How BrokerBase scores work: editorial, user and risk signals explained

Three independent signals power every profile: an editorial score from our analysts, a Bayesian-adjusted user rating, and a trust score computed from regulation and complaint data. Here is exactly how each one is built.

Every institution profile on BrokerBase shows three separate numbers, and we deliberately never merge them into one “overall” score. Blended scores are easy to game and hard to interpret; separated signals let you weigh what matters for your situation.

The editorial score (0–100)

Written by our analysts after testing the product: account opening, deposits, spreads during London and New York sessions, withdrawal turnaround, and support quality. Editorial scores are updated at every re-review and the date is shown on the profile.

The user rating (1–5 stars)

The star average is the plain arithmetic mean of published reviews. For ranking we additionally compute a Bayesian-adjusted value that blends an institution’s reviews with the platform-wide mean until it accumulates enough volume — this prevents five fresh 5-star reviews from outranking three hundred steady 4.3-star ones. Both numbers are always displayed together.

The trust score (0–100)

Computed automatically from verifiable facts: license quality (tier-1 licenses dominate), revoked or suspended licenses, official regulator warnings, complaint volume and resolution rate, and years in operation. The formula is deterministic — the same inputs always produce the same score — and sponsorship has no input into it whatsoever.

What sponsorship buys (and what it cannot)

Sponsors get a labelled placement. They cannot buy score changes, review removals, complaint deletions or warning suppressions. Every sponsored placement carries a visible label.