potential to show some pretty compelling stories.

C:\Users\asanders\AppData\Local\Microsoft\Windows\INetCache\Content.Word\color=health.png

Disliked:

  • Potential to break up content more with relevant imagery — the content was very dense.

Writers:

iPullRank Team (including Mike King):

  • Mike King has a few slide decks on the basics of machine learnings and AI
  • iPullRank has a few data scientists on staff

Links:

Tips on Reading:

  • Read chapters 1-6 and the rest depending upon personal interest.

Review Google PhD ↓

Need to Know: A two-hour presentation from Google’s 2017 IO conference that walks through getting 99% accuracy on the MNIST dataset (a famous dataset containing a bunch of handwritten numbers, which the machine must learn to identify the numbers).

Loved:

  • This talk struck me as very modern, covering the cutting edge.
  • Found this to be very complementary to Fast.ai, as it covered similar topics (e.g. ReLu, CNNs, RNNs, etc.)
  • Amazing visuals that help to put everything into context.

Disliked:

  • The presentation is only a short conference solution and not a comprehensive view of machine learning.
  • Also, a passive form of learning.

Presenter:

Martin Görner:

  • Developer Relations, Google (since 2011)
  • Started Mobipocket, a startup that later became the software part of the Amazon Kindle and its mobile variants

Links:

Tips on Watching:

  • Google any concepts you’re unfamiliar with.
  • Take your time with this one; 2 hours of screen time doesn’t count all of the Googling and processing time for this one.

Caltech Machine Learning iTunes ↓

Need to Know: If math is your thing, this course does a stellar job of building the mathematic intuition behind many machine learning models. Dr. Abu-Mostafa is a raconteur, includes useful visualizations, relevant real-world examples, and compelling analogies.

Loved:

  • First and foremost, this is a real Caltech course, meaning it’s not a watered-down version and contains fundamental concepts that are vital to understanding the mechanics of machine learning.
  • On iTunes, audio downloads are available, which can be useful for on-the-go learning.
  • Dr. Abu-Mostafa is a skilled speaker, making the 27 hours spent listening much easier!
  • Dr. Abu-Mostafa offers up some strong real-world examples and analogies which makes the content more relatable.
    • As an example, he asks students: “Why do I give you practice exams and not just give you the final exam?” as an illustration of why a testing set is useful. If he were to just give students the final, they would just memorize the answers (i.e., they would overfit to the data) and not genuinely learn the material. The final is a test to show how much students learn.
  • The last 1/2 hour of the class is always a Q&A, where students can ask questions. Their questions were useful to understanding the topic more in-depth.
  • The video and audio quality was strong throughout. There were a few times when I couldn’t understand a question in the Q&A, but overall very strong.
  • This course is designed to build mathematical intuition of what’s going on under the hood of specific machine learning models.
    • Caution: Dr. Abu-Mostafa uses mathematical notation, but it’s different from Andrew Ng’s (e.g., theta = w).
  • The final lecture was the most useful, as it pulled a lot of the conceptual puzzle pieces together. The course on neural networks was a close second!

Disliked:

  • Although it contains mostly evergreen content, being released in 2012, it could use a refresh.
  • Very passive form of learning, as it wasn’t immediately actionable.

Lecturer:

Dr. Yaser S. Abu-Mostafa:

  • Professor of Electrical Engineering and Computer Science at the California Institute of Technology
  • Chairman of Machine Learning Consultants LLC
  • Serves on a number of scientific advisory boards
  • Has served as a technical consultant on machine learning for several companies (including Citibank).
  • Multiple articles in Scientific American

Links:

Tips on Watching:

  • Consider listening to the last lesson first, as it pulls together the course overall conceptually. The map of the course, below, was particularly useful to organizing the information taught in the courses.

Image source: http://work.caltech.edu/slides/slides18.pdf

“Pattern Recognition & Machine Learning” by Christopher Bishop ↓

Need to Know: This is a very popular college-level machine learning textbook. I’ve heard it likened to a bible for machine learning. However, after spending a month trying to tackle the first few chapters, I gave up. It was too much math and pre-requisites to tackle (even with a multitude of Google sessions).

Loved:

  • The text of choice for many major universities, so if you can make it through this text and understand all of the concepts, you’re probably in a very good position.
  • I appreciated the history aside sections, where Bishop talked about influential people and their career accomplishments in statistics and machine learning.
  • Despite being a highly mathematically text, the textbook actually has some pretty visually intuitive imagery.

Disliked:

  • I couldn’t make it through the text, which was a bit frustrating. The statistics and mathematical notation (which is probably very benign for a student in this topic) were too much for me.
  • The sunk cost was pretty high here (~$75).

Writer:

Christopher Bishop:

  • Laboratory Director at Microsoft Research Cambridge
  • Professor of Computer Science at the University of Edinburgh
  • Fellow of Darwin College, Cambridge
  • PhD in Theoretical Physics from the University of Edinburgh

Links:

Tips on Reading:

  • Don’t start your machine learning journey with this book.
  • Get a friend in statistics to walk you through anything complicated (my plan is to get a mentor in statistics).
  • Consider taking a (free) online statistics course (Khan Academy and Udacity both have some great content on statistics, calculus, math, and data analysis).

Machine Learning: Hands-on for Developers and Technical Professionals ↓

Need to Know: A fun, non-intimidating end-to-end launching pad/whistle stop for machine learning in action.

Loved:

  • Talks about practical issues that many other sources didn’t really address (e.g. data-cleansing).
  • Covered the basics of machine learning in a non-intimidating way.
    • Offers abridged, consolidated versions of the content.
    • Added fun anecdotes that makes it easier to read.
    • Overall the writer has a great sense of humor.
    • Writer talks to the reader as if they’re a real human being (i.e., doesn’t expect you to go out and do proofs; acknowledges the challenge of certain concepts).
  • Covers a wide variety of topics.
  • Because it was well-written, I flew through the book (even though it’s about ~300 pages).

Disliked:

  • N/A

Writer:

Jason Bell:

  • Technical architect, lecturer, and startup consultant
  • Data Engineer at MastodonC
  • Former section editor for Java Developer’s Journal
  • Former writer on IBM DeveloperWorks

Links:

Tips on Reading:

  • Download and explore Weka’s interface beforehand.
  • Give some of the exercises a shot.

Introduction to Machine Learning with Python: A Guide for Data Scientists ↓

Need to Know: This was a was a well-written piece on machine learning, making it a quick read.

Loved:

  • Quick, smooth read.
  • Easy-to-follow code examples.
  • The first few chapters served as a stellar introduction to the basics of machine learning.
  • Contain subtle jokes that add a bit of fun.
  • Tip to use the Python package manager Anaconda with Jupyter Notebooks was helpful.

Disliked:

  • Once again, installation was a challenge.
  • The “mglearn” utility library threw me for a loop. I had to reread the first few chapters before I figured out it was support for the book.
  • Although I liked the book, I didn’t love it. Overall it just missed the “empowering” mark.

Writers:

Andreas C. MĂĽller:

  • PhD in Computer Science
  • Lecturer at the Data Science Institute at Columbia University
  • Worked at the NYU Center for Data Science on open source and open science
  • Former Machine Learning Scientist at Amazon
  • Speaks often on Machine Learning and scikit-learn (a popular machine learning library)
  • And he makes some pretty incredibly useful graphics, such as this scikit-learn cheat sheet:

http://1.bp.blogspot.com/-ME24ePzpzIM/UQLWTwurfXI/AAAAAAAAANw/W3EETIroA80/s1600/drop_shadows_background.png

Image source: http://peekaboo-vision.blogspot.com/2013/01/machin…

Sarah Guido:

  • Former senior data scientist at Mashable
  • Lead data scientist at Bitly
  • 2018 SciPy Conference Data Science track co-chair

Links:

Tips on Reading:

  • Type out code examples.
  • Beware of the “mglearn” utility library.

Udacity: Machine Learning by Georgia Tech ↓

Need to Know: A mix between an online learning experience and a university machine learning teaching approach. The lecturers are fun, but the course still fell a bit short in terms of active learning.

Loved:

  • This class is offered as CS7641 at Georgia Tech, where it is a part of the Online Masters Degree. Although taking this course here will not earn credit towards the OMS degree, it’s still a non-watered-down college teaching philosophy approach.
  • Covers a wide variety of topics, many of which reminded me of the Caltech course (including: VC Dimension versus Bayesian, Occam’s razor, etc.)
  • Discusses Markov Decision Chains, which is something that didn’t really come up in many other introductory machine learning course, but they are referenced within Google patents.
  • The lecturers have a great dynamic, are wicked smart, and displayed a great sense of (nerd) humor, which make the topics less intimidating.
  • The course has quizzes, which give the course a slight amount of interaction.

