Reputation math: how one star of rating change affects revenue for a $500k/year local business

You've seen the number before: one star of rating change is worth 5-9% of revenue. It shows up in nearly every reviews-management article written in the last decade, almost always as a footnote, almost never with the math worked through for a specific business. The number is real — it comes from a well-built Harvard study — but citing it isn't the same as understanding what it means for you.

This article does the math, honestly, for a single-location business doing $500,000 a year. You'll get the anchor number and exactly where it came from, the caveats the original researcher put on his own finding, a worked example at each half-star band, why the relationship between rating and revenue is stepped rather than smooth, and a break-even calculation on the time it actually takes to move your rating through asking.

One thing worth naming before you get into the arithmetic: not every rating shift is something you did. Sometimes a rating moves because Google removed a review — when Google removes a review you didn't choose, the math below still applies, but the direction of the shift isn't in your control. Keep that in mind as you read the rest of this as a planning tool, not a promise.

The one number everyone cites, and where it actually came from

The number traces back to Michael Luca, an economist at Harvard Business School, in Working Paper 12-016 — first published in 2011 and revised in March 2016. Luca combined Yelp review data with tax filings from the Washington State Department of Revenue for restaurants in Seattle. Because Yelp displays a restaurant's rounded average rating rather than the precise decimal, Luca used a regression discontinuity design around the rounding thresholds to isolate the causal effect of rating on revenue, separate from the effect of quality itself.

The finding, verbatim from the paper's abstract: a one-star increase in Yelp rating leads to a 5-9 percent increase in revenue. That's the anchor number this entire article works from.

Two caveats sit inside the paper itself, and you should carry both of them with you.

First, the effect is driven entirely by independent restaurants. Chain restaurants showed no measurable revenue response to rating changes at all. Luca's interpretation is that chains carry brand reputation built through advertising and familiarity, so a Yelp rating just doesn't move the needle the same way for a diner choosing between a national brand and a table they already know.

Second, the signal only works when it carries information. Luca found that a rating change had roughly 50% more revenue impact for a restaurant with at least 50 reviews than for one with fewer than 10. A high rating built on a handful of reviews doesn't move customer decisions the way a high rating built on genuine volume does — the number needs weight behind it.

There's also an honest transfer problem worth naming directly: Luca's data is Yelp, restaurants, and Seattle tax records from 2003 to 2009. It isn't Google, it isn't 2026, and it isn't your industry. Nobody has replicated a study of this quality specifically for Google Business Profile ratings across categories. The 5-9% figure is the best empirical anchor that exists, but treat it as directional — a working estimate for planning, not a guarantee for your business specifically. The direction has held up in every follow-up study that has looked at the question since, and the magnitude is defensible. The rest of this article treats 5-9% as the working number.

Applied to a $500k/year business, half-star by half-star

Picture a single-location, independently owned business doing $500,000 in annual revenue — a salon, a small dental practice, a family restaurant, a home services company. For the Luca finding to apply cleanly, assume it's independent, not a franchise, and assume it already has at least 50 Google reviews, since that's the threshold where the rating signal carries real weight.

Applying the 5-9% band to $500,000 in revenue:

• One full star = $25,000 to $45,000 per year
• Half a star = $12,500 to $22,500 per year
• A quarter star = $6,250 to $11,250 per year

Moving between common half-star bands looks like this:

• 3.5★ to 4.0★: $12,500 to $22,500 per year
• 4.0★ to 4.5★: $12,500 to $22,500 per year
• 4.5★ to 5.0★: $12,500 to $22,500 per year, though a true 5.0★ is rarely reached with real review volume — a 4.9★ sitting on hundreds of reviews reads as more credible to a consumer than a suspiciously perfect 5.0★ with only a handful behind it

The precision multiplier from Luca's paper matters here too. His finding of a 50% stronger effect above 50 reviews means the arithmetic above sits closer to the top of the range if you already have review depth. A business with 12 reviews sitting at 4.0★ won't see the same lift moving to 4.5★ that a business with 200 reviews sees making the identical jump, because the rating carries less trust when the sample behind it is thin.

There's also a real limit to what this arithmetic tells you. It doesn't tell you which specific customer decisions produce the revenue — some of it is new customers choosing you over a competitor, some is existing customers returning more often, some is word-of-mouth accelerating on top of a better public number. It doesn't tell you the timeline either. Luca's underlying data is quarterly; a rating change doesn't produce a revenue jump next week, it shifts conversion gradually as the new number surfaces across more searches over months. And it doesn't account for category — a high-consideration purchase like a dental procedure or a home renovation probably weights star rating more heavily in the decision than a quick coffee run does, and Luca's paper doesn't isolate that effect at the category level.

Treat the numbers above as a working range for planning, not a forecast you can bank on.

