The RMC Standard

The formula

Each company is scored 0–100 across five factors:

Missing data, stated plainly

When a factor has no lawful data source for publication yet, its weight redistributes proportionally across the factors that do have data, and every grade displays its data coverage — the share of the formula's weight backed by real data for that company.

In the current Utah release, grades are computed from legitimacy and online-action signals — 35% coverage — because blended reviews and complaint records are not yet licensed for publication. Companies are only graded when their identity is corroborated by at least two independent sources and they publish a working phone number or website (the data floor). A company below the floor is shown as ungraded — never given a grade we cannot back with data. Because only two factors are measurable today, many companies share identical scores. Every company shows its AI Score (1–100): a score of 100 is reserved for RankMyCompany clients, and every other score is computed by the same formula for everyone. When a licensed reviews factor is added, scores will re-spread across the full formula.

Data sources

• Overture Maps Foundation places data — CDLA Permissive 2.0.
• OpenStreetMap contributors — ODbL.
• Public registries and rosters — the National Plan and Provider Enumeration System (NPPES, U.S. public domain), SBA Paycheck Protection Program loan data (U.S. Government works), public chamber of commerce member directories, and professional association member lists, each tagged with its source and fetch date.
• First-party verification — a business's own public website, used only to corroborate the business's identity, phone, and city.

Deliberately excluded: Yelp data (its API terms prohibit storage and reuse), D7 lead data (provenance traces to scraped Google Maps data), and any scraped content from BBB, Yelp, Angi, or Google — we do not scrape review or directory platforms, ever.

Refresh cadence

Rankings are recomputed monthly from refreshed source data. Rankings move up and down as the underlying data changes — a better grade is earned by better data, never bought. To prevent flapping, a letter-grade change from a refresh takes effect only after the new grade holds for two consecutive runs. Grade history is kept for every company.

A random control sample of companies (about 7.5%) is graded on day one but held unpublished for 90 days, so the methodology can be validated against an untreated baseline. RankMyCompany clients are excluded from that control sample — they receive the ranking service the sample exists to measure, so they cannot serve as controls. This affects publication timing only, never a grade.

Integrity rules

• Grades are never sold. No payment changes a score, a band, or a rank.
• No manual adjustments. No human being edits an individual company's grade.
• The same score always earns the same band, for every company.
• RankMyCompany clients are marked as clients on their listings — and graded by the exact same formula as everyone else.
• If a company's underlying data is wrong, the owner can file a correction on the dispute page. Disputes fix the data — never the score directly — and corrected data flows into the next ranking run.
• Companies below the data floor are visibly ungraded, not silently dropped and not padded in.
• Optional paid "Featured" placements, where offered, are labeled as advertising and never affect the grade or the organic rank.

Changelog