Disliked:

  • Some videos were very long, which made the content a bit harder to digest.
  • The course overall was very time consuming.
  • Despite the quizzes, the course was a very passive form of learning with no assignments and little coding.
  • Many videos started with a bunch of content already written out. Having the content written out was probably a big time-saver, but it was also a bit jarring for a viewer to see so much information all at once, while also trying to listen.
  • It’s vital to pay very close attention to notation, which compounds in complexity quickly.
  • Tablet version didn’t function flawlessly: some was missing content (which I had to mark down and review on a desktop), the app would crash randomly on the tablet, and sometimes the audio wouldn’t start.
  • There were no subtitles available on tablet, which I found not only to be a major accessibility blunder, but also made it harder for me to process (since I’m not an audio learner).

Lecturer:

Michael Littman:

Charles Isbell:

  • Professor and Executive Associate Dean at School of Interactive Computing at Georgia Tech
  • Focus on statistical machine learning and “interactive” artificial intelligence.

Links:

Tips on Watching:

  • Pick specific topics of interest and focusing on those lessons.

Andrew Ng’s Stanford’s Machine Learning iTunes ↓

Need to Know: A non-watered-down Stanford course. It’s outdated (filmed in 2008), video/audio are a bit poor, and most links online now point towards the Coursera course. Although the idea of watching a Stanford course was energizing for the first few courses, it became dreadfully boring. I made it to course six before calling it.

Loved:

  • Designed for students, so you know you’re not missing out on anything.
  • This course provides a deeper study into the mathematical and theoretical foundation behind machine learning to the point that the students could create their own machine learning algorithms. This isn’t necessarily very practical for the everyday machine learning user.
  • Has some powerful real-world examples (although they’re outdated).
  • There is something about the kinesthetic nature of watching someone write information out. The blackboard writing helped me to process certain ideas.

Disliked:

  • Video and audio quality were pain to watch.
    • Many questions asked by students were hard to hear.
  • On-screen visuals range from hard to impossible to see.
  • Found myself counting minutes.
  • Dr. Ng mentions TA classes, supplementary learning, but these are not available online.
  • Sometimes the video showed students, which I felt was invasive.

Lecturer:

Andrew Ng (see above)

Links:

Tips on Watching:

  • Only watch if you’re looking to gain a deeper understanding of the math presented in the Coursera course.
  • Skip the first half of the first lecture, since it’s mostly class logistics.

Additional Resources


Motivations and inspiration

If you’re wondering why I spent a year doing this, then I’m with you. I’m genuinely not sure why I set my sights on this project, much less why I followed through with it. I saw Mike King give a session on Machine Learning. I was caught off guard, since I knew nothing on the topic. It gave me a pesky, insatiable curiosity itch. It started with one course and then spiraled out of control. Eventually it transformed into an idea: a review guide on the most affordable and popular machine learning resources on the web (through the lens of a complete beginner). Hopefully you found it useful, or at least somewhat interesting. Be sure to share your thoughts or questions in the comments!

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2IjrVG3

Advertisements

The GDPR Is Coming. Here’s How It Will Affect Your AdWords Account

The GDPR Is Coming. Here’s How It Will Affect Your AdWords Account

Allen Finn

Last updated :
May 14, 2018

If you’re based—or advertise to prospects—in Europe, there’s a pretty decent chance you’re familiar with the General Data Protection Regulation (GDPR).

It’s a package of new legislative rules being introduced by the European Union to make it easier for residents of EU countries to protect their personal data online. The regulation was officially approved on April 27, 2016, and will formally go into effect across the entirety of the EU by May 25, 2018.

And it’s being heralded as “the most important change in data privacy regulation in 20 years.”

We’ve talked about how this suite of historically restrictive (at least from the advertisers’ perspective) laws will impact your Facebook Advertising efforts.

Today, we tackle AdWords.

The Gist

Search, plain old intent-based search, requires no personally identifying information. Today, at least, a search query doesn’t constitute “personal,” regardless of its contents.

google adwords gdpr impact

Provided you aren’t using any kind of remarketing or conversion tracking, you won’t need to do anything at all. Google is the controller (the handler of personal data) and there is no processor (an entity that process data on behalf of a controller); you’re just along for the ride.

This approach is great for, say, Coca-Cola’s next branding campaign in which impressions are the only metric that matter; for small businesses, not so much.

gdpr impact on large advertisers like coca cola versus smbs

When you want to learn something—or create audiences—based on the tangible business value created by all those clicks you’re paying for, things get messier.

Cookies, Remarketing, and RLSA

Do you use Google Analytics, Tag Manager, or the AdWords Remarketing code on your site to build valuable, bottom-of-the-funnel audiences?

(Gosh, I hope the answer’s yes…)

google tracking suite impacted by the gdpr

If so, you must obtain consent.

Per Google, “Advertisers using AdWords will be required to obtain consent for the use of cookies where legally required, and for the collection, sharing, and use of personal data for personalized ads for users in the EEA. This includes use of remarketing tags and conversion tags. Where legally required, advertisers must also clearly identify each party that may collect, receive, or use end users’ personal data.”

In plain English, this means that if you’re using a Google product to track the on-site action of prospects in order to serve personalized ads down the line, you must acquire their consent to do so.

Exceptions: Customer Match and Store Sales

There are two instances—Customer Match and uploaded Store Sales data—in which Google acts as both a controller and a processor of personal data, meaning that they simultaneously determine the purposes of data while processing data you control.

The exact language they use is as follows (note that you are “the customer”):

“When we handle end user personal data, the customer and Google will each act as independent controllers under the GDPR, except for the Customer Match and Store sales (direct upload) features, where Google will act as the customer’s processor for customer-provided personal data.”

As such, in these situations you are responsible for ensuring that the data Google is processing complies with the GDPR.

Customer Match is a tool that allows you to upload a CSV file loaded with customer data to target specific groups within AdWords.

adwords custom audiences gdpr

Since you’re relying on data that’s by no means pseudonymous to create your Customer Match audience (email, phone, name, and zip code are all pretty identifying), you’ll need to be able to prove that you acquired explicit, opt-in consent from each member of your database; doing so simply isn’t Google’s problem.

Store Sales refers to the ability to the ability to import offline transaction data into AdWords, at which time Google matches transactions data with AdWords user information to create powerful audience for optimization, upselling and cross-selling.

adwords store sales gdpr impact

In addition to the same personally identifying information implicated in Custom Audiences, when it comes to Store Sales there’s also a chance that financial data could be appended and, thus, there is a clear need for informed consent under the GDPR.

Now, the majority of advertisers aren’t using either of these valuable tools, but the ones that are will need to be able to prove to prospective auditors that the information uploaded for Google to process on their behalf is kosher.

What’s Next?

The GDPR promises to be one of the most far-reaching and ambitious consumer protection programs ever devised.

Although the implementation of the GDPR is likely to cause some businesses more difficulty than others (such as enterprise firms that offer “big data” products), it’s important to remember that this legislation is being introduced to protect users’ rights in a time at which almost every conceivable aspect of our lives is stored online – and is highly vulnerable to exposure and exploitation.

The regulations officially go into effect on May 25th. If you haven’t started preparing your AdWords account (and landing pages) for the impending changes, you should probably get going. 

Allen Finn

Allen Finn writes many things at WordStream, where he reigned as fantasy football champion for some time. He likes marveling at funky beer labels, every beat on Liquid Swords, most cuts of beef, and New Hampshire.

from Internet Marketing Blog by WordStream https://ift.tt/2IHle42

An Exciting New Chapter for WordStream!

An Exciting New Chapter for WordStream!

Howard Kogan
Howard Kogan

Last updated :
May 10, 2018

When I joined WordStream last year, the company was already growing at an incredible rate. And yet we have big ambitions in order to fulfill our purpose – to help small and medium sized businesses and the agencies that serve them succeed and grow.

We’ve been working hard since then refining our vision, developing new products and services, and empowering our teams so we can better serve our customers. Today, I’m thrilled to announce that a new chapter for WordStream has begun. We’ve entered into an agreement to be acquired by Gannett, a move that positions us to provide even more value to our customers so they can build their businesses and achieve their most ambitious goals.

Gannett is a $3.1 billion media company that owns 270 media properties (both print and digital) across the U.S. and U.K., as well as a digital advertising division that includes services such as SEO, website development, marketing automation and lead management. You might already  be a reader or advertiser within Gannett’s network and I bet you know one of Gannett’s largest properties – USA TODAY.

