Score your ICP on stack density (L4)

The steps

1. 01

   Write your density index as a weighted formula

   Pick 8 to 12 technologies whose presence indicates a buyer ready for what you sell, and assign each a weight. Group them into three or four categories so the score means something: sales execution (a sales engagement platform, a conversation intelligence tool, a forecasting tool), data maturity (a cloud warehouse, a transformation layer, a BI platform), AI readiness (an LLM assistant deployed company-wide, an orchestration framework, a vector database), and scale signals (an identity provider, an HRIS). Write the formula down so it can be audited and argued with. A score nobody can inspect gets ignored the first time it contradicts a rep's favorite account. • Owner: RevOps • Tool options: a document, then your technographic provider's filter • Pitfall: shipping a formula nobody can inspect or argue with • Definition of done: a written weighted formula exists and sales leadership has reviewed and challenged it

2. 02

   Validate against 200 closed deals before it touches a territory

   Score the last 200 closed opportunities, won and lost, and compare average score for each group. Also check win rate by score decile, because a formula can separate the averages while being useless in the middle of the range. If won deals do not score meaningfully higher, the formula is wrong. Adjust weights and re-run. Do not proceed on a formula you would not defend in front of the CRO, because that is exactly where it ends up. • Owner: RevOps analyst • Tool options: your CRM export plus the technographic provider • Pitfall: proceeding on a formula that does not separate won from lost deals • Definition of done: the score separates won from lost by a margin you would defend, and win rate rises across deciles

3. 03

   Score the live database and set exactly three tiers

   Tool: Clay

   Score every account, then cut into three tiers on two axes, density score and a headcount band that historically converts for you. Tier 1: high density and in a converting size band. Tier 2: one of the two. Tier 3: neither. Three tiers, no more. Five tiers means nobody remembers what tier 4 meant. Push tier and score onto the account record where reps can see them, with the contributing technologies listed so the score is explainable on a call. • Owner: RevOps • Tool options: Clay, Clearbit, HG Insights or ZoomInfo pushing into your CRM • Pitfall: building five or more tiers instead of exactly three • Definition of done: every account carries a tier and score, the contributing signals are visible, and the tier distribution has been reported to leadership

4. 04

   Reallocate territories and quota to match the tiers

   Tool: Salesforce Maps

   Move tier 1 density toward your strongest reps, and set quota against realistic conversion per tier rather than a flat number per head. A rep holding mostly tier 3 accounts with the same quota as a rep holding tier 1 will miss, and you will conclude the rep is the problem. This is where scoring projects quietly die. Everything before this step is analysis. This is the step where the analysis costs someone something, which is why it needs the CRO's name on it and not RevOps' alone. • Owner: CRO plus RevOps • Tool options: Salesforce Maps, Anaplan, Fullcast, or your territory planning spreadsheet • Pitfall: leaving territories and quota untouched, letting analysis stop at the score • Definition of done: territories are redrawn and quotas reflect tier mix

5. 05

   Write two messaging variants, differentiated by what they already run

   Tool: Outreach

   Same product, two different first conversations. Tier 1 already has AI in the building, so lead with orchestration, governance and control, their pain is sprawl, cost and no visibility (playbook 9 and playbook 12 language). Tier 3 has nothing yet, so lead with the first workflow and a 30-day proof, their pain is not knowing where to start (playbook 1 language). One hard rule: never open by claiming they lack a technology. Technographic providers under-detect badly and unevenly, and being wrong about their stack in the first line ends the conversation in the room. • Owner: Product marketing • Tool options: your messaging framework, loaded into Outreach, Salesloft, HubSpot or Apollo • Pitfall: opening by claiming a prospect lacks a technology when detection may simply be wrong • Definition of done: two variants exist, are loaded into the sequencer, and are mapped to tier in the CRM

6. 06

   Recalibrate quarterly against actual results

   Tool: Looker

   Each quarter compare predicted conversion per tier against actual, then adjust the weights and document what changed and why. Also re-check provider coverage, because vendors get acquired, detection methods change, and a technology that scored high last year may simply stop being detected. A model nobody refreshes becomes folklore inside two quarters, and folklore is harder to remove than a bad model. • Owner: RevOps analyst • Tool options: closed-deal cohorts by tier, charted in Looker, Tableau or Power BI • Pitfall: leaving the model unrefreshed until it becomes folklore • Definition of done: two quarterly recalibrations are complete with weight changes documented

7. 07

   Instrument the result

   Tool: Looker

   Instrument: win rate and average deal size by tier. If tier 1 does not beat tier 3 on both, the index needs work rather than more accounts. • Where it breaks: absence bias. A provider that under-detects a vendor makes a good account look empty, so treat every low score as a hypothesis to test on the call rather than a fact to assert in an email. Second failure: territories never get redrawn, so the score becomes a field nobody looks at. Third: five tiers, or a formula nobody outside RevOps can explain, and reps override it on instinct within a month. • Visual guidance: not specified beyond the tier and score fields being explainable on a call; present the contributing signals alongside win rate and deal size by tier.

Tools in this playbook

• Clay
• Salesforce Maps
• Outreach
• Looker

Next playbooks

• Voice-of-Customer → Product & Enablement Loop (L2)
• Zapier Agents for sales ops chores (L2)
• Agentic Nightly CRM Hygiene + Ownership Sync (L3)
• AI-driven price-quote acceleration with CPQ (L3)