Open a listing with a 4.8 average and the first instinct is to trust it. Then look at the dates. If the newest review is from two years ago, the number stops meaning what it looked like it meant. A lifetime average carries no expiry date, but readers supply one anyway: has anyone actually been here lately?
That question is not answered by the star average at all. It is answered by the dates next to the reviews, a different field entirely, sitting in different software, and reported nowhere on most dashboards.
What a reader actually does with the dates
Nobody reads a profile top to bottom and averages it in their head. They read the first few visible reviews and form an impression from those, then glance at the dates almost as a formality, to confirm the impression is current. A business with a strong average but a stale top of page fails that formality check. The reviews might be glowing, but they read like a photograph of a place that has since changed hands or closed.
This is also how the platforms themselves behave, not just how people do. Google has said publicly that review signals, including how recently and how often a business receives them, factor into local search ranking. So staleness has a second, separate cost: it is a ranking problem before the reader ever reaches the page, not only a trust problem once they do.
Put those together and the lifetime average becomes close to the least informative number on the profile. It tells you what happened, summed across however many years the business has been collecting reviews. It says nothing about what is happening now, and now is the only thing a prospective customer can act on.
The number worth tracking instead
The useful figure is a trailing window, not a lifetime one: the average rating over roughly the last ninety days, alongside the date of the most recent review at each location. Ninety days is long enough to smooth out a bad week, short enough that it moves when something in the business actually changes.
The trailing-90 figure catches two things a lifetime average is built to hide. It catches decline early: a location sliding from 4.6 to 4.1 over a bad quarter is invisible in a lifetime average sitting on years of good history, but it shows up immediately in the trailing number. It catches improvement the same way. A location that fixed a real problem three months ago is still dragging a mediocre lifetime average behind it, sometimes for years, purely because of the volume of older reviews underneath. A rising trailing-90 next to a flat lifetime average is good news a lifetime-only dashboard actively hides from the team that earned it.
This is a different lens from review count, which is its own credibility signal. Volume tells a reader the average is backed by enough data to trust. Recency tells them the business behind that average is still the business operating today. A profile can be strong on one and weak on the other, and the fix for each differs.
What to measure per location
A chain with more than a handful of locations cannot watch this by eye across every site, so it has to be something a report surfaces automatically, per location, without anyone opening each profile in turn.
- Trailing 90-day average rating, alongside the lifetime average, so a gap between the two is visible at a glance
- Date of the newest review at each location, flagged once it passes a set threshold, such as 30 or 45 days
- Reviews per month, tracked as a rate rather than a running total, so a quiet month shows up the month it happens
None of these need a new number system, just the same review data most tools already collect, reported on a rolling window instead of collapsed into one all-time figure. Once a location's newest review is visibly stale, the fix is usually simple: ask more customers, more consistently, rather than relitigate reviews from years ago. A trailing metric turns "keep the average up" into a habit with a deadline, which is the version of the goal that actually gets acted on.