This move makes perfect sense for WordStream because of the alignment in our visions: Gannett is a next-generation media company that empowers communities to thrive, and WordStream’s software and services do the same for small businesses. Our offerings will strengthen Gannett’s portfolio and market presence in digital advertising, and together we will be the leader in this space.

How this benefits our customers

Everyone at WordStream knows I live by a two-word phrase: Customers First! Much of my focus in the past year has been making sure that we’re building a customer-first company. Back in November, we had our first Customer Insight Roundtable, where we invited a group of customers from a range of verticals (from health care to home furnishings) into our office for a day of knowledge sharing and workshops. It was an amazing view into what motivates them, and an excellent reminder of why we all come into work every day. It was such a success, we’re doing it again in a few weeks.

Our customers have many challenges and barriers to growth, and our job is to lessen those challenges so they can focus on what they love about their work (whether it’s a doggy day care or a travel agency). As part of Gannett, we’ll be able to offer our customers a greater range of solutions to help them succeed and grow. We’ll be able to leverage Gannett’s expansive media network as well as their strong presence in the UK, Australia, and New Zealand to better serve our growing international client base, and Gannett will invest in our products so we can more quickly expand our cross-platform capabilities, our offerings for agencies and more.

The same but better

I’ve barely scratched the surface of the positive change I believe this acquisition holds in store. At the same time, we will retain the elements that have made WordStream a market leader as we will continue to operate independently inside of Gannett. Our executive team remains in place and we continue to aggressively hire and grow our teams across the company. Our core values have not changed. Our brand, our reputation, and our thought leadership will all continue, and continue to thrive.

Our acquisition by Gannett creates incredible opportunities both for all our people here at WordStream and for our customers. We thank you for being part of the journey so far, and we can’t wait to show you what we’ve got planned!

Howard Kogan

Howard Kogan

Howard M. Kogan joined WordStream as president in June 2017 and became CEO in January 2018. He is responsible for overseeing the company’s day-to-day operations and internal strategy with a focus on rapid growth.

from Internet Marketing Blog by WordStream https://ift.tt/2jQ6C4x

Google Gives Advertisers the Ability to Target Cord-Cutters Through YouTube

Google Gives Advertisers the Ability to Target Cord-Cutters Through YouTube

Allen Finn

Last updated :
May 9, 2018

Live television audiences might not be what they used to, friend, but that doesn’t mean millennial eyeballs aren’t glued to screens during primetime (and every other hour of the day). As such, businesses that once leaned on commercials to build brand awareness and promote new products—whether nationally or regionally—are looking for impressions elsewhere.

Enter Google.

youtube trueview cord cutter targeting

Everyone’s favorite monopoly-haver just released a suite of new features designed to help advertisers reach the coveted cord-cutting subset of humanity (i.e. people who don’t pay for cable) while they catch up on the latest episode of SVU. Why? Because, per Google, “18- to 49-year-olds in the US are either light viewers of TV or do not subscribe to TV; but over 90 percent of these people watch YouTube.”

Here’s what’s being rolled out to help you ply your prospects away from the NBA playoffs.

Target YouTube Audiences on TV Screens

Despite not having a cable box tucked into the entertainment center, people are still using televisions. While recent years have seen mobile devices take the forefront as the medium for video viewership, things are swinging back towards TV.

This is due, in large part, to the rise of gaming consoles-as-entertainment-hubs and the proliferation of services like Netflix, Hulu, Amazon and, of course, YouTube (where Google reports that more than 150 million hours of video are consumed each day…on televisions alone!).  In response to advertisers clamoring to reach these show-binging, sport-crushing viewers, Google has rolled out a new device type to target…

TV Screens.

youtube device targeting tv screens

At some point within the next few months, Google will add television sets to the current list of targetable devices available in both AdWords and DoubleClick. That list currently includes computers, phones, and tablets. In beta tests, ads targeted to TV’s yielded an average lift of 47% in ad recall and a 35% uptick in purchase intent.

Not too shabby.

New Audience: Light TV Viewers

This one’s probably going to matter more to the majority of the advertising community…

If you’re looking to reach cord cutters—either as a standalone audience or paired with other traits that might help you define your target demographic—Google will now offer a new audience segment in AdWords.

It’s called light TV viewers.

Unfortunately, it’s only available as an audience on YouTube. Personally, I’d love to see it rolled out across the Display Network, too; I’d wager that if it sees use on YouTube, it’ll be opened up across the GDN. By pairing the light TV viewer audience with other affinities, geotargeting, and gender/age targeting, you’ll be able to get even closer to homing in on granular subsets of the clicking population, a la Facebook.

And finally, the pièce de résistance.

…. YouTube Gets TV Commercials! (sort of)

One for the big boys.

After launching YouTube TV in 2017. To say the service sees its fair share of use would be the understatement of the century. YouTube TV now reaches 85% of US households and has partnerships with a handful of pro sports leagues; it also just announced a partnership with cable networks to bring even more programming to the platform. As such, it makes complete sense for Google to begin offering advertisers the ability to interrupt the viewing experience of millions of Americans with the commercials they thought they’d left behind.

google preferred youtube trueview premium placements

If you currently leverage television commercials in any capacity—or would like to fold them into your brand-building efforts—you can now use Google Preferred (a pre-packaged set of 12 high-quality channels designed to help advertisers succeed in the coveted 18-34 year-old group) to advertise on both top-performing YouTube content and traditional TV content from one place.

Game. Changer.

More Toys for Advertisers

With Facebook’s recent suite of community-centric tools and AR integrations across Instagram and Messenger, it makes complete sense for Google to begin leveraging their most valuable subsidiary (YouTube) in ways that improve targeting while providing a more engaging, relevant industry for consumers.

Allen Finn

Allen Finn writes many things at WordStream, where he reigned as fantasy football champion for some time. He likes marveling at funky beer labels, every beat on Liquid Swords, most cuts of beef, and New Hampshire.

from Internet Marketing Blog by WordStream https://ift.tt/2IHlakO

Google Locks up Bail Bonds Ads

Google Locks up Bail Bonds Ads

Mark Irvine
Mark Irvine

Last updated :
May 8, 2018

Every day, 3.5 billion searches occur on Google, and each has its own intent. While most searches are trivial or mundane, some ask potentially life-changing questions. The results to those queries have potentially serious consequences.

Google recognizes this and frequently introduces new advertising policies to prevent advertisers from manipulating its users when they’re searching for these sensitive services. In the past, Google has introduced restrictions for ads from locksmiths, medical services, payday loans, and rehab services.

Yesterday, Google announced that it will soon ban all bail bond ads from its platforms. In announcing this new policy, Google stated that “Studies show that for-profit bail bond providers make most of their revenue from communities of color and low-income neighborhoods when they are at their most vulnerable, including through opaque financing offers that can keep people in debt for months or years.”

Google’s newest ban on bail bonds ads goes into effect this July and will restrict advertisers from advertising for the following services across all of Google’s properties, including Google Search, Google Display Network, YouTube, and Gmail.

  • Bail bond agents
  • Bail bondsmen
  • Bail bond financing
  • Bounty hunters

Google’s newest ad policy will not affect advertising for legal services, including criminal defense.

Although these bail bonds services will no longer be able to advertise on Google, there are currently no restrictions preventing them from advertising on Bing or Facebook.

Google’s strong stance on the predatory practice of bail bonds financing was heralded as “…The largest step any corporation has taken on behalf of the millions of women who have loved ones in jails across this country,” by Gina Clayton, executive director of the Essie Justice Group. It’s worth noting that this stance is one that Google will also pay a financial burden for by shunning a good chunk of expensive ad revenue. In 2017, bail bond related keywords were the 2nd most expensive keywords for advertisers on the Google SERP, averaging over $58 per each click!

google adwords most expensive keywords in the usa bail bonds
Mark Irvine

Mark Irvine

Mark is a Senior Data Scientist at WordStream, focused on research and training for the everchanging world of PPC. He was named the 5th Most Influential PPC Expert of 2017 by PPC Hero.

from Internet Marketing Blog by WordStream https://ift.tt/2Kjy5GT

Monitoring Featured Snippets – Whiteboard Friday

Posted by BritneyMuller

We’ve covered finding featured snippet opportunities. We’ve covered the process of targeting featured snippets you want to win. Now it’s time for the third and final piece of the puzzle: how to monitor and measure the effectiveness of all your efforts thus far. In this episode of Whiteboard Friday, Britney shares three pro tips on how to make sure your featured snippet strategy is working.

https://fast.wistia.net/embed/iframe/rcgls1y8tz?seo=false&videoFoam=true

https://fast.wistia.net/assets/external/E-v1.js

Monitoring featured snippets

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today we are going over part three of our three-part series all about featured snippets. So part one was about how to discover those featured snippet opportunities, part two was about how to target those, and this final one is how to properly monitor and measure the effectiveness of your targeting.

