RATINGS/DECODED

Explainers / MECHANICS

Do old reviews still count toward a star rating?

It depends on the platform. Google describes a plain average of all published ratings, so an old review counts like a new one. Trustpilot, Amazon, Tripadvisor and Meta each say newer reviews carry more weight, though none publishes the exact decay.

A restaurant changes chefs. A plumbing company is sold. A hotel is renovated. In each case, reviews from before the change describe a business that no longer exists, and the star rating may or may not have noticed.

What each platform says about time

PlatformStated treatment of older reviews
GoogleNone stated. The score is "the average of all ratings published".
Trustpilot"Newer, recent reviews hold more weight in the TrustScore than older ones."
AmazonModels "consider factors such as how recent the rating or review is".
TripadvisorIn the Popularity Ranking, recent reviews count for more; older ones carry less weight but stay visible.
Meta business ratingsAn average "weighted by recency and other factors".
YelpNot stated for the star rating.
ThumbtackNot stated for the profile rating; reviews must be posted within set windows after the job.

Sources: 11 Google. Google Business Profile Help, Understand review scores for local places and businesses 22 Trustpilot. Trustpilot Help Center, TrustScore and star rating explained 33 Amazon. Amazon Customer Service, Understanding Customer Reviews and Ratings 44 Tripadvisor. Tripadvisor Insights, Everything you need to know about the Tripadvisor Popularity Ranking 55 Meta. Meta Business Help Center, About business ratings and reviews 66 Thumbtack's posting windows. Thumbtack Help, Content and Review Policy

Four of these say recency matters and none gives a half-life. If you want to know whether a three-year-old review counts for half, a tenth or nothing on Trustpilot, the platform does not disclose this.

Why platforms weight recency

The stated reason is the obvious one. Trustpilot calls recent reviews "a good indication of current customer satisfaction".22 Trustpilot Help Center, TrustScore and star rating explained Tripadvisor says newer reviews better reflect the current experience.44 Tripadvisor Insights, Everything you need to know about the Tripadvisor Popularity Ranking

There is a second effect worth understanding. When recent reviews dominate, the score depends on a steady flow. Trustpilot says as much: a TrustScore "is most stable when reviews come in regularly", and "whether or not a company asks for reviews can impact the TrustScore". A business that stops collecting reviews does not freeze its score; the score becomes hostage to whichever few reviewers happen to turn up next. In a recency-weighted system, review volume is a form of insurance.

The plain-average problem

Under Google's stated method, history has full weight forever. Consider a business with 400 reviews averaging 4.7 that goes downhill and collects twenty consecutive two-star reviews. The new average is (400 × 4.7 + 20 × 2) / 420, which is 4.57. Displayed to one decimal, it has slipped from 4.7 to 4.6 while every recent customer has been unhappy.

That inertia cuts both ways. A business that had a terrible first year and then fixed its problems drags the old reviews behind it indefinitely.

The count in parentheses is your tool here. A large count on a plain-average platform means the displayed number is slow to react, so sorting by newest becomes more important as the count grows.

Ratings drift on their own

Even when the business does not change, ratings tend to move with time and sequence. Godes and Silva studied book ratings and separated two processes: one tied to how long an item has been available and one tied to the order of reviews. Controlling for time, they found ratings decrease with sequence, so that the nth rating is on average lower than the one before it.77 Godes and Silva, Marketing Science 31(3), 2012, Sequential and Temporal Dynamics of Online Opinion

Li and Hitt offer one mechanism: early buyers are not typical buyers. If early customers hold different preferences from later ones, early reviews will mislead later readers, who do not fully correct for the difference.88 Li and Hitt, Information Systems Research 19(4), 2008, Self-Selection and Information Role of Online Product Reviews The first fifty reviewers of a new restaurant are often the neighborhood's enthusiasts. The next five hundred are everyone else.

Moe and Trusov add a social layer. They found that rating behavior "is significantly influenced by previously posted ratings", although the sales effects of these dynamics were relatively short lived once indirect effects were considered.99 Moe and Trusov, Journal of Marketing Research 48(3), 2011, The Value of Social Dynamics in Online Product Ratings Forums People write differently when they can see what others wrote.

Put these together and a gentle downward slope in a young business's rating is normal. It does not prove decline. A sharp break, with a clear before and after, is what deserves attention.

How to read recency yourself

  1. Sort by newest and read the last ten to twenty. Compute a rough average in your head. Compare it with the headline.
  2. Find the date of the oldest review on the first page. If twenty reviews span four years, the business gets about five reviews a year and the profile is thin, whatever the total count says.
  3. Look for step changes. Owner replies that mention "new management" or reviews that say "it used to be great" mark the point where older reviews stop being relevant.
  4. On Tripadvisor, check the ranking and the rating separately. The Popularity Ranking is the recency-sensitive figure, and it is recalculated daily.44 Tripadvisor Insights, Everything you need to know about the Tripadvisor Popularity Ranking
  5. Discount very old praise for staff-dependent services. A salon review is about a stylist. A review of a hardware store's opening hours ages more slowly.

Questions readers ask

Do Google reviews expire?

Google's help pages describe no expiry. The score is an average of all published ratings, and a rating stays published unless it is removed for a policy violation or deleted by its author.11 Google Business Profile Help, Understand review scores for local places and businesses

Is a business with only old reviews a bad sign?

It is an information gap more than a bad sign. It may mean the business does not ask for reviews, or that it has slowed down. On recency-weighted platforms it also means the score can swing sharply with the next review.

How recent is recent enough?

For services that depend on particular people, the past year is a reasonable window. For anything that has changed ownership, only reviews after the change tell you much.

Method Each recency statement is taken from the platform's help page. The discussion of rating drift uses two peer-reviewed studies of how ratings change over time. See how we source and check.

Sources

Every source below was opened and read for this page. Links go to the publisher's own copy.

  1. Google Business Profile Help. Understand review scores for local places and businesses
  2. Trustpilot Help Center. TrustScore and star rating explained
  3. Amazon Customer Service. Understanding Customer Reviews and Ratings
  4. Tripadvisor Insights. Everything you need to know about the Tripadvisor Popularity Ranking
  5. Meta Business Help Center. About business ratings and reviews
  6. Thumbtack Help. Content and Review Policy
  7. Godes and Silva, Marketing Science 31(3), 2012. Sequential and Temporal Dynamics of Online Opinion
  8. Li and Hitt, Information Systems Research 19(4), 2008. Self-Selection and Information Role of Online Product Reviews
  9. Moe and Trusov, Journal of Marketing Research 48(3), 2011. The Value of Social Dynamics in Online Product Ratings Forums