Define the record and decision being scored
Score the buying account for ICP fit. Score dated activity for engagement. Keep the contact’s role and request available to the seller. Once a buying process exists, use opportunity evidence and the DealPack win proxy to review that deal.
Keep the measures visible separately. A fit score of 80 is not an 80% win probability, and a large potential contract value is not a bookings forecast.
| Measure | Evidence | Decision |
|---|---|---|
| Account fit | Firmographics, technographics, use case, and ability to adopt | Eligibility and territory composition. |
| Lead or account engagement | Dated buyer actions, relevant product activity, and conversations | Response and weekly work priority. |
| Potential contract value | Pricing units, likely scope, and comparable won contracts | Economic potential and book balance. |
| Opportunity win proxy | Buyer replies, meetings, people, deal size, and close-date movement | Deal review and the evidence still needed. |
Build fit criteria from the customer problem
Start with customers the business can win, retain, and serve. Choose the firmographic and technographic facts that explain the need and implementation requirements. Company size or a technology tag needs a business reason to earn points.
- Write observable scoring anchors, an evidence source, refresh rule, and weight for each criterion.
- Use a consistent scale and weights totaling 100%. Keep hard exclusions separate from the weighted fit total.
- Treat missing required evidence as pending. Do not score an unknown as a confirmed mismatch or silently redistribute its weight.
| Input family | What to establish |
|---|---|
| Firmographic | Which size, industry, region, business model, or growth pattern changes the need or ability to buy? |
| Technographic | Which existing tools create a compatible workflow, a replacement need, or a real integration constraint? |
| Problem and operating motion | Which recurring work, team structure, or use case makes the product valuable? |
| Adoption requirements | Which implementation, data, staffing, or operating conditions are necessary? |
Score dated engagement without counting the same event twice
A buyer request or substantive reply supports a different action from a page view. Define the signal, its observation date, the period it remains relevant, and the action it supports. Keep passive activity from overwhelming more informative evidence.
Deduplicate repeated deliveries of the same event. Cap repeated low-information events and define how contact activity rolls up to the buying account. Preserve who acted and what they requested so the seller can respond in context.
- A demo request supports a timely response; the buying process still needs qualification.
- Several active contacts may indicate broader evaluation, but confirm their buying roles.
- Expire or decay signals according to their useful window. A buyer-agreed return date remains visible alongside the score.
Set the trigger for rep attention
Prioritize accounts where ICP fit, recent engagement, and the buying cycle line up. Screen the account and signal before routing it. The trigger should explain why this account merits seller attention now and what the rep should follow up on.
A direct buyer request needs a timely response. The rep uses that conversation to confirm the problem, motivation, decision timing, and next step. Do not make the buyer wait while the team completes a second qualification form.
- For each trigger, define the qualifying account group, event, evidence window, expiry, receiving owner, and response expectation.
- Route signals on an active opportunity to its current owner with the deal context. Create a new opportunity only when a distinct buying process is established.
- Keep high-fit accounts outside a current buying window in planned outbound or nurture. Record a confirmed return date when one exists.
- Review whether triggered accounts become accepted opportunities and progress through the expected buying window, alongside missed or late handoffs.
| Screen | Evidence and treatment |
|---|---|
| ICP fit | Confirm the use case and required company characteristics. Research missing facts instead of treating them as a failed fit. |
| Shopping signal | Identify the actual request or evaluation activity, who generated it, and when it happened. Passive activity can support research without establishing a buying process. |
| Buying window | Record the expected evaluation or purchase timing, its source, and whether the buyer confirmed it. An inferred renewal or planning cycle remains a hypothesis. |
| Motivation | Give the rep the likely business reason for acting. In discovery, confirm the outcome the buyer wants and why the decision matters now. |
Route the lead through the account record
Match a new lead or contact to its buying account before assigning sales work. Keep one account owner and a clear next action. If matching is uncertain, resolve the relationship rather than creating another company or parallel outreach thread.
| Situation | Treatment |
|---|---|
| Known account with an owner | Send the request and activity to that owner. |
| Known account with an active opportunity | Connect the contact and evidence to the current buying work; avoid a duplicate opportunity. |
| Unowned eligible account | Assign under segment and capacity rules, then verify the saved owner. |
| High activity with weak or uncertain fit | Answer the request and establish the use case before treating it as qualified pipeline. |
| Strong fit with no current signal | Use planned outbound, research, or nurture according to the motion. |
Estimate contract value from scope and pricing
Estimate the contract the account could plausibly buy: priced units, likely users or usage, required product scope, and comparable won contracts. Record the amount basis, range, assumptions, source date, and number of comparable contracts.
When a predictive value model is justified, use fields available at the estimate date and test it on later contracts. Until there is enough evidence, a transparent pricing-based range is more useful than an unexplained precise number.
- Separate annual recurring value from one-time fees and total multiyear contract value.
- Do not count the full account potential and its open opportunity as two revenue contributions.
- Use potential value to compare books and prioritize research; forecast only the scoped opportunity with supported buying timing.
Connect account priority to opportunity review
A strong ICP match can still have a stalled deal. High engagement can reflect evaluation work, administrative traffic, or uncertainty. Review the actual buying process before interpreting a changed score.
DealPack starts from a historical win rate and adjusts it using observed behavior. Keep that win proxy, the suspected blocker, and the seller’s next action in separate fields. A score change prompts inspection; buyer evidence determines stage, timing, and disposition.
Validate each model against the outcome it estimates
Check known fits, poor fits, incomplete records, and accounts where fit and engagement disagree. Then save a scoring-date cohort and inspect later outcomes by segment and version. Retain the attention each account received, since routing and seller effort affect observed results.
- Keep source evidence, model version, observation date, and missing-data status in the existing record.
- Inspect routing impact before changing thresholds; preserve active buyer relationships under ROE.
- Use the recurring exceptions to change a specific criterion, source, or workflow.
| Model | Review |
|---|---|
| Fit | Accepted opportunities, wins, retention, and repeated seller objections by band. |
| Engagement | Whether the signal led to relevant conversations or evaluation within its stated window. |
| Contract value | Estimated range versus actual won value for comparable scope and pricing. |
| Opportunity win proxy | Later win frequency by saved score band, with sample count and follow-up window. |