So we’ll jump right in. So there are a couple different steps and things you can do to go through this.

I. Manually resubmit URL and check SERP in incognito

First is just to manually resubmit a URL after you have tweaked that page to target that featured snippet. Super easy to do. All you do is go to Google and you type in “add URL to Google.” You will see a box pop up where you can submit that URL. You can also go through Search Console and submit it manually there. But this just sort of helps Google to crawl it a little faster and hopefully get it reprioritized to, potentially, a featured snippet.

From there, you can start to check for the keyword in an incognito window. So, in Chrome, you go to File > New Incognito. It tends to be a little bit more unbiased than your regular browser page when you’re doing a search. So this way, you’d start to get an idea of whether or not you’re moving up in that search result. So this can be anywhere from, I kid you not, a couple of minutes to months.

So Google tends to test different featured snippets over a long period of time, but occasionally I’ve had experience and I know a lot of you watching have had different experiences where you submit that URL to Google and boom — you’re in that featured snippet. So it really just depends, but you can keep an eye on things this way.

II. Track rankings for target keyword and Search Console data!

But you also want to keep in mind that you want to start also tracking for rankings for your target keyword as well as Search Console data. So what does that click-through rate look like? How are the impressions? Is there an upward trend in you trying to target that snippet?

So, in my test set, I have seen an average of around 80% increase in those keywords, just in rankings alone. So that’s a good sign that we’re improving these pages and hopefully helping to get us more featured snippets.

III. Check for other featured snippets

Then this last kind of pro tip here is to check for other instances of featured snippets. This is a really fun thing to do. So if you do just a basic search for “what are title tags,” you’re going to see Moz in the featured snippet. Then if you do “what are title tags” and then you do a -site:Moz.com, you’re going to see another featured snippet that Google is pulling is from a different page, that is not on Moz.com. So really interesting to sort of evaluate the types of content that they are testing and pulling for featured snippets.

Another trick that you can do is to append this ampersand, &num=1, &num=2 and so forth. What this is doing is you put this at the end of your Google URL for a search. So, typically, you do a search for “what are title tags,” and you’re going to see Google.com/search/? that typical markup. You can do a close-up on this, and then you’re just going to append it to pull in only three results, only two results, only four results, or else you can go longer and you can see if Google is pulling different featured snippets from that different quota of results. It’s really, really interesting, and you start to see what they’re testing and all that great stuff. So definitely play around with these two hacks right here.

Then lastly, you really just want to set the frequency of your monitoring to meet your needs. So hopefully, you have all of this information in a spreadsheet somewhere. You might have the keywords that you’re targeting as well as are they successful yet, yes or no. What’s the position? Is that going up or down?

Then you can start to prioritize. If you’re doing hundreds, you’re trying to target hundreds of featured snippets, maybe you check the really, really important ones once a week. Some of the others maybe are monthly checks.

From there, you really just need to keep track of, “Okay, well, what did I do to make that change? What was the improvement to that page to get it in the featured snippet?” That’s where you also want to keep detailed notes on what’s working for you and in your space and what’s not.

So I hope this helps. I look forward to hearing all of your featured snippet targeting stories. I’ve gotten some really awesome emails and look forward to hearing more about your journey down below in the comments. Feel free to ask me any questions and I look forward to seeing you on our next edition of Whiteboard Friday. Thanks.

Video transcription by Speechpad.com

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2jN8pal

How to Discover and Monitor Bad Backlinks

Posted by rjonesx.

Identifying bad backlinks has become easier over the past few years with better tool sets, bigger link indexes, and increased knowledge, but for many in our industry it’s still crudely implemented. While the ideal scenario would be to have a professional poring over your link profile and combing each link one-by-one for concerns, for many webmasters that’s just too expensive (and, frankly, overkill).

I’m going to walk through a simple methodology using Link Explorer and Excel (although you could do this with Google Sheets just as easily) to combine together the power of Moz Link Explorer, Keyword Explorer Lists, and finally Link Lists to do a comprehensive link audit.

The basics

There are several components involved in determining whether a link is “bad” and should potentially be removed. Ultimately, we want to be able to measure the riskiness of the link (how likely is Google to flag the link as manipulative and how much do we depend on the link for value). Let me address three common factors used by SEOs to determine this score:

Trust metrics:

There are a handful of metrics in our industry that are readily available to help point out concerning backlinks. The two that come to mind most often are Moz Spam Score and Majestic Trust Flow (or, better yet, the difference between Citation Flow and Trust Flow). These two scores actually work quite differently. Moz’s Spam Score predicts the likelihood a domain is banned or penalized based on certain site features. Majestic Trust Flow determines the trustworthiness of a domain or page based on the quality of links pointing to it. While calculated quite differently, the goal is to help webmasters identify which sites are trustworthy and which are not. However, while these are a good starting point, they aren’t sufficient on their own to give you a clear picture of whether a link is good or bad.

Anchor text manipulation:

One of the first things an SEO learns is that using valuable anchor text can help increase your rankings. The very next thing they learn is that using valuable anchor text can bring on a penalty. The reason for this is pretty clear: the likelihood a webmaster will give you valuable anchor text out of the goodness of their heart is very rare, so over-optimization sticks out like a sore thumb. So, how do we measure anchor text manipulation? If we look at anchor text with our own eyes, this seems to be rather intuitive, but there’s a better way to do it in an automated, at-scale fashion that will allow us to better judge links.

Low authority:

Finally, low-authority links — especially when you would expect higher authority based on the domain — are concerning. A good link should come from an internally well-linked page on a site. If the difference between the Domain Authority and Page Authority is very high, it can be a concern. It isn’t a strong signal, but it is one worth looking at. This is especially obvious in certain types of spam, like paginated comment spam or forum profile spam.

So, let’s jump into how we can pull together a quick backlink analysis taking into account these various features of a bad backlink profile. If you’d like to follow along with this tutorial, hop into Link Explorer in another tab:

Follow along with Link Explorer

Step 1: Get the backlink data

The first and easiest step is just to get your backlink data from Link Explorer’s huge backlink index. With nearly 30 trillion links in our index, you can rest assured that we will find most of the bad backlinks with which you should be concerned. To begin, visit the Link Explorer > Inbound Links section and enter in the domain or page which you wish to analyze.

How to Find Bad Backlinks

Because we aren’t concerned with nofollow links, you will want to set the “follow” filter so that we only export followed links. We also aren’t concerned with deleted links, so we can set the Link Status to “Active.”

How to Find Bad Backlinks

Once you have set these filters, hit the “Export” button. You will have a couple of choices. If your site has fewer than 1,000 backlinks, go ahead and choose the immediate download. However, if your link profile is larger, choose the largest setting and be patient for the download to be prepared. We can keep going with other steps of the project in the meantime, but you don’t want to miss out on bad links, which means you need to export them all.

A lot of SEOs will stop at this point. With PA, DA, and Spam Score included in the standard export, you can do a damn good job of finding bad links. Link Explorer does all of that out-of-the-box for you. But for our purposes here, we wan’t to go a step further and do “anchor text qualification.” This is especially valuable for large link profiles.

Step 2: Get anchor text

Getting anchor text out of the new Link Explorer is incredibly simple. Just visit Link Explorer > Anchor Text and hit the Export button. No extra filters will be needed here.

How to Find Bad Backlinks

Step 3: Measure anchor text value

Now here is a quick trick where we can take advantage of Moz Keyword Explorer’s Keyword Lists to find anchor text that appears to be manipulated. First, we want to remove some of the extraneous anchor text which we know absolutely won’t be concerning, such as URLs as anchor text. This step isn’t completely necessary, but will save you some some credits in Moz Keyword Explorer, so it might be worth it.

How to Find Bad Backlinks

After you’ve removed the extraneous anchor text, we’ll just copy and paste our anchor text into a new keyword list for Keyword Explorer.

How to Find Bad Backlinks

By putting the anchor text into Keyword Explorer, we’ll be able to sort anchor text by search volume. It isn’t very common that anchor text happens to have a high search volume, but when webmasters are trying to manipulate search results they often use the keyword for which they’d like to rank in the anchor text. Thus, we can use the search volume of anchor text as a proxy for manipulated anchor text. In fact, when working with Remove’em before I joined Moz, we discovered the anchor text manipulation was the most predictive factor in link penalties.

