Our rating method

How a database gets on the list

DataBuss doesn't review databases because a vendor asked to be included. Candidates surface during ordinary research: a tool mentioned in a client's stack, a name that keeps coming up while comparing directories for an unrelated task, or a database flagged in reader mail sent to the editors. None of that guarantees a spot on the list, though. Before a write-up is drafted, someone on the editorial side has to open the database and run it through a set of real search queries, because a homepage and a pricing page rarely say much about how the search actually behaves.

A candidate has to meet a few basic conditions before it's considered a fair match for the rating at all.

• It's a database of companies rather than a general-purpose data tool with company records as a side feature.
• It offers some form of search or filtering, not just a static export file or a single bulk download.
• It's accessible enough that editors can test it directly, whether through a free tier, a trial, or a demo account.

The 44 criteria and their weight

Every database in the rating is scored against the same 44 criteria, which is what makes two write-ups genuinely comparable even when the databases themselves serve very different industries. Those 44 criteria sit within ten broader groups, and each group carries its own weight in the final number rather than counting equally.

• Database Coverage — 10%
• Data Quality — 10%
• Search Capabilities — 15%
• Integration — 10%
• Verification — 10%
• Compliance — 5%
• Pricing — 10%
• User Experience — 10%
• Support — 10%
• Additional Features — 10%

Where the data comes from

Most of the figures behind a review come from direct use of the database itself: search results logged against real queries, filters tested one by one, export formats opened and checked, support tickets sent and timed for a response. Pricing is read off the vendor's own published plans at the time the review is written, not from a sales call or a quote that might change by the next inquiry. Claims made in a database's own marketing are checked against what the interface actually delivers, and the two don't always line up.

Some information simply isn't available through normal use. A vendor might hide pricing behind a 'contact sales' form, decline to disclose how often its records are refreshed, or keep its data sourcing methodology undocumented. When that happens, the gap isn't filled in with a guess. The relevant criterion is scored conservatively, the write-up notes the item as unverifiable or 'pricing on request', and the final score reflects that missing piece rather than papering over it.

How the final score is calculated

Each of the 44 criteria is scored on its own, based on what testing actually showed rather than on impressions from a product page. Those individual scores are then averaged within their group, so the Search Capabilities group, for instance, ends up as a single figure built from every criterion that sits under it. Each group's average is multiplied by its assigned weight, and the ten weighted results are added together to produce the database's overall score.

Because Search Capabilities carries 15% against the 5% assigned to Compliance, a database with a clever query engine but a thin compliance story will still outscore one with the reverse profile, and that's by design: search quality tends to matter more to day-to-day use than a compliance checklist most buyers never read closely.

Keeping the rating current

The rating isn't rebuilt on a fixed calendar, and that's intentional. A review gets revisited whenever the database itself changes — a new pricing tier, a redesigned search interface, a feature that quietly disappeared since the last pass — rather than being left untouched simply because a year hasn't gone by yet. Older write-ups that haven't changed in a while aren't necessarily outdated; they're databases that, as far as the editors can tell, haven't changed either.

Anyone who spots something that looks out of date, or who runs a database differently from how a write-up describes it, can flag it by email at reviews@databaseofcompanies.online. The editors read mail on Mondays and Wednesdays, and a note pointing at a stale figure or a changed feature is usually what triggers the next round of retesting.