Methodology

Revenue estimation: the model, its signals, and its confidence

Twenty-five infrastructure signals, additively scored into five revenue brackets, with a published confidence value and a full evidence trail for every record.

The brackets

Bracket Revenue Employees
Micro <$1M 1–10
Small $1M–$10M 11–50
Mid-market $10M–$100M 51–500
Enterprise $100M–$1B 501–5000
Large enterprise $1B+ 5001+

How it works

No financial data is used — none is publicly available for most companies. Instead the model reads what a company's infrastructure choices reveal about its budget. Each signal contributes points to one or more brackets; points are summed, normalised, and the leading bracket wins if it clears the confidence threshold.

The strongest signals

Signal Weight What it reveals
Tranco rank strongest The single best predictor. Top-1k contributes 20 points toward large enterprise.
Registrar tier high MarkMonitor and CSC serve enterprise brand-protection clients; GoDaddy and Namecheap serve everyone else.
Email provider + SPF high Proofpoint and Mimecast are enterprise procurement. A long SPF include chain means many paid SaaS tools.
Marketing automation high Marketo or Salesforce in the SPF record implies a funded marketing team; Mailchimp implies a small one.
Subdomain count medium Hundreds of subdomains means many internal systems, which means many people.
CMS and CDN tier medium Sitecore or Adobe plus Akamai is a six-figure stack. Wix is not.
Enterprise pages medium /investors, /security, /partners and /careers appear when a company has those functions.
Domain age + DMARC medium A 1998 domain with an enforcing DMARC policy is a mature, governed organisation.

Confidence — and when we publish nothing

Confidence is the normalised margin between the winning bracket and the rest. Two guards apply: the estimate must reach 0.40 confidence and be supported by at least three independent signals. Below either threshold no estimate is published.

A blank is more honest than a guess

Roughly a third of records carry no revenue estimate. That is intentional. A wrong bracket sends a rep into a call with the wrong pitch and the wrong price — far more expensive than an empty field they can see and route around.

The evidence trail

Every record carries the list of signals that fired and the bracket each favoured, for example tranco:top_100k | registrar:markmonitor | spf:marketo | pages:/investors. You can audit any estimate, and you can filter on the evidence itself when a particular signal matters more to you than the bracket.

Stability across re-crawls

Rank data wobbles daily around the bracket boundaries. Rather than let that flip a company between brackets, the estimate only moves when the new reading jumps two brackets or is meaningfully more confident. You get a stable field you can build a workflow on.

Why this decides your pricing and your pitch

Territory and quota design

Bracket plus country plus industry is enough to build territories from a raw domain list — no manual research pass. Employee ranges are derived per business model using sector-appropriate revenue-per-employee, so a SaaS and an ecommerce brand at the same revenue get realistically different headcounts.

Knowing when to walk away

The most valuable use is negative. If your product needs a $10M+ customer, filtering out micro and small removes most of a list before a human ever reads it — and the confidence value tells you which exclusions to trust.

Other scoring methodology

SEO Score Reputation Signals Revenue Estimation Shopify Store Score
Start free — no credit card required