Step 4: Merge, filter, sort, & model

We will now merge the data (backlinks export and keyword list export) to finally get that list of concerning backlinks. Let’s start with the backlink export. We’ll open it up in Excel and then remove duplicate domain-anchor text pairs.

I’ll start by showing you a quick trick to extract out the domains from a long list of URLs. I copied the list of URLs from the first column to the last column in Excel, and then chose Data > Text to Columns > Delimited > Other > /. This will cause the URLs to be split into different columns wherever the slash occurs, leaving you with the 4th new column being just the domain names.

How to Find Bad Backlinks

Once you have completed this step, we are going to remove duplicate domain-anchor text pairs. Notice that we aren’t going to limit ourselves to one link per domain, which is what many SEOs do. This would be a mistake, since there could be multiple concerning links on the site with different anchor text.

How to Find Bad Backlinks

After choosing Data > Remove Duplicates, I select the column of Anchor Text and the column of Domain. With the duplicates removed, we are now left with the links we want to judge as good or bad. We need one more thing, though. We need to merge in the search volume data we got from Keyword Explorer. Hit the export button on the keyword list you created from anchor text in Keyword Explorer:

How to Find Bad Backlinks

Open up the export and then copy and paste the data into a second sheet in Excel, next to the backlinks sheet you already created and filtered. In this case, I named the two sheets “Raw Data” and “Anchor Text Data”:

How to Find Bad Backlinks

You’ll then want to do a VLOOKUP on the backlinks spreadsheet to create a column with the search volume for the anchor text on each link. I’ve taken a screenshot of the VLOOKUP formula I used, but yours will look a little different depending upon the the names of the sheets and the exact columns you’ve created.

Excel formula: =IF(ISNA(VLOOKUP(C2,'Anchor Text Data'!$A$1:$I$402,3,FALSE)),0,VLOOKUP(C2,'Anchor Text Data'!$1:$I$402,3,FALSE))

=IF(ISNA(VLOOKUP(C2,’Anchor Text Data’!$A$1:$I$402,3,FALSE)),0,VLOOKUP(C2,’Anchor Text Data’!$1:$I$402,3,FALSE))

It looks a little complicated, but that’s simply because I’m using two VLOOKUPs simultaneously to replace N/A results with the number 0. You can always manually put in 0 wherever N/A shows up.

Now it’s time for the fun part: modeling. First, I recommend sorting by the volume column you just created just so you can see the most concerning anchor text at the top. It’s amazing to see links with anchor text like “ring” or “jewelry” automatically populate at the top of the list, since they’re also keywords with high search volume.

How to Find Bad Backlinks

Second, we’ll create a new column with a formula that takes into account the quality of the link, the riskiness of the anchor text, and the Spam Score:

Excel formula: =D11+(F11-E11)+(LOG(G11+1)*10)+(LOG(O11+1)*10)

=D11+(F11-E11)+(LOG(G11+1)*10)+(LOG(O11+1)*10)

Let’s break down that formula real quickly:

  • D11: This is simply the Spam Score
  • (F11-E11): This is the Domain Authority minus the Page Authority. (This is a bit debatable — some people might just prefer to choose 100-E11)
  • (Log(G11+1)*10): This is a fancy way of converting the number of times this anchor text link occurs into a consistent number for our equation. Without taking the log(), having a high number here could overcome the other signals.
  • (Log(O11+1)*10): This is a fancy way of converting the search volume to a number consistent for our equation. Without taking the log(), having a high search volume could also overcome other signals.

Once we run this equation and create a new column, we can sort by “Riskiness” and find the links with which we should be most concerned.

How to Find Bad Backlinks

As you can see, examples of comment spam and paid links popped to the top of the list because the formula gives a higher value to low-quality, spammy links with risky anchor text. But wait, there’s more!

Step 5: Build a Link List

Link Explorer doesn’t just leave you hanging after doing analysis. Our goal is to help you do SEO, not just analyze it. Your next step is to start a new Link List.

The Link List feature allows you to track whether certain links are alive. If you embark on a campaign to try and remove some of these spammier links, you can create a Link List and use it to monitor the status of those links. Just create a new list by naming it, adding your domain, and then copying and pasting the concerning links.

How to Find Bad Backlinks

You can now just monitor the Link List as you do your outreach to remove bad links. The Link List will track all the metrics, including whether the link has been removed.

How to Find Bad Backlinks

Wrapping up

Whether you want to do a cursory backlink audit by just looking at Spam Score and PA, or a deep-dive taking into account anchor text qualification, Link Explorer + Keyword Explorer and Link Lists make it possible. With our greatly improved backlink index, you can now rest assured that the data you need is right at your finger tips and, if you need to get down-and-dirty in Excel, you can readily export it to do deeper analysis.

Find your spammy links!

Good luck hunting bad backlinks!

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2rwNJHf

How to Discover and Monitor Bad Backlinks

Posted by rjonesx.

Identifying bad backlinks has become easier over the past few years with better tool sets, bigger link indexes, and increased knowledge, but for many in our industry it’s still crudely implemented. While the ideal scenario would be to have a professional poring over your link profile and combing each link one-by-one for concerns, for many webmasters that’s just too expensive (and, frankly, overkill).

I’m going to walk through a simple methodology using Link Explorer and Excel (although you could do this with Google Sheets just as easily) to combine together the power of Moz Link Explorer, Keyword Explorer Lists, and finally Link Lists to do a comprehensive link audit.

The basics

There are several components involved in determining whether a link is “bad” and should potentially be removed. Ultimately, we want to be able to measure the riskiness of the link (how likely is Google to flag the link as manipulative and how much do we depend on the link for value). Let me address three common factors used by SEOs to determine this score:

Trust metrics:

There are a handful of metrics in our industry that are readily available to help point out concerning backlinks. The two that come to mind most often are Moz Spam Score and Majestic Trust Flow (or, better yet, the difference between Citation Flow and Trust Flow). These two scores actually work quite differently. Moz’s Spam Score predicts the likelihood a domain is banned or penalized based on certain site features. Majestic Trust Flow determines the trustworthiness of a domain or page based on the quality of links pointing to it. While calculated quite differently, the goal is to help webmasters identify which sites are trustworthy and which are not. However, while these are a good starting point, they aren’t sufficient on their own to give you a clear picture of whether a link is good or bad.

Anchor text manipulation:

One of the first things an SEO learns is that using valuable anchor text can help increase your rankings. The very next thing they learn is that using valuable anchor text can bring on a penalty. The reason for this is pretty clear: the likelihood a webmaster will give you valuable anchor text out of the goodness of their heart is very rare, so over-optimization sticks out like a sore thumb. So, how do we measure anchor text manipulation? If we look at anchor text with our own eyes, this seems to be rather intuitive, but there’s a better way to do it in an automated, at-scale fashion that will allow us to better judge links.

Low authority:

Finally, low-authority links — especially when you would expect higher authority based on the domain — are concerning. A good link should come from an internally well-linked page on a site. If the difference between the Domain Authority and Page Authority is very high, it can be a concern. It isn’t a strong signal, but it is one worth looking at. This is especially obvious in certain types of spam, like paginated comment spam or forum profile spam.

So, let’s jump into how we can pull together a quick backlink analysis taking into account these various features of a bad backlink profile. If you’d like to follow along with this tutorial, hop into Link Explorer in another tab:

Follow along with Link Explorer

Step 1: Get the backlink data

The first and easiest step is just to get your backlink data from Link Explorer’s huge backlink index. With nearly 30 trillion links in our index, you can rest assured that we will find most of the bad backlinks with which you should be concerned. To begin, visit the Link Explorer > Inbound Links section and enter in the domain or page which you wish to analyze.

How to Find Bad Backlinks

Because we aren’t concerned with nofollow links, you will want to set the “follow” filter so that we only export followed links. We also aren’t concerned with deleted links, so we can set the Link Status to “Active.”

How to Find Bad Backlinks

Once you have set these filters, hit the “Export” button. You will have a couple of choices. If your site has fewer than 1,000 backlinks, go ahead and choose the immediate download. However, if your link profile is larger, choose the largest setting and be patient for the download to be prepared. We can keep going with other steps of the project in the meantime, but you don’t want to miss out on bad links, which means you need to export them all.

A lot of SEOs will stop at this point. With PA, DA, and Spam Score included in the standard export, you can do a damn good job of finding bad links. Link Explorer does all of that out-of-the-box for you. But for our purposes here, we wan’t to go a step further and do “anchor text qualification.” This is especially valuable for large link profiles.

Step 2: Get anchor text

Getting anchor text out of the new Link Explorer is incredibly simple. Just visit Link Explorer > Anchor Text and hit the Export button. No extra filters will be needed here.

