- Run a structured survey for top AEs and CSMs

- Ask who buys fast, who churns, where deals stall, who champions

- Cluster answers into patterns, not isolated stories

- Use LLMs (Claude/GPT) to summarize traits of best and worst-fit customers

- Output: a list of qualitative ICP signals (good + bad)

‍

- Export 12–24 months of Closed-Won + Closed-Lost deals

- Enrich each account in Clay with firmographics (industry, size, funding, geo)

- Add technographics (stack, CRM, data tools, competitors, integrations)

- Optionally enrich growth + intent (hiring, traffic, jobs, reviews)

- Output: a rich historical dataset that explains why certain deals win

‍

- Define Tier 1 / Tier 2 / Tier 3 based on fit and value

- Choose firmographic criteria (sub-industry, headcount, region, funding, growth)

- Add technographic criteria (complementary tools, maturity, stack alignment)

- Include account fit signals and negative indicators

- Turn this into a clear, shareable tiering rubric for the whole GTM org (example here)

- Implement the scoring logic as rules inside Clay

- Run it across historical deals to assign Tier 1/2/3 to every past account

- Check: do wins cluster in Tier 1 and losses in Tier 3?

- Refine signals and thresholds until tiers line up with real ACV, win rates, and cycle length

- Outcome: a model that predicts correlates with close rate and contract size, not based on assumptions

- Push tiers into CRM and make them visible on every account

- Acquisition: focus paid + ABM on Tier 1

- Manual prospecting: SDRs start with Tier 1

- Automated outbound: Tier 2 and 3

- Marketing: build messaging and positioning around Tier 1 accounts

‍