Skip to content

Facebook Recommendations Instead of Star Reviews: What Changed and How to Adapt

Facebook business pages have been shifting from star ratings toward a Yes/No recommendation model. That change matters for anyone who collects customer feedback, monitors reputation or reports on review performance. This guide explains the recommendation flow, the new content types around it and the practical choices that follow.

What changed on Facebook business pages

Facebook began rolling out a Yes/No recommendations system to local business pages. The move was understood as replacing the existing star-rating system, although the average rating shown on a page would remain and would draw on past reviews and new recommendations.

The shift reduces the quantitative range from five stars to two choices. That makes the written content of a recommendation more important: a customer can say yes while adding a caveat, or say no while describing a fixable problem. Reputation managers need to read the message, not just the binary answer.

Facebook had previously tested a 10-point review scale, so the direction of travel was not a simple extension of the five-star model. The recommendation approach moved further away from graded ratings.

Rich endorsements: photos, text and attribute tags

The recommendations process introduced rich endorsements, which let users add photos and text. At the time of the change, this function was exclusive to restaurants and coffee shops but was expected to reach other industries.

Rich endorsements make leaving a recommendation easier by suggesting attributes, or tags, for the user to select. The approach resembles the subjective and objective attributes used on Google My Business. For a business, this means the recommendation can carry structured signals alongside free text.

A practical consequence is that reputation managers should watch both the Yes/No answer and the selected attributes. A positive recommendation with a mismatched tag can still misrepresent the experience, and a negative one with specific tags can point to an operational issue.

What the Yes/No model does to review volume and content

Removing the middle of the scale changes who is likely to respond. Previously, a customer with mixed feelings could leave a two- or three-star review. Under a Yes/No model, people without a strong opinion may choose not to respond at all. That could lower the rate of recommendations, although the simpler process could also raise it.

Because there is less grey area in the rating, reputation managers should focus on the content of each recommendation. Expect more messages that begin with a recommendation and continue with a qualification. Those qualifications are where useful customer-experience detail sits.

A recommendation is not the only thing that can appear on the page. Users can answer the question ‘How could [Business Name] improve?’ and publish that message if they choose. They can also send private feedback instead, although that route involves adjusting recommendation privacy and is less straightforward.

Reporting recommendations and handling authenticity

Facebook introduced reporting options intended to keep recommendations authentic. Users and page owners could report a recommendation under categories including:

  • Recommendation not relevant
  • Unfair recommendation
  • Nudity
  • Violence
  • Harassment
  • Suicide or self-injury
  • Spam
  • Drugs
  • Hate Speech

There is no dedicated category for fake reviews left by review networks, former employees or rivals. In practice, the closest available categories are likely to be ‘unfair’, ‘not relevant’ or ‘harassment’, depending on the content.

This reporting route is worth distinguishing from the process for reporting fake Google reviews, which has been described as difficult, clunky and underused. The Facebook categories are broader, so a page owner needs to choose the closest match and explain the problem clearly.

API access and reputation tooling

At the time of the change, Facebook had not opened the recommendations functionality to its API. Once reviews were officially turned off, reputation-management software providers would need to find ways to cover the lost data.

That limitation affects how teams plan their monitoring. A tool that relies on the review API may not receive recommendation data in the same way. Teams should check what their platform actually collects and how it treats a Yes/No recommendation in aggregate ratings.

There is also a reporting question: whether reputation tools would treat a Facebook recommendation as equivalent to a five-star rating, inflating averages, or separate recommendations from the five-star pool. The answer affects any cross-platform rating comparison.

A practical checklist for handling Facebook recommendations

Use this checklist when adapting a business page or client account:

  1. Confirm what your page currently displays. Check whether the page shows a Yes/No recommendation prompt, a star average, or both.
  2. Read the text, not only the answer. Log caveats and improvement requests alongside the Yes/No result.
  3. Watch attribute tags. Note which tags appear and whether they match the written feedback.
  4. Separate public and private feedback. A public recommendation message and a private improvement note are different signals.
  5. Use the closest report category. For suspected fake or unfair recommendations, choose ‘unfair’, ‘not relevant’ or ‘harassment’ as appropriate and include context.
  6. Check API coverage. Ask your reputation platform whether it ingests Facebook recommendations and how it counts them.
  7. Review aggregate reporting. Decide whether recommendations should sit inside or outside a five-star average.
  8. Adjust request wording. Asking for a recommendation is simpler than asking for a star rating, but it may attract more polarised responses.

Worked example: reading a qualified recommendation

Suppose a customer leaves a Yes recommendation with the message: ‘I would recommend this place but the wait on Saturday was long.’ The binary answer is positive, but the text identifies a service issue.

A useful response separates the two signals. The Yes counts toward the recommendation total. The wait-time comment belongs in an operations log and may deserve a public reply that acknowledges the delay. If the same issue appears in several recommendations, it becomes a pattern rather than a one-off.

This is the practical difference between a star rating and a recommendation. A three-star review carried a built-in middle position. A Yes/No answer does not, so the written qualification does more of the work.

Frequently asked questions

What is a Facebook recommendation?

A Facebook recommendation is a Yes/No response to a prompt about a business page. It replaced the star-rating input in the rollout described here. The page can still show an average rating, but that average draws on past reviews and new recommendations.

Are Facebook star ratings gone completely?

The recommendation system was understood as replacing the star-rating input, but the average rating out of five shown on the page would remain. That average would take past reviews and new recommendations into account.

What are Facebook rich endorsements?

Rich endorsements let users add photos and text to a recommendation. They also suggest attributes, or tags, for the user to select. At the time of the change, the function was exclusive to restaurants and coffee shops but was expected to expand to other industries.

How do I report a fake or unfair Facebook recommendation?

Use the reporting options on the recommendation. The categories include ‘Recommendation not relevant’, ‘Unfair recommendation’, ‘Harassment’, ‘Spam’ and others. There is no dedicated fake-review category, so choose the closest match and explain the issue.

Can a customer send private feedback instead of a public recommendation?

Yes. A user can answer ‘How could [Business Name] improve?’ and choose to send private feedback rather than publish it. That route involves adjusting recommendation privacy and is less straightforward than a public message.

Does Facebook offer API access for recommendations?

At the time of the change, Facebook had not opened the recommendations functionality to its API. Reputation-management providers therefore had to account for a gap in review data.