How to Find Bad Backlinks

Step 3: Measure anchor text value

Now here is a quick trick where we can take advantage of Moz Keyword Explorer’s Keyword Lists to find anchor text that appears to be manipulated. First, we want to remove some of the extraneous anchor text which we know absolutely won’t be concerning, such as URLs as anchor text. This step isn’t completely necessary, but will save you some some credits in Moz Keyword Explorer, so it might be worth it.

How to Find Bad Backlinks

After you’ve removed the extraneous anchor text, we’ll just copy and paste our anchor text into a new keyword list for Keyword Explorer.

How to Find Bad Backlinks

By putting the anchor text into Keyword Explorer, we’ll be able to sort anchor text by search volume. It isn’t very common that anchor text happens to have a high search volume, but when webmasters are trying to manipulate search results they often use the keyword for which they’d like to rank in the anchor text. Thus, we can use the search volume of anchor text as a proxy for manipulated anchor text. In fact, when working with Remove’em before I joined Moz, we discovered the anchor text manipulation was the most predictive factor in link penalties.

Step 4: Merge, filter, sort, & model

We will now merge the data (backlinks export and keyword list export) to finally get that list of concerning backlinks. Let’s start with the backlink export. We’ll open it up in Excel and then remove duplicate domain-anchor text pairs.

I’ll start by showing you a quick trick to extract out the domains from a long list of URLs. I copied the list of URLs from the first column to the last column in Excel, and then chose Data > Text to Columns > Delimited > Other > /. This will cause the URLs to be split into different columns wherever the slash occurs, leaving you with the 4th new column being just the domain names.

How to Find Bad Backlinks

Once you have completed this step, we are going to remove duplicate domain-anchor text pairs. Notice that we aren’t going to limit ourselves to one link per domain, which is what many SEOs do. This would be a mistake, since there could be multiple concerning links on the site with different anchor text.

How to Find Bad Backlinks

After choosing Data > Remove Duplicates, I select the column of Anchor Text and the column of Domain. With the duplicates removed, we are now left with the links we want to judge as good or bad. We need one more thing, though. We need to merge in the search volume data we got from Keyword Explorer. Hit the export button on the keyword list you created from anchor text in Keyword Explorer:

How to Find Bad Backlinks

Open up the export and then copy and paste the data into a second sheet in Excel, next to the backlinks sheet you already created and filtered. In this case, I named the two sheets “Raw Data” and “Anchor Text Data”:

How to Find Bad Backlinks

You’ll then want to do a VLOOKUP on the backlinks spreadsheet to create a column with the search volume for the anchor text on each link. I’ve taken a screenshot of the VLOOKUP formula I used, but yours will look a little different depending upon the the names of the sheets and the exact columns you’ve created.

Excel formula: =IF(ISNA(VLOOKUP(C2,'Anchor Text Data'!$A$1:$I$402,3,FALSE)),0,VLOOKUP(C2,'Anchor Text Data'!$1:$I$402,3,FALSE))

=IF(ISNA(VLOOKUP(C2,’Anchor Text Data’!$A$1:$I$402,3,FALSE)),0,VLOOKUP(C2,’Anchor Text Data’!$1:$I$402,3,FALSE))

It looks a little complicated, but that’s simply because I’m using two VLOOKUPs simultaneously to replace N/A results with the number 0. You can always manually put in 0 wherever N/A shows up.

Now it’s time for the fun part: modeling. First, I recommend sorting by the volume column you just created just so you can see the most concerning anchor text at the top. It’s amazing to see links with anchor text like “ring” or “jewelry” automatically populate at the top of the list, since they’re also keywords with high search volume.

How to Find Bad Backlinks

Second, we’ll create a new column with a formula that takes into account the quality of the link, the riskiness of the anchor text, and the Spam Score:

Excel formula: =D11+(F11-E11)+(LOG(G11+1)*10)+(LOG(O11+1)*10)

=D11+(F11-E11)+(LOG(G11+1)*10)+(LOG(O11+1)*10)

Let’s break down that formula real quickly:

  • D11: This is simply the Spam Score
  • (F11-E11): This is the Domain Authority minus the Page Authority. (This is a bit debatable — some people might just prefer to choose 100-E11)
  • (Log(G11+1)*10): This is a fancy way of converting the number of times this anchor text link occurs into a consistent number for our equation. Without taking the log(), having a high number here could overcome the other signals.
  • (Log(O11+1)*10): This is a fancy way of converting the search volume to a number consistent for our equation. Without taking the log(), having a high search volume could also overcome other signals.

Once we run this equation and create a new column, we can sort by “Riskiness” and find the links with which we should be most concerned.

How to Find Bad Backlinks

As you can see, examples of comment spam and paid links popped to the top of the list because the formula gives a higher value to low-quality, spammy links with risky anchor text. But wait, there’s more!

Step 5: Build a Link List

Link Explorer doesn’t just leave you hanging after doing analysis. Our goal is to help you do SEO, not just analyze it. Your next step is to start a new Link List.

The Link List feature allows you to track whether certain links are alive. If you embark on a campaign to try and remove some of these spammier links, you can create a Link List and use it to monitor the status of those links. Just create a new list by naming it, adding your domain, and then copying and pasting the concerning links.

How to Find Bad Backlinks

You can now just monitor the Link List as you do your outreach to remove bad links. The Link List will track all the metrics, including whether the link has been removed.

How to Find Bad Backlinks

Wrapping up

Whether you want to do a cursory backlink audit by just looking at Spam Score and PA, or a deep-dive taking into account anchor text qualification, Link Explorer + Keyword Explorer and Link Lists make it possible. With our greatly improved backlink index, you can now rest assured that the data you need is right at your finger tips and, if you need to get down-and-dirty in Excel, you can readily export it to do deeper analysis.

Find your spammy links!

Good luck hunting bad backlinks!

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2rwNJHf

Moz’s Link Data Used to Suck… But Not Anymore! The New Link Explorer is Here – Whiteboard Friday

Posted by randfish

Earlier this week we launched our brand-new link building tool, and we’re happy to say that Link Explorer addresses and improves upon a lot of the big problems that have plagued our legacy link tool, Open Site Explorer. In today’s Whiteboard Friday, Rand transparently lists out many of the biggest complaints we’ve heard about OSE over the years and explains the vast improvements Link Explorer provides, from DA scores updated daily to historic link data to a huge index of almost five trillion URLs.

https://fast.wistia.net/embed/iframe/tw8djfqtem?seo=false&videoFoam=true

https://fast.wistia.net/assets/external/E-v1.js

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Howdy, Moz fans, and welcome to another edition of Whiteboard Friday. This week I’m very excited to say that Moz’s Open Site Explorer product, which had a lot of challenges with it, is finally being retired, and we have a new product, Link Explorer, that’s taking its place. So let me walk you through why and how Moz’s link data for the last few years has really kind of sucked. There’s no two ways about it.

If you heard me here on Whiteboard Friday, if you watched me at conferences, if you saw me blogging, you’d probably see me saying, “Hey, I personally use Ahrefs, or I use Majestic for my link research.” Moz has a lot of other good tools. The crawler is excellent. Moz Pro is good. But Open Site Explorer was really lagging, and today, that’s not the case. Let me walk you through this.

The big complaints about OSE/Mozscape

1. The index was just too small

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

Mozscape was probably about a fifth to a tenth the size of its competitors. While it got a lot of the quality good links of the web, it just didn’t get enough. As SEOs, we need to know all of the links, the good ones and the bad ones.

2. The data was just too old

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So, in Mozscape, a link that you built on November 1st, you got a link added to a website, you’re very proud of yourself. That’s excellent. You should expect that a link tool should pick that up within maybe a couple weeks, maybe three weeks at the outside. Google is probably picking it up within just a few days, sometimes hours.

Yet, when Mozscape would crawl that, it would often be a month or more later, and by the time Mozscape processed its index, it could be another 40 days after that, meaning that you could see a 60- to 80-day delay, sometimes even longer, between when your link was built and when Mozscape actually found it. That sucks.

3. PA/DA scores took forever to update

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

PA/DA scores, likewise, took forever to update because of this link problem. So the index would say, oh, your DA is over here. You’re at 25, and now maybe you’re at 30. But in reality, you’re probably far ahead of that, because you’ve been building a lot of links that Mozscape just hasn’t picked up yet. So this is this lagging indicator. Sometimes there would be links that it just didn’t even know about. So PA and DA just wouldn’t be as accurate or precise as you’d want them to be.

