What happened when we turned off AMP
We went into this experiment knowing there was some risk, but haven't seen anything to make us reconsider the move.
A little less than two months ago Search Engine Land made the decision to stop publishing versions of our content using Google’s Accelerated Mobile Pages. For us, it boiled down to wanting to simplify our reporting and our desire to end the process of hosting Search Engine Land content on 3rd party servers.
Since then, a lot has happened, but the bottom line is we have seen very little disruption to our traffic and have reaped the benefit of having a clearer picture of our audience analytics.
What happened to traffic? For us, it is difficult to draw any major conclusions about traffic changes since we turned off AMP. Search Engine Land is a media website that primarily produces journalism, so we are very much tied to the news that emerges. As you would expect, when big news like core updates or major Google Ads changes happens our traffic jumps. But as the news dies down during the holiday season we usually see month-to-month declines. That why year-over-year benchmarking is generally favored by news organizations.
We did not see any year-over-year declines in traffic that we could tie to AMP aside from the loss of pageviews to a handful of pieces that routinely spike for organic traffic. For example, an older article about Google SERP Easter Eggs ranks highly for us and usually spikes a few times during the year (including Easter time!). Mobile traffic to that post was previously going to the AMP version. However, we turned off AMP at a time that piece was spiking on mobile and did not see that traffic shift back to our native page. The page itself has never really driven quality traffic so the lost traffic isn’t really a problem.
Safeguarding. Around the time we shut off AMP we also took a few steps that could safeguard us in case the experiment caused a big traffic decline. We increased our publishing volume for starters. We also adjusted the strategy in our newsletters to better optimize for click-through rate. That move was also in response to Apple’s privacy change in iOS 15 that now makes open rates a less reliable metric.
The big win. One of the main reasons for turning off AMP was to better understand our metrics. Despite several failed attempts at AMP stitching in Google Analytics, we never could tell how our audience moves from our AMP pages to our native ones. Users were undoubtedly being double-counted as unique in both the AMP and our native website dashboards. The clearest indicator that this was true is in the change we’ve seen in return visitors since we turned off AMP. The number of sessions by return visitors has jumped by 30% since we made the change, and now we have a far better picture of our most valuable audience set.
Why we care. We went into this experiment knowing there was some risk, but haven’t seen anything to make us reconsider the move. The biggest question mark had always been around the Page Experience Update. AMP pages were as fast as they come, so the worry was that our native pages that don’t benchmark as fast as AMP would lose out. It is true that we saw the percentage of pages with “good” Core Web Vitals scores in Goole Search Console plummet when we turned off AMP, but we do not believe it hurt traffic or rankings. It makes sense because many SEOs are still struggling to tie their own wins or losses directly to the Page Experience Update.
So we’re not looking back. And if you have your story about turning off AMP we’d love to hear it.