Google recently disclosed that 30% of search queries are the result of an algorithm that maps search queries to concepts and then uses those concepts to match with relevant web pages. Artificial Intelligence ranks web pages in many ways. Is the concept of 200 ranking factors still relevant? What should be considered a ranking factor? The answer may help define the direction that SEO is headed.
200 Ranking Factors
The 200 ranking factors were signals that Google said were in use around 2006. Google has changed so it’s time to reconsider what ranking factors are and how that affects SEO.
Nearly 20 years ago, ranking a web page was straightforward. It was a recipe. Here were the ingredients:
- Title Tag
- Add keywords to web pages and meta data
- Link out to authority sites using important keywords
- Use the main keywords on home page and increasingly specific (and longer) keywords the deeper you clicked through the website.
When Googlers said there were over 200 ranking factors it was a concept the SEO community could easily understand. It fit the ranking recipe of links, title tags, headings, etc.
Yet for the past few years we’ve known that Google ranks web pages that don’t even contain the keywords on the page. Today, Google and the SEO industry focuses on satisfying users. It really may be time to reconsider what we mean when we talk about ranking factors. It’s entirely possible that Google has left the room and turned off the lights on the anchor text and keyword SEO party.
Bill Slawski on Ranking Factors
Bill Slawski, of the SEO company Go Fish Digital offered these observations:
When Microsoft came out with RankNet about 10 years ago, they mentioned in that paper that they were using a little more than 500 ranking factors.
I’m not sure if there were ever 200 ranking factors at Google or if there are many that come into play under specific instances, such as when a word is a compound word… like data center or datacenter might.
Treating those as the same unit of meaning would be one possible ranking factor that applies in a relatively small instance. Is it possible that many of Google’s ranking signals are things that are used in relatively small circumstances?
It’s also possible that Google may have certain thresholds in place that certain ranking factors may not take effect unless those thresholds are met. For instance, if a certain category of result is selected 500 times within an hour on a search for something with multiple meanings, such as “lincoln.”
Then that algorithm can say that at the 500 selection mark, within an hour, all of the lincoln results that are about lincoln cars will be the highest ranked lincoln results. Unless there are 750 searches within the next hour for Abraham Lincoln, and then the results on a search for lincoln will be topped by results for the former president.
With a multiple part algorithm like that with two different thresholds, would that count as one ranking factor or two ranking factors?
Google scores pages based upon an authority signal such as PageRank. And they have mentioned in many patents an information retrieval (IR) score based upon relevance.
The details about those are limited, and it is debatable how many ranking signals might be contained within the calculation of an IR score.
One of the new blog posts from Google on ranking images told us that those would also be ranked based upon freshness and upon how high up the images are placed within pages. Another one may have said that Google would also be looking at the authority of the page it is on.”
Re-ranking Algorithms and Ranking Factors
I then asked Bill about re-ranking algorithms. Those are the algorithms that are said kick in after the core ranking algorithm has done it’s thing, in order to rank pages with synonyms or for other reasons.
Bill Slawski on Re-ranking Algorithms
“I have written about different signals that Google may use to re-rank pages after an initial ranking based upon an authority and IR (Information Retrieval) score.
There were likely at least 40 of those that there were patents or papers on by 2007.
Do re-ranking factors count as ranking factors? Something like phrase-based indexing, which with the last continuation patent that came out with new claims, I would no longer call a re-ranking algorithm.
I suspect that phrase-based indexing changed from a re-ranking approach to a ranking approach based upon the new claims in the latest continuation patent. That may be true with other re-ranking approaches as well..
Are Factors in Re-ranking Algorithms Ranking Factors?
I then asked Bill if factors in the re-ranking part of the algorithm that exist outside of the core ranking algorithm should be considered ranking factors?
“Yes, I think it is fair to question what is a ranking factor.
When you look at the large data sets patent, the patent describes millions of possible instances involving searches, documents, and queries. Do those count as ranking factors? If so, that was originally filed in 2003, so there might have been over a million ranking factors for a long time.”
Martin Macdonald on the State of Ranking Factors Today
I next turned Martin Macdonald, a search marketing expert with a distinguished career managing the in-house search marketing of many major brands (learn more at his Bay Area SEO website and on his blog, WebMarketingSchool.com).
Here is what Martin had to say about ranking factors:
“Studying and understanding Ranking Factors are absolutely core for all practitioners of SEO, regardless of whether they’re seasoned professionals, or just starting out in their Search careers. I don’t want to diminish the importance of knowing and understanding what metrics have the potential to move site rankings – but, there’s a lot more nuance than people ever really discuss or publish.
The “200 ranking factors”: the number itself originates from a Press Pack issued by Google way back in 2006 where they claimed to have “in excess of 200 ranking factors.”
That was 12 years ago, and we all know that Google has improved considerably over the past decade. Therefore, assuming that those same 200 factors haven’t changed would likely be dangerous.
Furthermore, we don’t know the definition of ‘Ranking Factor’ that Google is using here. There is a strong argument that in reality, there are just three ranking factors (on-site, off-site & user metrics). Simultaneously you can make a strong argument that there are 40,000+ ranking factors by combining other known metrics.
In order to arrive at the much higher numbers, you need to consider the interplay between factors – for instance: If a page has high levels of content, yet a relatively small amount of time on site, that could be a ranking factor.
If a site has high numbers of backlinks, but the originating source of those links isn’t semantically related to the target, that might be a ranking factor.
If a site has a huge bounce rate, but a strong repeat visitor profile, then that could be ranking factor… and so on.
The overall point here is: if you assume there are more than 200 ranking factors, and you assume that the interplay between each and any of those might also be a ranking factor, you’re looking at “200 factors squared” so, 40,000+.
That’s not to mitigate the importance of studying everything we can about contributory factors to a pages rank – but trying to figure out why a page ranks where it does, using just those items we typically track, is unlikely to provide anything other than the most basic of optimization paths.
The TL;Dr: “It Depends”!”
How a New Perspective on Ranking Factors Impacts SEO
It may be the time to retire the idea of 200 plus ranking factors. Doing so may open your perspective to better ways of ranking a web page.
For example, complacency with content is a common oversight. A publisher may aspire to rank for the two word keyword phrase, but their content is really about a less popular three word phrase. This is usually a disconnect between the most popular user intent and the publisher’s opinion of what is best.
In the medical niche this can manifest in a failure to rank for medical related phrases because the content is about alternative medical solutions. The old way of doing SEO, focused on the Ranking Factors, don’t help.
So adding the two word phrase in the title, in the headings and throughout the page, in addition to the anchor text of the inbound links may no longer help that site rank for the two word phrase. The reason is because the page is about the two word phrase plus the additional word that is less popular.
Moving beyond the ranking factors mindset can be helpful for better understanding why a page doesn’t rank and what needs to be done to make it rank better.
More and more it seems like those old 200 ranking factors aren’t as useful for ranking. An updated approach to content may be called for. As Martin Macdonald advised, “…there’s a lot more nuance than people ever really discuss or publish.”
Images by Shutterstock, Modified by Author
Subscribe to SEJ
Get our weekly newsletter from SEJ’s Founder Loren Baker about the latest news in the industry!