The math isn't linear — it's stepped

Half-star arithmetic makes the relationship look smooth. It isn't. BrightLocal's Local Consumer Review Survey 2026, based on 1,002 US adult consumers surveyed in February 2026, shows exactly where the real cliffs sit:

• 31% of consumers will only use businesses rated 4.5★ or higher — up from 17% in 2025, nearly doubling in a single year
• 68% of consumers will only use businesses rated 4.0★ or higher — up from 55% in 2025
• 92% of consumers say star ratings affect their decision when choosing a local business

Here's why that changes the arithmetic from Section 2. Moving from 3.9★ to 4.0★ crosses the 68% threshold — you go from invisible to 68% of the consumer market to visible to them, in one-tenth of a star. Moving from 4.4★ to 4.5★ crosses the 31% threshold the same way. Moving from 4.1★ to 4.3★ crosses neither cliff at all. Same size of rating change, wildly different revenue consequence depending on where it lands.

For a $500,000/year business, the practical read looks like this: if you're currently at 3.8★, getting to 4.0★ is worth substantially more than the raw per-half-star arithmetic suggests, because it opens the 68% consumer pool that was ignoring you entirely below that line. If you're at 4.3★, getting to 4.5★ opens the 31% pool the same way. If you're at 4.1★, the work to reach 4.4★ is real, but the revenue lift is smaller, because you're moving within a band instead of crossing a cliff.

The threshold logic runs in both directions, and this is where losing a review costs more than gaining one seems to. A single review that drops you from 4.5★ to 4.4★ costs more than a review that drops you from 4.3★ to 4.2★, because the first one pushes you back across the cliff and the second one doesn't. The deeper breakdown of what a single lost review actually costs walks through that asymmetry in full.

Recency adds a second dimension on top of the threshold effect. The same BrightLocal survey found that 74% of consumers only care about reviews from the last 90 days. A 4.7★ average built from 200 reviews with nothing new in six months reads worse to a consumer today than a 4.4★ average built on a steady weekly stream of new reviews. Old ratings decay, even when the number itself hasn't moved.

The break-even on time spent asking

Take the arithmetic from Section 2 again: half a star is worth $12,500 to $22,500 a year for a $500,000 business. The next question is what it actually costs, in hours, to produce that half-star.

Here's a realistic scenario. Say the business serves 50 review-worthy customers a month, and well-timed review asks — made in person or by text, with a direct link, at the right moment — convert at roughly 20%, well below the 60%+ figures that badly-timed-ask articles tend to over-claim. That's 10 new reviews a month, or 120 a year. Starting from 4.0★ on 80 existing reviews, 120 new reviews averaging 4.7★ pulls the running average toward 4.5★ over the course of a year.

The time cost of that asking, measured honestly: about 30 seconds per customer to make the ask and hand over a review link. Fifty customers a month, twelve months, thirty seconds each comes to roughly 5 hours a year of direct asking time. Add another 2 hours upfront for setting up templates, a QR code, and a follow-up sequence, and the total lands around 7 hours a year of review-request labor.

Divide the revenue: $12,500 to $22,500 in attributable annual revenue across roughly 7 hours of work comes out to somewhere between $1,780 and $3,200 an hour. That's about as high-value an hour as a small business owner can spend — not because asking for reviews is glamorous, it isn't, but because it compounds. A rating built to 4.5★ this year keeps working next year without additional effort behind it.

The math breaks in a few predictable ways. If the asks are generic — "please leave us a review" with no specifics — conversion drops to 3-5% and the arithmetic above collapses. If the underlying service quality isn't there, the reviews that come in average 3.5★ instead of 4.7★, and the rating moves the wrong direction regardless of how many asks go out. And if there's no actual system behind it — no template, no link, no follow-up — the habit gets forgotten by week two of month two, and the 7-hour estimate balloons into something closer to 20 hours of remembering, forgetting, and re-starting.

That last failure mode is the reason I built Ominvo. The math above justifies systemizing the ask — not because asking itself is hard, but because remembering to ask consistently, for every customer, every single day, is the part that quietly stops happening once the first few weeks of enthusiasm wear off. Ominvo runs the follow-ups so the 7 hours stays 7 hours instead of turning into 20 hours of manual reminders, and if the numbers in this article hold up for your business, that's what pricing covers. The full breakdown of how to structure the ask itself — timing, wording, in-person versus digital — covers the mechanics this section assumes.

What the math doesn't cover

The 5-9% figure is a working estimate, not a forecast — it's the best empirical anchor available, built on Yelp restaurant data from over a decade ago, applied here to Google reviews across other categories. The threshold cliffs at 4.0★ and 4.5★ are real, and they move: the 4.5★ cliff nearly doubled in size in a single year. The break-even math in Section 4 assumes the underlying service is actually worth reviewing well — a bad customer experience run through a well-built review-ask system just produces bad reviews faster and more consistently.

The value in doing this math isn't precision. It's what the number does to the decision. If a half-star is worth roughly $15,000 a year and about 7 hours of consistent asking, building a system to do that asking stops being a debate and starts being an obvious call.