Account scoring models

Build account scoring around a decision your team can act on. Keep account quality separate from timing. A single account score can mix whether you should sell to an account with whether you can reach it this year.¹ A high-fit account with weak access needs a different plan from a reachable account with weak fit. Give each score one job and combine them only when the action calls for it. Use the score to rank accounts, not to make the decision for you.

Decide what the score must decide

Start with the decision the score will support. If it cannot tell you what to do next, it is only a sorting exercise.

For account prioritization, use the size of the potential prize and the likelihood of closing as separate inputs.² Define fit as the account's match against your firmographic, technographic, and operational criteria.³

Write the decision in plain language before choosing fields:

• Should this account enter the target list?
• Should it receive focused coverage?
• Should it move into active outreach now?
• Should it stay in a lower-touch queue?

Keep the decision narrow enough for one score band to lead to one clear action. If you need to answer several of these questions, keep separate outputs for fit, value, and readiness.

Choose the inputs

Choose fields that explain the decision and that your team can check consistently. Each extra field makes the model harder to maintain, so give every field a job.

Account assessment uses a collection of different data points.⁴ A common model can include firmographic fit, technographic fit, and intent or activity.⁵ Fit scoring can use company properties such as company size and annual revenue.⁶ Activity scoring for companies can use website visits, newsletter subscriptions, CTA clicks, and marketing email opens.⁷

For account selection, include current revenue, account-level growth opportunity, cost to serve, and geographic location when those factors affect the decision.⁸ Technographic data can also help determine equitable account books.⁹

Keep account facts separate from signals about current behavior. A company can fit your target profile while showing little current activity. It may deserve a place on the list without earning immediate outreach priority.

Set weights and scoring rules

Weights tell the model what to do when signals point in different directions. Set them after choosing the decision and fields, then write down how each value earns its score.

Assign weights to the account-assessment data points.¹⁰ After selecting the fields, determine how to score the values in those fields and the combinations of those values.¹¹ Decide whether a strong signal can compensate for a weak one or whether certain combinations should cap the result.

Use a consistent rating guide for contextual factors so different people reach a common, defensible score.¹² Describe what each score means with observable evidence. Use weighted firmographic and technographic fit, plus intent or activity, to prioritize and tier accounts, and document thresholds and examples for each band.¹³

Keep the component scores visible even if you publish a combined score. A single total is easier to sort, while the components show why an account landed there and which input should change before the next review.

Turn the score into a queue

A score earns its place when it changes the order of work. Connect each band to a coverage action, an owner, and a review condition.

Sales representatives use scores to prioritize outreach and follow-ups.¹⁴ In enterprise selling, account fit and stakeholder mapping can matter more than volume, and one well-worked account can outweigh twenty thin ones.¹⁵ Let the score protect time for accounts that deserve deeper work.

Run two capacity checks before setting the size of the target list.¹⁶ The sales model asks how many accounts you need to hit the target.¹⁷ The capacity model asks how many accounts one rep can handle.¹⁸ Compare the results before assigning coverage. A list can support the target and still exceed the team's ability to work it properly.

Account count can assign the same weight to a 40-employee prospect and a 4,000-employee prospect even though the accounts differ materially.¹⁹ Use expected value, fit, and the amount of work required to shape coverage. Let lower-priority accounts wait when the list exceeds workable capacity.

Review the model

Review the score when outcomes or account conditions expose a weak assumption. Improve the rule, the inputs, or the action attached to the result.

Inaccurate technology and firmographic data can prevent a company from identifying real ICP-fit accounts.²⁰ Check whether the highest-scoring accounts still match the kind of business you can serve and whether the signals still predict useful work.

Keep the prospect score separate from a score for existing-customer engagement. Sales and marketing teams tracking prospects may need a different score from a support team tracking existing customers.²¹

When the model produces too many similar scores, inspect the weights and combinations before adding more fields. When it produces a queue no one can work, revisit the capacity check and the thresholds.

What not to do

Use these checks when the model starts producing familiar-looking lists or vague priorities.

• Do not let sales intuition, existing relationships, or a minimally updated prior-year list drive account selection.²²
• Do not use industry averages to decide how many accounts each person should target.²³
• Do not run outreach without a systematic scoring method, because SDRs waste time without one.²⁴
• Do not use the same score for prospect prioritization and existing-customer engagement when those teams track different behaviors.²¹