Google recently announced that it would use https as a ranking signal. Reason enough for us to look at our data and analyze whether there are any current effects – and whether it is already evident that https sits could have a ranking advantage over http-URLs.
The news was not entirely unexpected as Google has for some time being doing a lot of high profile work for more security in the web. I do not want to comment on what may lay behind this. I have my own opinion on this issue. In any case, we have not only compared a few data points, but also performed a time series analysis to set the PR-effective statement by Google in context with historical data.
Actually, we had already considered https as feature for this year’s ranking factors study, but we eventually left it out due to the not especially informative results and for reasons of space. And even if Google now says that https is a rather lightweight signal, which currently affects just around 1% of worldwide search queries, we have taken this opportunity to conduct a couple of more fundamental analyses of our data.
If you would like to read a short introduction to the area and some tips on changing over to SSL/HTTPS for website operators, please follow this link: HTTPS as ranking factor – how to handle it.
Yes, no, maybe? On the way to the right data
We have looked at the average rankings of http and https – on a URL basis, directories respectively. Ultimately, only certain domain areas can be encrypted and others not. From these rankings, we have calculated cumulative SEO visibilities and compared the two curves with each other over time.
Initially it looked as if there was indeed a boost for encrypted URLs at a certain point in time – although this was not at the time of Google’s announcement but somewhat earlier – in the last week of June and even once in May 2013. The basis for the following graphics are rankings of hundreds of thousands of keywords:
Some outliers – care required with interpretation
However, it is clear that a handful of domains or individual directories, respectively, were responsible for these outliers (for example play.google.com/ or games.yahoo.com/). If these are discounted, the following picture is obtained:
In this case, the – previously very prominent – outliers have disappeared. The curves follow each other very smoothly.
Focus on differences between HTTPS and HTTP
In this graphic, everything has been reduced to the differences between the two curves:
A difference in the vicinity of the value “2” on the Y axis could be considered as significant, however the actual fluctuations rarely exceed a value of 0.2.
Clear conclusion: HTTPS – to date no relation with rankings
In a nutshell: No relationships have been discernible to date from the data analyzed by us between HTTPS and rankings nor are there any differences between HTTP and HTTPS. In my opinion therefore, Google has not yet rolled out this ranking factor – and/or this factor only affects such a small section of the index to date that it was not possible to identify it with our data.
On the subject of ranking factors! This year’s study is just about ready for release. Those who just can’t wait, can preregister here.
This year’s study is twice as big as last year’s, because we have included several important new factors in the analysis. How SEO develops depends on the interplay of many factors and we want to make this clear.
Have you recently conducted your own analyses or a migration to HTTPS and already have results? Perhaps some of our subscribers have been using HTTPS for some time and have noticed recent changes in rankings? I am always happy to get your feedback and hear about your experience – and of course your opinion of our analyses. Subjectively speaking, I consider https as important and believe that much will change in this area in the future. Objectively speaking, there is no evidence for this trend yet.