Reviews and online reputation: the numbers that contradict what almost everyone does

What matters more: the rating or how fresh the reviews are?

Freshness matters more, and the field evidence is uncomfortable for anyone who has spent years showing off a 4.8.

A 90-seat grill house in Bogotá lost 11% of its visits year over year with the same kitchen and the same chef, and the finding sat in a Google listing that had gone fourteen months without a new review, where the last seven talked about parking. Some 76% of diners check reviews before booking or ordering, according to BrightLocal 2026, and the real cutoff is 4.0: below that, the listing stops making the short list before your food even exists for that customer. Concrete decision: if your latest review is older than 45 days, this week the problem is not the menu, it is review flow, and you fix it with a monthly target and a named owner. Twenty-five reviews a month tell the local algorithm that the business is alive, and that freshness signal moves map ranking faster than pushing the rating from 4.4 to 4.6.

Monthly flow beats the decimal point

Run the math on your own count: going from 4.4 to 4.6 with 300 accumulated reviews takes roughly 120 straight perfect ratings, four or five months of flawless work, while a flow of 25 a month comes from a card on the check and a trained server. The rating is a weighted average you do not control; cadence you do. I got this wrong for years, recommending campaigns to chase tenths of a point, when the number that moves cash is how many reviews came in last quarter. Operating target: new reviews per week on the morning board, next to food cost. The listing card shows three things together — rating, delivery time and shipping cost — and there the star behaves like a price variable. Moving from 4.2 to 4.6 on a delivery app shifts conversion by 8 to 14 percentage points depending on the listing, with no menu change and no discount, and that happens over a market that reached close to US$96 billion in prepared-food delivery in the United States (Statista 2024).

In delivery the rating is price in disguise

Translate it: a location billing US$40,000 monthly in delivery at 3% conversion that climbs to 3.3% adds roughly US$4,000 a month with zero kitchen investment. That is why reputation gets managed with a calendar, a target and an owner, exactly like inventory, and not as customer service. A restaurant sitting at exactly 5.0 sells less than one at 4.6, and the reason is trust, not quality. Diners discount the perfect score because it smells of filtering, bought reviews or relatives chiming in, whereas a 4.6 with answered negative reviews signals a real business that sometimes fails and fixes it. The sweet band runs from 4.3 to 4.7. Tension resolved: you want high stars, yet credibility is bought with visible imperfection, and the bridge between the two is the REPLY. A two-star complaint answered inside 24 hours with a concrete fix — not the we appreciate your feedback template — converts the new reader, who is almost never the person who complained.

The paradox of the perfect 5.0

Concrete decision: ban filtering requests to happy customers only, and answer 100% of reviews under 3 stars. Set against the channels you actually pay for, reputation is the cheapest asset in your funnel. Email marketing returns US$36 for every US$1 invested with a 43.6% average open rate in restaurants (Stripo 2025), though its click sits around 1.06% and click-to-open at 3.28%, among the lowest by industry (Mailchimp 2025). SMS opens at close to 98%, with 90% read within one to three minutes (Constant Contact 2024) and 18% clicking through (Tabular 2025). Influencer work pays between US$5.78 and US$7.65 per dollar per Socially Powerful 2025 and iQFluence 2026. Every one of those channels talks to someone who already knows you; the review talks to someone who does not. Allocation rule: if the 43.6%-open email has budget and reviews have no owner, your funnel is upside down.

How to read these numbers in YOUR operation?

Three scenarios, three different targets. Small single-location spot, up to 60 seats: 12 to 18 monthly reviews are enough to hold the freshness signal, with a shift lead who asks for the review when handing over the check;

zero budget, real cost fifteen minutes a day. Mid-size restaurant with active delivery, 60 to 120 seats: aim for 25 a month and watch each app's rating separately, because a 4.2 on the delivery platform costs you more cash than a 4.5 on Google. Group of three locations or more: consolidate a weekly board with new reviews, 90-day rolling rating and average reply time per site, then tie the manager's bonus to that figure rather than to the historical rating, already frozen by accumulated volume. Methodological honesty first: the channel figures cited here come from public reports by Stripo 2025, Mailchimp 2025, Constant Contact 2024, Tabular 2025, Socially Powerful 2025, iQFluence 2026 and Statista 2024, aggregators that average thousands of accounts across different industries, not your city and not your average check.

Where these benchmarks come from and what they leave out?

The 43.6% open rate blends large chains with independents; the US$36 per dollar return excludes the cost of the team writing the emails.

Delivery conversion ranges come from behavior observed by listing and shift with category and time of day. Use them as an order of magnitude to set priorities, never as a promise. Your only clean data point is your own: measure four weeks of review flow and delivery conversion before and after, and that number rules. Online reputation is not customer service and not putting out fires on social media; it is the top layer of the funnel that decides how many people ever get to see your menu. Diego F. Parra works it at Masterestaurant with the same discipline as inventory: calendar, monthly target, a named owner and one figure on the board. Follow the counterfactual all the way.

Reputation as the top layer of the funnel

If you do nothing for six months, your listing accumulates rising average age, drops in local ranking, makes the short list less often among the 76% who check before deciding, loses organic map traffic, and then you buy ads to patch the hole, which raises your acquisition cost precisely when margin is already tight. Start tomorrow: set the weekly review target and hang it on one person's name. The difference is flow, not stars. A listing pulling 25 reviews a month tells the local algorithm the business is alive, and that freshness signal outweighs climbing from 4.4 to 4.6, a move that takes months and depends on decimals nobody on your team controls. On delivery the effect is harsher and faster, because the rating shows up on the listing card right next to delivery time and fee. Going from 4.2 to 4.6 inside a delivery app shifts conversion by 8 to 14 percentage points depending on the listing, with no menu change and no discount.

Where the real money difference sits?

There is a trade paradox worth settling here: the restaurant at a clean 5.0 sells less than the one at 4.6.

Guests discount perfection because it smells bought or filtered, and e-commerce behavior research puts the conversion sweet spot between 4.2 and 4.7. Chasing 5.0 means chasing a number that works against you. The third gap is guest LTV. A customer who wrote a review and got a reply using their name comes back more often, and that repeat visit is worth more than the first: at an 18 USD check and 2.4 extra visits a year, every well-answered review carries 43 USD of incremental sales with zero ad spend. Then there is the creative use, where almost nobody looks. Reviews are the only free, literal, date-stamped market research your restaurant owns: the exact words your guest uses to describe your dish are the best TikTok script you will ever get, far better than any agency idea.