Aug 15 2026
Your Facebook page proudly shows 227 reviews. Your website widget says "Based on 172 reviews". Or worse - "Based on 1 review", sitting right above a list of twenty-five perfectly real ones. If you've ever connected Facebook recommendations to a website, there's a good chance you've seen some version of this. It is not your widget misbehaving, and it is not your page. It is Facebook's API - and today we want to tell the whole story: what exactly Facebook returns, why it's been broken for years, how the latest plugin updates work around it, and what to do if your count still needs a nudge.
The mismatch comes in several flavours, and over the years we've seen every one of them: the widget shows fewer reviews than the page; the widget shows a tiny nonsense number like 1 or 2; the widget shows zero; or the number is right today and different tomorrow without a single new review appearing. All of these have the same root cause.
When a widget asks Facebook's Graph API for a page's recommendations, the API returns two things: the list of reviews themselves, and a single field called rating_count - Facebook's own statement of how many ratings the page has. The list is generally fine. The counter is not.
We recently audited every Facebook page connected to our plugins - hundreds of live pages - and compared rating_count against the reviews the API itself delivered in the very same response. The results are remarkable:
rating_count is simply zero - while the reviews are right there in the response.rating_count exactly matched a non-empty review list was… zero. Not once.This is not new. We reported it to Meta back in 2022 (bug 570160061284085), with concrete pages showing 48 in the API against 64 on facebook.com. The ticket was eventually closed with the support channel itself being retired - the bug was never fixed, and our audit shows it has only gotten worse since.
Knowing the counter is unreliable, earlier plugin versions tried to be clever: when the plugin had managed to load all of a page's reviews, it counted them and used that; otherwise it fell back to Facebook's counter. Reasonable on paper - but the two numbers rarely agree, so whenever the plugin crossed from one mode to the other (for example, after a full sync finished in the background), the visible count jumped. And if Facebook's counter happened to be a nonsense 1, that 1 could win.
The latest updates - Business Reviews Bundle 2.0.4 and Facebook Reviews plugin 2.8.1 - rebuild this logic around a simple principle: trust what can be counted, not what Facebook claims.
No settings to change, no cache to clear - existing installations pick the new behaviour up automatically after updating.
There is one thing no plugin can fully work around: for some pages, Facebook's API refuses to deliver every recommendation the page shows. The API might hand over 48 reviews for a page whose public counter says 64 - the remaining ones simply never come through (this affects certain recommendation types and very old reviews). For exactly this case, the Facebook connection has a small option: Rating count adder.
How to use it - once:
facebook.com/<your-page>/reviews) and note the real count.Two things worth knowing about this field in the new versions. First, it always applies now - in older versions it silently stopped working once a page was fully synced, which is precisely when most users tried to use it. Second, it accepts negative values: occasionally the API returns more reviews than the page publicly shows (removed or hidden recommendations can still come through), and entering, say, -5 brings the count back in line.
And because the base now grows on its own, this really is a one-time correction: you're entering the fixed gap between what the API delivers and what your page shows - not the total. When your page collects new recommendations, the widget's count rises with them, and your correction stays valid. No weekly babysitting.
If your count still looks wrong after updating and setting the adder, write to us at support@richplugins.com with your page link - these cases are exactly the data that helps us keep improving this. Thank you for your attention!