Account Scoring for B2B: A Simple Model You Can Build

What is account scoring, and how is it different from lead scoring?

Lead scoring rates a person. Account scoring rates a company. In B2B, several people usually weigh in on a purchase, so one keen contact at a poor-fit company is worth less than quiet interest from three people at a perfect-fit one. Account scoring captures that.

The output is simple: a score per account and a tier that tells your team what to do next. The top tier gets personal outreach from a salesperson. The middle tier gets a researched sequence. The bottom tier gets nothing for now, or light marketing. The value is not in the number. It is in the decision the number makes easy.

Fit, intent and engagement answer three different questions

Most scoring models go wrong by mixing these up. Keep them separate, at least while you build.

• Fit asks: should this company ever buy from us? Industry, size, region, business model, the tools they use, and whether the right buyer role exists. Fit changes slowly.
• Intent asks: is there a reason to buy now? Hiring for a role your product supports, a new leader in the buyer seat, new funding, expansion into a new market, or a public statement about the problem you solve. Intent fades over weeks.
• Engagement asks: have they interacted with us? Replies to outreach, visits to key pages, form fills, webinar attendance, or a meeting taken. Engagement is the strongest signal and the easiest to misread, because a student or a competitor can read your pricing page too.

A simple weighted model you can build this week

Here is a starting model on a small points scale. The points below are illustrative choices for a typical B2B company, not benchmarks. Nobody can tell you the right weights for your market from the outside. Treat these as a first draft, then change them after you test the model against your own closed deals.

Each signal is yes or no. If yes, the account gets the points. Fit signals add up to a maximum of 10. Intent and engagement add on top.

How scores should route work

A score only helps if it changes what someone does on Monday. Set tiers with simple rules. Again, the cut-offs below are illustrative. Pick your own after looking at where past wins would have landed.

• Tier A: fit of 8 or more plus any intent or engagement signal. A salesperson researches the account and reaches out personally, across email and LinkedIn, within days.
• Tier B: fit of 6 or more with no intent yet. The account goes into a researched email sequence. AI drafts from a note on each company, and a person approves every message.
• Tier C: fit below 6. No outbound for now. Keep the account on file and rescore it when new signals appear.
• Any tier: a reply or a meeting request jumps the account straight to a salesperson, whatever the score says.

A worked example: two accounts, two different next steps

Take two companies from the same list, using the illustrative points above. Company one is in your best industry, in your size range, has a reachable head of sales, and uses a CRM you integrate with. That is a fit score of 10. It posted a job for two new SDRs last week, which adds 3 intent points. It lands in Tier A, and a salesperson opens with a note about the hiring plan.

Company two is in your best industry and has the buyer role, but it is outside your size range and uses no tool you connect to. Its fit score is 5. Someone from the company read your pricing page, which adds 2 engagement points. It still lands in Tier C, because fit comes first. One page visit does not make a poor-fit account worth a salesperson's afternoon. If that visitor replies to anything, the reply rule moves the account up.

How do you pick the weights without guessing?

Use your own history. Pull your last twenty or so closed-won deals and the same number of closed-lost or stalled ones. For each account, note which signals were true at the time you first engaged. Then compare the two groups.

Signals that show up far more in wins deserve more points. Signals that show up equally in both are noise, even if they feel important. If you have few closed deals, use the accounts that took meetings as a stand-in, and revisit the model once you have more wins.

Then run the model backwards. Score last year's pipeline and check whether your wins would have landed in Tier A. If many landed in Tier C, your fit criteria are wrong, not your salespeople.

Keep it simple first

The most common failure is a model with dozens of signals that nobody trusts or understands. Start with six to eight signals you can actually collect for most accounts. A spreadsheet is fine for the first version. Move it into your CRM once the tiers hold up.

Watch three other traps. First, stale intent: a hiring post from months ago is not a reason to buy now, so let intent points expire. Second, bad data: a score built on wrong company sizes or dead contacts routes work to the wrong place, so enrich and verify before you score. Third, no feedback loop: ask your salespeople every month which Tier A accounts were a waste of time, and adjust.

Where Trexinet fits

Trexinet builds scoring into the outbound programs we run for B2B companies. AI enriches each account, researches the signals, and drafts a note on why it scored where it did. A person reviews the tiers before any outreach starts, and a human approves every send. Replies are triaged and synced to your CRM, so engagement feeds back into the score.

If you want a first scoring model and target list for your market, we write a free 30-day pipeline plan within one business day.