4. Some scores were really confusing and out of date

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

MozRank and MozTrust relied on essentially the original Google PageRank paper from 1997, which there’s no way that’s what’s being used today. Google certainly uses some view of link equity that’s passed between links that is similar to PageRank, and I think they probably internally call that PageRank, but it looks nothing like what MozRank was called.

Likewise, MozTrust, way out of date, from a paper in I think 2002 or 2003. Much more advancements in search have happened since then.

Spam score was also out of date. It used a system that was correlated with what spam looked like three, four years ago, so much more up to date than these two, but really not nearly as sophisticated as what Google is doing today. So we needed to toss those out and find their replacements as well.

5. There was no way to see links gained and lost over time

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

Mozscape had no way to see gained and lost links over time, and folks thought, “Gosh, these other tools in the SEO space give me this ability to show me links that their index has discovered or links they’ve seen that we’ve lost. I really want that.”

6. DA didn’t correlate as well as it should have

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So over time, DA became a less and less indicative measure of how well you were performing in Google’s rankings. That needed to change as well. The new DA, by the way, much, much better on this front.

7. Bulk metrics checking and link reporting was too hard and manual

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So folks would say, “Hey, I have this giant spreadsheet with all my link data. I want to upload that. I want you guys to crawl it. I want to go fetch all your metrics. I want to get DA scores for these hundreds or thousands of websites that I’ve got. How do I do that?” We didn’t provide a good way for you to do that either unless you were willing to write code and loop in our API.

8. People wanted distribution of their links by DA

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

They wanted distributions of their links by domain authority. Show me where my links come from, yes, but also what sorts of buckets of DA do I have versus my competition? That was also missing.

So, let me show you what the new Link Explorer has.

Moz's new Link Explorer

Click on the whiteboard image above to open a high-resolution version in a new tab!

Wow, look at that magical board change, and it only took a fraction of a second. Amazing.

What Link Explorer has done, as compared to the old Open Site Explorer, is pretty exciting. I’m actually very proud of the team. If you know me, you know I am a picky SOB. I usually don’t even like most of the stuff that we put out here, but oh my god, this is quite an incredible product.

1. Link Explorer has a GIANT index

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So I mentioned index size was a big problem. Link Explorer has got a giant index. Frankly, it’s about 20 times larger than what Open Site Explorer had and, as you can see, very, very competitive with the other services out there. Majestic Fresh says they have about a trillion URLs from their I think it’s the last 60 days. Ahrefs, about 3 trillion. Majestic’s historic, which goes all time, has about 7 trillion, and Moz, just in the last 90 days, which I think is our index — maybe it’s a little shorter than that, 60 days — 4.7 trillion, so almost 5 trillion URLs. Just really, really big. It covers a huge swath of the web, which is great.

2. All data updates every 24 hours

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So, unlike the old index, it is very fresh. Every time it finds a new link, it updates PA scores and DA scores. The whole interface can show you all the links that it found just yesterday every morning.

3. DA and PA are tracked daily for every site

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

You don’t have to track them yourself. You don’t have to put them into your campaigns. Every time you go and visit a domain, you will see this graph showing you domain authority over time, which has been awesome.

For my new company, I’ve been tracking all the links that come in to SparkToro, and I can see my DA rising. It’s really exciting. I put out a good blog post, I get a bunch of links, and my DA goes up the next day. How cool is that?

4. Old scores are gone, and new scores are polished and high quality

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So we got rid of MozRank and MozTrust, which were very old metrics and, frankly, very few people were using them, and most folks who were using them didn’t really know how to use them. PA basically takes care of both of them. It includes the weight of links that come to you and the trustworthiness. So that makes more sense as a metric.

Spam score is now on a 0 to 100% risk model instead of the old 0 to 17 flags and the flags correlate to some percentage. So 0 to 100 risk model. Spam score is basically just a machine learning built model against sites that Google penalized or banned.

So we took a huge amount of domains. We ran their names through Google. If they couldn’t rank for their own name, we said they were penalized. If we did a site: the domain.com and Google had de-indexed them, we said they were banned. Then we built this risk model. So in the 90% that means 90% of sites that had these qualities were penalized or banned. 2% means only 2% did. If you have a 30% spam score, that’s not too bad. If you have a 75% spam score, it’s getting a little sketchy.

5. Discovered and lost links are available for every site, every day

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So again, for this new startup that I’m doing, I’ve been watching as I get new links and I see where they come from, and then sometimes I’ll reach out on Twitter and say thank you to those folks who are linking to my blog posts and stuff. But it’s very, very cool to see links that I gain and links that I lose every single day. This is a feature that Ahrefs and Majestic have had for a long time, and frankly Moz was behind on this. So I’m very glad that we have it now.

6. DA is back as a high-quality leading indicator of ranking ability

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

So, a note that is important: everyone’s DA has changed. Your DA has changed. My DA has changed. Moz’s DA changed. Google’s DA changed. I think it went from a 98 to a 97. My advice is take a look at yourself versus all your competitors that you’re trying to rank against and use that to benchmark yourself. The old DA was an old model on old data on an old, tiny index. The new one is based on this 4.7 trillion size index. It is much bigger. It is much fresher. It is much more accurate. You can see that in the correlations.

7. Building link lists, tracking links that you want to acquire, and bulk metrics checking is now easy

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

Building link lists, tracking links that you want to acquire, and bulk metrics checking, which we never had before and, in fact, not a lot of the other tools have this link tracking ability, is now available through possibly my favorite feature in the tool called Link Tracking Lists. If you’ve used Keyword Explorer and you’ve set up your keywords to watch those over time and to build a keyword research set, very, very similar. If you have links you want to acquire, you add them to this list. If you have links that you want to check on, you add them to this list. It will give you all the metrics, and it will tell you: Does this link to your website that you can associate with a list, or does it not? Or does it link to some page on the domain, but maybe not exactly the page that you want? It will tell that too. Pretty cool.

8. Link distribution by DA

Moz's Link Data Used to Suck... But Not Anymore! The New Link Explorer is Here - Whiteboard Friday

Finally, we do now have link distribution by DA. You can find that right on the Overview page at the bottom.

Look, I’m not saying Link Explorer is the absolute perfect, best product out there, but it’s really, really damn good. I’m incredibly proud of the team. I’m very proud to have this product out there.

If you’d like, I’ll be writing some more about how we went about building this product and a bunch of agency folks that we spent time with to develop this, and I would like to thank all of them of course. A huge thank you to the Moz team.

I hope you’ll do me a favor. Check out Link Explorer. I think, very frankly, this team has earned 30 seconds of your time to go check it out.

Try out Link Explorer!

All right. Thanks, everyone. We’ll see you again for another edition of Whiteboard Friday. Take care.

Video transcription by Speechpad.com

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2HP4f04

Efficient Link Reclamation: How to Speed Up & Scale Your Efforts

Posted by DarrenKingman

Link reclamation: Tools, tools everywhere

Every link builder, over time, starts to narrow down their favorite tactics and techniques. Link reclamation is pretty much my numero-uno. In my experience, it’s one of the best ROI activities we can use for gaining links particularly to the homepage, simply because the hard work — the “mention” (in whatever form that is) — is already there. That mention could be of your brand, an influencer who works there, or a tagline from a piece of content you’ve produced, whether it’s an image asset, video, etc. That’s the hard part. But with it done, and after a little hunting and vetting the right mentions, you’re just left with the outreach.

Aside from the effort-to-return ratio, there are various other benefits to link reclamation:

  1. It’s something you can start right away without assets
  2. It’s a low risk/low investment form of link building
  3. Nearly all brands have unlinked mentions, but big brands tend to have the most and therefore see the biggest routine returns
  4. If you’re doing this for clients, they get to see an instant return on their investment

Link reclamation isn’t a new tactic, but it is becoming more complex and tool providers are out there helping us to optimize our efforts. In this post, I’m going to talk a little about those tools and how to apply them to speed up and scale your link reclamation.

Finding mentions

Firstly, we want to find mentions. No point getting too fancy at this stage, so we just head over to trusty Google and search for the range of mentions we’re working on.

As I described earlier, these mentions can come in a variety of shapes and sizes, so I would generally treat each type of mention that I’m looking for as a separate project. For example, if Moz were the site I was working on, I would look for mentions of the brand and create that as one “project,” then look for mentions of Followerwonk and treat that as another, and so on. The reasons why will become clear later on!

So, we head to the almighty Google and start our searches.

To help speed things up it’s best to expand your search result to gather as many URLs as you can in as few clicks as possible. Using Google’s Search Settings, you can quickly max out your SERPs to one hundred results, or you can install a plugin like GInfinity, which allows you to infinitely scroll through the results and grab as many as you can before your hand cramps up.

