Rand Fishkin uses the ludicrous title, Wizard of Moz. He co-authored/co-founded the Art of SEO, Inbound.org, and Moz (he clearly likes doing stuff with other people). Rand’s an addict of all things content, search, & social on the web, from his multiple blogs to Twitter, Google+, Facebook, LinkedIn, and FourSquare. In his minuscule spare time, Rand enjoys the company of his amazing wife, Geraldine, whose serendipitous travel blog chronicles their journeys.
Presentation link: bit.ly/twoalgo
“Just Google it!” is a simple statement that millions of people would say every day when they seek information. It is also a statement that testifies to the overwhelming power of the Google search algorithm. No matter how big Google’s competitors are, not one can hope to come even close to it when it comes to the Internet search. Microsoft’s Bing is perpetual second, capturing a tiny market share. In fact, when I worked in Silicon Valley in 2003 at IBM Almaden Research center, I tried to get a job with a coolest new research project at the time, the search engine. The current team had the best PhD level computer scientists IBM employed at the time, including a person that won Alan Turing award for the work on relational databases. The best and brightest people tried to design and develop a computing system that is fast enough to process millions of search queries, has huge storage capacity to store information on internet pages and is intelligent enough to understand human queries. Needless to say, my Master in Computer Science from NCSU did not pass the mustard and I did not get a job with that group. But, neither did IBM. We all know today that IBM never launched a search engine. On the other hand, I will never forget my interview experience because it exposed me to unbelievable complexity of a search engine design and algorithms. Today, we all know that Google’s engineers have cracked this difficult problem.
Reverse Engineering of Google Search Engine Algorithm
Needless to say, Google search algorithm itself is highly guarded proprietary secret. I am not even sure that anyone at Google can fully comprehend its architecture today. Us, outsiders, can only watch and observe the search engine output results and try to figure out its inner workings so we can rank our pages higher. When it comes to black box reverse engineering of Google search engine results ranking, Rand Fishkin, is the ultimate guru (his presentation on SEO in two algorithm world).
For the longest part, SEO experts relied on link building practices, keyword stuffing to ensure their client’s web page is shown at the top of the SERP. However, in the last 3 years, Google erased all bad SEO practices regarding manipulative link building. For example, the query “best beef in Seattle” no longer shows only results for these keywords; it shows steakhouses, recipes, anything. Google engineers predicted that we want diverse results and fresh results. Another example is query for “best marketing conferences” — search engine knows that we are not interested in past conferences even though it had a ton of good links. It connects topics and keywords, like marvel super hero. For the first time people can actually believe that link building doesn’t work any more for SERP.
Early on, in 2008 Google rejected machine learning techniques as part of it’s search engine algorithm. That has changed in 2012 as the company started to use machine learning for paid click ads. With machine learning taking over SERP, SEO is changing.
Machine learning, or artificial intelligence algorithms are being dusted off from obscure computer science research papers as they enter commercialization phase and become more powerful. Rand speculates that Google uses variety of inputs, such as: colors, shapes, gradients, perspective, interlacing, text, human label images into its machine learning process to find the best matching algorithm. Using this technique, Google for photos on Android can search for all your photos of cat and show them to you.
2015 Google Algorithm Changes and SEO Impact
As a user, you don’t know why something ranks; you are not even aware of the input variables as the system only gives good/bad search results. Did you click and get satisfied with the information presented? If so, it’s good search result. If you click back and forward, then it’s bad. Click trough rate (CTR), user engagement matter more than ever. Because of that, make sure to optimize the landing page for searcher output, not keywords, links, or anchor text.
Since 2015, its hard to move up the rankings with more traffic visiting your web site and increasing CTR because Google apparently doesn’t use CTR any more. It used to work. Use Google Trends to see the search volume. Bunch of clicks from a robot may drive results for few hours, but that’s it; it’s just a temporary bump.
Panda, Penguin, Hummingbird, Pigeon are Google’s artificial “animals” battling web spam. Google Panda algorithm introduced first in 2012 has a goal to crack down on spam, content farms, scrapers, and websites with a high ad-to-content ratio. Penguin, released in 2012 and updated in 2014 focuses on crushing spam. 2013 release of the Hummingbird uses semantic search and integrate better with the Knowledge Graph. Finally, Pigeon focuses on local search. In 2015 Google made significant strides by adding artificial intelligence to its SERP trying to predict what a person, not a computer is trying to find out.
Best SEO: optimize for people + Google at the same time
Unfortunately, web masters and small business owners still have to use old SEO tactics — keywords, unique content, snippet optimization, but in the future AI could matter more.
2016 SEO optimization goals
- Increase CTR. Optimize title, metadata description, URL, keyword, but your ultimate goal is to have get a lot of clicks. If your CTR goes up, your rank can go up. If you don’t do well, you will fall back down. Every single element in the search engine results matter, so do not ignore any data point.
- Beat your fellow SERP competitors on engagement. The longer your visitors stay on the page and read it, the better you will rank. If people click on your link, but quickly bounce back to the search results, your ranking will suffer.
- Speed, speed and more speed. If your web site does not load immediately, no one will stick around and wait for images and text to slowly render.
- Fill gaps in your visitors’ knowledge. Compel visitors to dig deeper in your web site and learn more about your business by positioning yourself as a subject matter expert. Content marketing matters, doesn’t it?
- Earn more shares and likes on social media. Google looks at engagement, how long does person stays on it, i.e. links shared but not read are not relevant. Also, the content needs to be read on Chrome or Android.
- Fulfill searchers task. Ultimately, Google wants people to quickly and efficiently find the information they are looking for and not go back over and over back to the search engine screen.
Create 10x better content
It’s difficult to create completely original, unique content. However, that should not be the goal, the goal is to create more consumable content with better visuals, better writing and updated recent information on the popular topic. News and current events get AI treatment, not so much medical info. We don’t need better content, but 10X content. Need to have 10x better answer that existing results, that’s the only way to stay out from the crowd. I highly recommend watching Rand’s whiteboard session that will give you a blueprint for creating 10x better, good unique content.
Lastly, if you want to get your hand dirty with AI, check out Alchemy API and Monkey Learn. Enter a keyword of your choice and their AI engine will show you the related AI keywords. Enjoy! I wasn’t really impressed with Money Learn output, but Alchemy API gave me some very insightful information. When I entered my own web page anagard.com for analysis, the tool displayed sentiment as perceived by people for certain phrases. For example, strategic management consultant was perceived favorably, but marketing manager did not. I guess I have some housekeeping to do myself.