Does FAQ schema still do anything for SEO, or is it just dead markup?

Google cut FAQ rich results back to mostly government and health sites in 2023, so the visible payoff is basically gone. Plenty of agencies still stick FAQPage markup on every service page by default anyway. I think it does very little for rankings now, though I'm not sure it's actually zero. The only place I'd still bother is where the markup matches the visible text exactly and the questions are ones customers really ask, not keyword padding. Has anyone measured a real change in Search Console impressions or clicks after pulling it, or is everyone going on gut feel?

Pulling FAQ markup as a test is mostly measuring noise, since the rich result was already gone for most sites. Rule of thumb: keep it only where the questions already sit on the page as visible copy and match what sales or support actually hears. Don't add it to pages just because a template has a slot for it.

Agreed on the visible-copy rule, but I'd skip the Search Console test. The rich results report will show zero FAQ items for most sites whether the markup is good or bad, and any impression change over a few weeks gets buried under seasonality and whatever else you changed. You can't learn much from that comparison.

I can't prove FAQ markup moves rankings, but I'd still keep it where it's honest. Say a dental clinic has "Is teeth whitening painful?" and "Do you take my insurance?" on the page, pulled from front-desk calls. Marking those up takes an hour and describes the page accurately for any tool that reads structured data. Adding twelve keyword-shaped questions to a crown page costs the same hour, but then the markup no longer matches what a visitor actually sees, and that mismatch is exactly what Google's structured data policies go after.

Skipping the test is a mistake, in my view. Seasonality does wreck a plain before-and-after on a single page, but it mostly washes out if you compare against similar pages that didn't change in the same window. Take twenty service pages, add markup to ten, and track CTR per page in Search Console for six to eight weeks. Since the rich result is mostly gone, don't expect anything visible in the SERP. All you're checking is whether the markup changes click behavior at all, which is a much narrower question than whether it moves rankings, and even a messy test can answer that.

Sasha's control-group idea is sound, but whether it's worth running depends on how many pages you have with real traffic. On a site with twenty service pages each pulling a few clicks a week, six to eight weeks won't separate a real CTR change from noise.