build-and-fill-a-tam

Use this when the team needs to size a market and turn that estimate into an executable account-coverage program. Produce labeled assumptions, real account counts where possible, and a burn-up plan that never confuses loaded records with valid TAM coverage.

Freeze the ICP definition

Write the account inclusion and exclusion rules before counting. Specify industry, company size, geography, business model, technology or operating conditions, and disqualifiers. Separate required filters from preferences and mark which filters can actually be measured by each source.

Define the unit of analysis. Count companies, not people, when estimating account TAM. Decide how subsidiaries, franchises, agencies, parent companies, and multi-brand groups are treated. Preserve one canonical domain per operating account and document exceptions.

Estimate the account universe

Prefer a bottom-up count from a source capable of returning the matching companies. Record the source, query date, applied filters, omitted filters, caps, sampling, and estimated duplication. If the source exposes only a count, label the result as a count estimate rather than a verified account list.

Cross-check with a second method: industry registry, known-account sample, CRM classification, or another data source. Investigate large differences instead of averaging incompatible populations. A provider cap is a floor, not the size of the market.

Use references/tam-model.md for the calculation and uncertainty bands.

Add revenue and contact targets

Use a real annual contract value: the median or trimmed mean of comparable closed-won deals, annualized from the recorded contract period. Do not substitute a price-page maximum or an unlabeled band.