How the BSU Score Works — Our Rating Methodology Explained

The BestServiceUAE Score — BSU Score for short — is a single 0–100 number assigned to every active workshop in our directory. It exists to do one job: let an owner glance at a list and immediately see which shops have the strongest combination of public signals, without having to compare ratings, review counts, and profile completeness manually.

This article explains the formula in full. No black box, no proprietary secrets — the whole calculation lives in scripts/calculate-scores.ts in our parser, runs every weekly sync, and is reproducible from the same Google Maps data anyone else could access.

The four input signals

Every workshop is evaluated on four signals, each contributing a fixed maximum to the final score:

That sums to 100. A small micro-differentiator (under 2 points) is added to break ties between otherwise-identical scores.

How rating is scored

The rating component is a straight-line scale: (rating ÷ 5) × 40. A 4.5★ workshop earns 36 points; a 4.0★ earns 32; a 3.5★ earns 28. We don't apply any adjustment for review volume here — the rating signal is the rating signal, full stop. Volume is handled separately.

Why review volume is logarithmic

The review-volume score is min(30, log₁₀(reviews) × 10). This matters because review counts in the UAE auto market are deeply skewed: a typical workshop might have 30 reviews, while a hub workshop in Al Quoz can have 3,000. A linear scale would let those mega-hubs swamp every other signal and dominate the rankings just by having existed longer.

The logarithmic curve says: going from 10 to 100 reviews matters a lot (10 points), going from 100 to 1,000 matters less (also 10 points), and going from 1,000 to 10,000 matters even less. By the time you cap out at 30 points, you've reached 1,000 reviews. Above that, volume stops being meaningful — the rating is the rating.

Profile completeness — fifteen points for actually being open

Five binary checks, three points each:

• Phone — a working number listed on Google.
• Website — any external site, even a single-page Linktree counts.
• Hours — published opening hours.
• Main photo — at least one photo of the workshop.
• Verification — a Google-verified business listing.

This is the boring part of the score, but it's also the most predictive of whether you can actually get hold of the workshop. A 4.8★ shop with no phone, no hours, and no website is functionally invisible — the score reflects that.

Quality bonus — the up-to-15 point reward

This is the only conditional part of the formula. It rewards workshops that hit both rating and volume thresholds simultaneously. The logic, in plain English:

• If rating ≥ 4.5 and reviews ≥ 50: 12 points base, scaling up to 15 with more reviews.
• If rating ≥ 4.0 and reviews ≥ 20: 8 points base, scaling up to 12.
• Otherwise: a partial credit that scales with how close the workshop got to those thresholds.

The point of the quality bonus is to separate workshops that are merely complete from workshops that are also demonstrably good at what they do. A 4.8★ / 200-review shop and a 4.8★ / 8-review shop are not the same, even if their average rating is identical — the volume tells you whether the rating is statistically meaningful.

Tier thresholds

Once the 0–100 score is calculated, workshops are bucketed into four tiers:

The thresholds aren't arbitrary — they roughly correspond to the natural break-points where workshops separate by overall public signal strength. Anything Platinum has cleared all five completeness checks, has a strong rating, has enough reviews to make that rating credible, and has earned the quality bonus.

Why we don't show AggregateRating until ten reviews

Schema.org AggregateRating is what surfaces the gold star count in Google search results. We deliberately omit it from any workshop with fewer than ten reviews. Three reasons:

• A 5★ average from two reviews is not a meaningful signal — it's a coin flip.
• Google's own guidelines warn against publishing AggregateRating from fake or synthetic sources.
• Including it indiscriminately would let new or thin profiles hijack search visibility on the strength of statistically meaningless data.

The threshold is conservative on purpose. Some workshops with 5–9 genuinely positive reviews don't get the rating snippet — we'd rather under-credit a real shop than risk over-crediting a fake one.

Why we don't accept payment

BestServiceUAE has no "featured" tier, no sponsorship slots, no pay-to-rank arrangement. The scoring code doesn't have a flag for it because there's nowhere to insert payment into the formula — every input is a public Google Maps signal that anyone can verify independently.

This isn't a marketing line. It's a structural choice. The directory is funded by display advertising on aggregate listing pages and by an eventual workshop-management subscription product (separate from rankings). Workshops can claim their listing, edit their description, and respond to reviews. They cannot pay to move up in the ranking, and they can't pay to remove a competitor from it either.

The data refresh cycle

The full pipeline runs weekly:

1. Pull fresh data from Google Maps for every active workshop.
2. Re-calculate ratings, review counts, and profile signals.
3. Re-run the BSU Score formula across the entire directory.
4. Re-categorize tier assignments.
5. Rebuild the static site and redeploy.

That means a workshop's position can shift up or down between weeks as reviews accumulate, hours change, or Google verification status changes. There's no editorial finger on the scale between syncs — the data wins.

An example, walked end-to-end

Take a hypothetical Dubai workshop: 4.7★ rating, 280 reviews, has a phone, website, hours, and photo, and is Google-verified.

• Rating: (4.7 ÷ 5) × 40 = 37.6
• Volume: log₁₀(280) × 10 = 2.45 × 10 = 24.5
• Completeness: 5 checks × 3 = 15
• Quality bonus: rating ≥ 4.5 + reviews ≥ 50 path → 12 + log₁₀(280/50) × 3 = 12 + 2.25 = 14.25 (capped at 15)
• Micro-differentiator: ~0.5

Total: 91.85 / 100 → Platinum tier.

The same workshop with only 12 reviews instead of 280 would score 37.6 + 10.8 + 15 + 4.5 + 0.5 = 68.4 → Silver. Same rating, same profile completeness — but the lower volume materially changes the credibility of that rating, and the score reflects that.

What the score doesn't capture

Three honest limitations:

• Specialty fit. The BSU Score doesn't know whether a workshop is good at your car. A 92-score general-repair shop may not be the right call for a Lamborghini. Use the brand and category pages to filter first.
• Recent quality changes. A shop that was great two years ago and has slipped lately is partly captured by the rating, but old reviews still count toward volume.
• Real-world experience. Public signals are public signals. Visit the workshop, get a written estimate, ask questions — the score is a starting filter, not a substitute for due diligence.

For the full input data, formula source, and tier-assignment logic, the canonical reference lives at the methodology page. The top end of every category is permanently visible at the top-rated list, and the broader site philosophy is on the about page.