Now we want to start copying as many of these results as possible into an Excel sheet, or wherever it is you’ll be working from. Clicking each one and copying/pasting is hell, so another tool to quickly install for Chrome is Linkclump. With this one, you’ll be able to right click, drag, and copy as many URLs as you want.

Linkclump Pro Tip: To ensure you don’t copy the page titles and cache data from a SERP, head over to your Linkclump settings by right-clicking the extension icon and selecting “options.” Then, edit your actions to include “URLs only” and “copied to clipboard.” This will make the next part of the process much easier!

Filtering your URL list

Now we’ve got a bunch of URLs, we want to do a little filtering, so we know a) the DA of these domains as a proxy metric to qualify mentions, and b) whether or not they already link to us.

How you do this bit will depend on which platforms you have access to. I would recommend using BuzzStream as it combines a few of the future processes in one place, but URL Profiler can also be used before transferring your list over to some alternative tools.

Using BuzzStream

If you’re going down this road, BuzzStream can pretty much handle the filtering for you once you’ve uploaded your list of URLs. The system will crawl through the URLs and use their API to display Domain Authority, as well as tell you if the page already links to you or not.

The first thing you’ll want to do is create a “project” for each type of mention you’re sourcing. As I mentioned earlier this could be “brand mentions,” “creative content,” “founder mentions,” etc.

When adding your “New Project,” be sure to include the domain URL for the site you’re building links to, as shown below. BuzzStream will then go through and crawl your list of URLs and flag any that are already linking to you, so you can filter them out.

Next, we need to get your list of URLs imported. In the Websites view, use Add Websites and select “Add from List of URLs”:

The next steps are really easy: Upload your list of URLs, then ensure you select “Websites and Links” because we want BuzzStream to retrieve the link data for us.

Once you’ve added them, BuzzStream will work through the list and start displaying all the relevant data for you to filter through in the Link Monitoring tab. You can then sort by: link status (after hitting “Check Backlinks” and having added your URL), DA, and relationship stage to see if you/a colleague have ever been in touch with the writer (especially useful if you/your team uses BuzzStream for outreach like we do at Builtvisible).

Using URL Profiler

If you’re using URL Profiler, firstly, make sure you’ve set up URL Profiler to work with your Moz API. You don’t need a paid Moz account to do this, but having one will give you more than 500 checks per day on the URLs you and the team are pushing through.

Then, take the list of URLs you’ve copied using Linkclump from the SERPs (I’ve just copied the top 10 from the news vertical for “moz.com” as my search), then paste the URLs in the list. You’ll need to select “Moz” in the Domain Level Data section (see screenshot) and also fill out the “Domain to Check” with your preferred URL string (I’ve put “Moz.com” to capture any links to secure, non-secure, alternative subdomains and deeper level URLs).

Once you’ve set URL Profiler running, you’ll get a pretty intimidating spreadsheet, which can simply be cut right down to the columns: URL, Target URL and Domain Mozscape Domain Authority. Filter out any rows that have returned a value in the Target URL column (essentially filtering out any that found an HREF link to your domain), and any remaining rows with a DA lower than your benchmark for links (if you work with one).

And there’s my list of URLs that we now know:

1) don’t have any links to our target domain,

2) have a reference to the domain we’re working on, and

3) boast a DA above 40.

Qualify your list

Now that you’ve got a list of URLs that fit your criteria, we need to do a little manual qualification. But, we’re going to use some trusty tools to make it easy for us!

The key insight we’re looking for during our qualification is if the mention is in a natural linking element of the page. It’s important to avoid contacting sites where the mention is only in the title, as they’ll never place the link. We particularly want placements in the body copy as these are natural link locations and so increase the likelihood of your efforts leading somewhere.

So from my list of URLs, I’ll copy the list and head over to URLopener.com (now bought by 10bestseo.com presumably because it’s such an awesome tool) and paste in my list before asking it to open all the URLs for me:

Now, one by one, I can quickly scan the URLs and look for mentions in the right places (i.e. is the mention in the copy, is it in the headline, or is it used anywhere else where a link might not look natural?).

When we see something like this (below), we’re making sure to add this URL to our final outreach list:

However, when we see this (again, below), we’re probably stripping the URL out of our list as there’s very little chance the author/webmaster will add a link in such a prominent and unusual part of the page:

The idea is to finish up with a list of unlinked mentions in spots where a link would fit naturally for the publisher. We don’t want to get in touch with everyone, with mentions all over the place, as it can harm your future relationships. Link building needs to make sense, and not just for Google. If you’re working in a niche that mentions your client, you likely want not only to get a link but also build a relationship with this writer — it could lead to 5 links further down the line.

Getting email addresses

Now that you’ve got a list of URLs that all feature your brand/client, and you’ve qualified this list to ensure they are all unlinked and have mentions in places that make sense for a link, we need to do the most time-consuming part: finding email addresses.

To continue expanding our spreadsheet, we’re going to need to know the contact details of the writer or webmaster to request our link from. To continue our theme of efficiency, we just want to get the two most important details: email address and first name.

Getting the first name is usually pretty straightforward and there’s not really a need to automate this. However, finding email addresses could be an entirely separate article in itself, so I’ll be brief and get to the point. Read this, and here’s a summary of places to look and the tools I use:

  • Author page
  • Author’s personal website
  • Author’s Twitter profile
  • Rapportive & Email Permutator
  • Allmytweets
  • Journalisted.com
  • Mail Tester

More recently, we’ve been also using Skrapp.io. It’s a LinkedIn extension (like Hunter.io) that installs a “Find Email” button on LinkedIn with a percentage of accuracy. This can often be used with Mail Tester to discover if the suggested email address provided is working or not.

It’s likely to be a combination of these tools that helps you navigate finding a contact’s email address. Once we have it, we need to get in touch — at scale!

Pro Tip: When using Allmytweets, if you’re finding that searches for “email” or “contact” aren’t working, try “dot.” Usually journalists don’t put their full email address on public profiles in a scrapeable format, so they use “me@gmail [dot] com” to get around it.

Making contact

So, because this is all about making the process efficient, I’m not going to repeat or try to build on the other already useful articles that provide templates for outreach (there is one below, but that’s just as an example!). However, I am going to show you how to scale your outreach and follow-ups.

Mail merges

If you and your team aren’t set in your ways with a particular paid tool, your best bet for optimizing scale is going to be a mail merge. There are a number of them out there, and honestly, they are all fairly similar with either varying levels of free emails per day before you have to pay, or they charge from the get-go. However, for the costs we’re talking about and the time it saves, building a business case to either convince yourself (freelancers) or your finance department (everyone else!) will be a walk in the park.

I’ve been a fan of Contact Monkey for some time, mainly for tracking open rates, but their mail merge product is also part of the $10-a-month package. It’s a great deal. However, if you’re after something a bit more specific, YAMM is free to a point (for personal Gmail accounts) and can send up to 50 emails a day.

You’ll likely need to work through the process with the whatever tool you pick but, using your spreadsheet, you’ll be able to specify which fields you want the mail merge to select from, and it’ll insert each element into the email.

For link reclamation, this is really as personable as you need to get — no lengthy paragraphs on how much you loved [insert article related to my infographic] or how long you’ve been following them on Twitter, just a good old to the point email:

Hi [first name],

I recently found a mention of a company I work with in one of your articles.

Here’s the article:[insert URL]

Where you’ve mentioned our company, Moz, would you be able to provide a link back to the domain Moz.com, in case users would like to know more about us?

Many thanks,
Darren.

If using BuzzStream

Although BuzzStream’s mail merge options are pretty similar to the process above, the best “above and beyond” feature that BuzzStream has is that you can schedule in follow up emails as well. So, if you didn’t hear back the first time, after a week or so their software will automatically do a little follow-up, which in my experience, often leads to the best results.

When you’re ready to start sending emails, select the project you’ve set up. In the “Websites” section, select “Outreach.” Here, you can set up a sequence, which will send your initial email as well as customized follow-ups.

Using the same extremely brief template as above, I’ve inserted my dynamic fields to pull in from my data set and set up two follow up emails due to send if I don’t hear back within the next 4 days (BuzzStream hooks up with my email through Outlook and can monitor if I receive an email from this person or not).

Each project can now use templates set up for the type of mention you’re following up. By using pre-set templates, you can create one for brand mention, influencers, or creative projects to further save you time. Good times.

I really hope this has been useful for beginners and seasoned link reclamation pros alike. If you have any other tools you use that people may find useful or have any questions, please do let us know below.

Thanks everyone!

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!

from Moz Blog https://ift.tt/2rg8xm0