How to Score and Route B2B Signals Without Treating Every Event as Intent

Why do sales teams stop trusting scores?

Usually because the score cannot be explained. A record marked ‘hot’ turns out to be a student, a competitor or a job seeker, and after a few of those, people stop believing the label. The usual cause is an opaque model: points assigned by rules nobody can see, fit and timing mixed into one number, and no feedback from what happened.

When a seller cannot answer ‘why is this account here?’, they fall back on their own judgment, which is what the score was meant to support.

Key pointA score nobody can explain is a score nobody uses.

How should you separate fit from timing?

Fit asks whether an account matches your ideal customer profile. Timing asks whether there is current evidence that it may deserve attention. They are different questions, and blending them hides the answer to both.

Key pointShow fit and timing as two dimensions, so the next action follows from both.

What should a transparent score show?

Not a single number, but the reasons behind it, in words a seller can check:

• Which fit criteria the account meets.
• Which signals were observed, and when.
• What context was added: the people, the CRM history, the situation.
• What the score does not know.

Reasons do two jobs. They let people trust a good score because the logic is visible, and they make a bad score easy to challenge, which is how the model improves.

Key pointReasons make a score checkable, and a checkable score can improve.

How do you avoid treating every event as intent?

Score signals as evidence of context, not as proof of purchase. A few working rules:

• Weigh combinations over single events: one visit is weak context; several related signals at the same account are stronger.
• Let signals decay: evidence from months ago says less about now.
• Keep a ‘not enough evidence yet’ outcome, so the system can decide to wait.
• Match the action to the evidence: a light signal earns a relevant, low-pressure touch, not an urgent sales call.

Key pointA score should say how much context there is, not how sure you are of a sale.

What should you define before automating routing?

A good score is wasted if routing sends it to the wrong place. Before building workflows, define the business rules:

Lead or account routing

Whether the unit of work is a person or the account, and what happens when several people from one account appear.

Ownership

Who owns the account by territory, segment or existing relationship, and who covers when they are unavailable.

Next action

What the owner is expected to do, and what happens if nothing is done.

States

Which CRM states routing reads and writes, including the difference between not ready and wrong fit.

Automation should implement approved logic, not hide an unclear process. That is the core of CRM / RevOps Orchestration.

Key pointDefine who acts and what happens next, then automate it.

How does a scoring model improve over time?

By recording what happened after each action and feeding it back. In the D.A.R.K Loop™ that is Kaizen: learn from outcomes and improve scoring, routing, messaging and prioritization. Without it, a scoring model keeps the assumptions of the day it was built.

Key pointScores improve only when outcomes are allowed to correct them.

Frequently asked questions