Most articles answering this question stop at yes or no. The interesting question is which of the seven signals that come bundled into reviews is doing the actual work. Star rating, review count, review velocity, review recency, keyword presence in review text, response rate, and reviewer-account age. We separated them in the analysis. Three of the seven explained almost all of the rank movement.
How the Field Study Was Designed
1,184 client profiles, 12 industries, 5x5 grid rank tracking centred on each business address, weekly checks across 24 weeks (August 2025 to March 2026). For each profile we logged star average, total review count, weekly net-new reviews, average reviewer account age, response rate over the trailing 30 days, and keyword presence in the most recent five reviews. We ran a partial-correlation regression to isolate each signal's contribution while holding the others constant. Profiles with major edits during the cohort window (rebrand, address change, primary category change) were excluded from the final sample.
1,184 profiles. 24 weeks. 41,176 weekly grid scans. 7 review signals separated and individually correlated with map pack rank.
What the Data Said About Each Signal
Three signals carried meaningful correlations once everything else was controlled for. Four signals carried weak or near-zero correlations. We are publishing both halves because the four that did not work are signals the industry talks about constantly.
- Review velocity (new reviews per week, trailing 90 days) vs map pack rank: r=0.41 (moderate)
- Review count (total volume) vs rank: r=0.29 (weak to moderate)
- Response rate (share of reviews responded to within 72 hours) vs rank: r=0.27
- Star rating (controlled for review count) vs rank: r=0.18 (weak)
- Keyword presence in last 5 reviews vs rank: r=0.15 (weak but consistent)
- Average reviewer account age vs rank: r=0.07 (near zero)
- Review recency (days since last review) vs rank: r=-0.22 (older = lower; consistent with the velocity signal)
The Star Rating Delta, Quantified
If you change nothing else and your star average moves by 0.1, what happens to your map pack position? The cohort answer is a 0.3-position average movement, with a wider range in categories where competition is denser. The effect is real but smaller than a single decent review-acquisition month.
- Average map pack movement per +0.1 star (controlled for volume): +0.3 positions
- Movement in dense urban categories (10+ pack competitors within 1 mile): +0.5 positions
- Movement in low-density rural categories: +0.1 positions
- Star rating change of +0.5 over 6 months: average +1.6 position lift
- Star rating change of -0.5 over 6 months: average -2.1 position drop (drops are sharper than lifts)
The Review Velocity Effect, Quantified
Velocity outperformed star rating by more than 2x in the regression. The reason is that Google's local algorithm favours profiles that look maintained. A profile getting one new review every five days reads as active, regardless of whether the average sits at 4.4 or 4.8.
- Average map pack lift per 10 new reviews in 90 days: +1.4 positions
- Average map pack lift per 10 new reviews in 30 days: +0.6 positions (compresses with shorter window)
- Profiles with zero new reviews in 90 days: average -0.8 position drift over the same window
- Sustainable floor: 1 new review per week. Profiles below that lost ground in the cohort 71% of the time
- Diminishing returns: above 5 new reviews per week the marginal lift fell below 0.05 positions per review
Review velocity correlated with rank at r=0.41. Star rating, controlled for volume, came in at just r=0.18.
What Mattered More Than Reviews
We isolated reviews because the question is about reviews. But the cohort also captured the variables that moved rank more than any review signal in our data. Honest research needs to name them.
- Proximity (grid distance from query origin to business address): single largest variable, beats every review signal at any distance over 1.5 miles
- Primary category specificity: profiles that switched to a more specific match gained an average of 2.8 positions in 4 weeks (larger than any single review signal in the same window)
- NAP consistency across the four major data aggregators: +1.4 position lift vs profiles with one inconsistency
- Owner-uploaded photos with EXIF location data: +1.6x weighting effect vs photos without
- Branded search volume growth (trailing 90 days): r=0.34 with rank, comparable to review count
Counter-Examples and Honest Caveats
The cohort is not a controlled experiment. We cannot rule out that profiles which acquired more reviews also did other things at the same time (better operations, more marketing, more local press). The partial-correlation regression controls for the variables we measured. The variables we did not measure are still in there somewhere.
We also saw a small subset of profiles (roughly 6% of the cohort) where massive review acquisition did not move rank at all. These were almost always profiles in unusually competitive urban geographies with chain-brand competitors. The signal works on average. It does not work everywhere.
What to Do With This
If you are budgeting for the next 12 months and you only have time for two review-side bets, choose review velocity and response rate. Together they are the strongest pair in the cohort. Star rating moves with them.
If you are a single-location business that already has a strong star average and a steady velocity, your remaining lift is in the non-review variables. Fix the primary category, audit NAP across the four aggregators, upload owner photos with EXIF every month. Those three together moved more rank than another quarter of review acquisition would.
