Study: Google penalizes text hidden with CSS and JavaScript

A recent study challenges Matt Cutts’s statements about text hidden on a web page in order to improve the user experience (via JavaScript, CSS, or textarea). Back in 2013, Google’s former head of the Webspam team took a reassuring stance, stating that Google did not penalize hidden content as long as it wasn’t done in a “spammy” way.

However, over the years many SEOs have noticed that hidden text carries less weight in Google’s eyes than text shown to the user by default. These beliefs are confirmed by a study that has just been published and that contradicts Matt Cutts. The sole purpose of this study was to prove that hidden text hurts the ranking of that content in Google.

SEO vs UX

The English SEO agency Reboot Online Marketing observed that there are still differences of opinion between SEOs and web developers about whether to display all of the text (for search visibility) or only part of it (for the page’s UX), adding a link to expand the rest of the text. Web developers generally justify their approach by arguing that hiding text has little or no impact on a page’s ranking on Google or on its content.

Fetch as Google

When Google made the famous “fetch as Google” tool available in its Search Console, we got a glimpse of the direction Google was heading, taking into account not only the code but also how that code was rendered — in other words, how the page displayed on users’ screens. This meant Google was able to tell what was hidden from users on display, and part of the SEO community didn’t see why Google would give as much weight to hidden text as to text that was clearly visible to everyone. However, since there was no study confirming or refuting this theory, it was high time someone set out to demonstrate it.

Conclusions of the SEO study on hidden text

Let’s get to the conclusions: This study clearly shows that after 6 months of indexing and 3,300 indexing readings, among the 20 domain names used for the test on a new query (a phrase that was not indexed at the start of the test: “andefabribiles”), the domain names displaying the full text (on average 400 words) rank better than the domain names that show only the first paragraph of the text and hide the rest via CSS, JavaScript (with a link to expand the text), or Textarea (with scroll).

After 6 months of testing, experimentation, and data analysis, the numbers are clear and the conclusion is easy to draw… Google prefers visible text!

Results of the hidden text study on Google © Reboot Online Marketing

The only caveat we would add to this study is that it covers desktop data only, since a technical issue meant that mobile data was not recorded.

Indeed, we are more likely to hide text by default from mobile users to spare them from having to scroll through long pages of text. The person in charge of the study nevertheless notes that the data observed (before the technical issue) was substantially similar to that of desktop. It will be interesting to run this study again once Google has switched to “Mobile first” indexing.