Some correlation studies have found interesting and contradictory things about topicality, which led me to do some experiments with surprising results. It appears that there is a cap to topicality: at some point your page is considered optimal for a given keyword, and further topicality improvements won’t help.
And that topicality cap isn’t as high as most SEOs think.
This point of optimization saturation has a big impact on how much effort we need to put into our topicality optimization, and means that there is no reward for perfect. It turns out just getting pretty close is actually just as good as perfect.
Correlation Studies & SEO Testing
I read a correlation study in 2012 that found no correlation between the title tag and rankings. This surprised me since it’s generally considered one of the more important topicality signals. More importantly, I have personally seen substantial increases in ranking just by improving title tags. Not just once, but many, many times. All SEOs have.
So I set out to prove this study wrong.
I approached it in two ways: I took pages that were ranking highly for a keyword, and completely removed the keyword from the title tag; then I built new pages with no keyword in the title tag, waited for them to rank, then optimized the title. I did this for a over dozen sites in all industries, with all traffic levels. I did this both on in incredibly competitive keywords and fairly noncompetitive ones (one of the advantages of working in an agency is the ability to test on a lot of different sites — with their permission, of course).
What I learned surprised me. Indeed, for page after page the title appeared to have no impact on rankings! This was shocking. It ran in the face of common belief and personal experience. But then I started encountering some pages where, as I expected, the title was hugely important. I removed the keyword from the title and when the page was reindexed the rankings fell precipitously.
I eventually realized what was happening: the pages that were well-optimized (which was most of them I was working on, of course) had no ranking change when the title was changed. But pages that were not well optimized — ones with only a short paragraph using the keyword once or twice — were hugely influenced by the title tag.
With this hypothesis in mind, I then expanded the test and was able to verify that this was in fact the case. Not just with the title tag, but also with the URL. As long as the content of the page itself was of a decent length, on topic, and had good keyword (and synonym/variant) usage, the other topicality elements didn’t matter at all.
And while this was a departure from common SEO thought, it also makes perfect sense.
Topicality is essentially just telling Google whether or not the topic of the pages matches the search query. Some pages might just mention it once, others might be pretty relevant while the best Google is confident are entirely about the query.
But at some point, when your page is entirely about the keyword, it can’t be more about it. And there are enough pages out there with poor URL structure or flavorful titles that aren’t keyword rich that these can’t be required signals if Google is to do its job well.
Implications of Optimization Saturation
This is why I take a somewhat more laid back approach to on-page optimization than you’ll see in a lot of other SEO guides and books: they stress the importance of optimizing every factor to the smallest degree. But we now know that while every factor can matter, if we’re doing our jobs well not every factor will matter.
If you have genuinely good content with good keyword usage, that is very likely enough right there. Including the keyword in the title and URL is still absolutely something you should do, but it’s more of a safety net than something that’s going to make a big difference. Definitely don’t waste your time going back and changing the URL of every page to squeeze the last ounce of SEO juice out — because you probably have all the topicality juice you can get.
If you’re very confident in the quality of your content optimization and length, you can even use the title tag to target a related keyword, rather than your primary keyword (as long as it still makes a good title for CTR purposes, of course).
Confirmation on Wayfair.com
Years after doing this research, I encountered some big confirmation of this while leading the SEO team at Wayfair.com.
An accidental code change altered the title tags of every page on every site. For one site (Wayfair) this made no difference to rankings and traffic. For another site (AllModern) rankings feel hugely.
In the case of AllModern, we used the title tag to optimize the word “Modern” for our topicality: because the pages themselves often never even used the word modern (generally for good reasons), or used it seldom. So with poor “modern” optimization, the title tag was vital. But on Wayfair a page selling Beds was well optimized for “beds” by default, so the title tag was irrelevant.
I wrote up this “accidental SEO test” along with a summary of other findings over at Moz.com.
Oh, and here’s one final fascinating tidbit on the topic: the only on-page factor that does not appear to have a cap is word count. More words is always better.