Right Suite logoRight Suite
AudienceNovember 22, 2026

How to Score Target Companies Against Your ICP (Match-Score Sheet With Evidence)

Score each company on 5 to 7 weighted ICP criteria, and write the evidence and source link next to every score. No evidence means a zero, not a guess.


To score companies against your ICP, turn it into 5 to 7 weighted criteria, score each company 0 to 3 per criterion, and put the evidence and source link next to every score. No evidence counts as zero. Sort by total and work the top tier first. RightAudience helps you set the weights.

Check your ICP criteria in RightAudience before you score 200 companies against the wrong ones.

Who this is for

You have an ICP and a list of companies. Maybe 50, maybe 2,000. You need to decide who to contact first, and you want the ranking to hold up when a co-founder, an investor or a future SDR asks "why this company?"

This is account-level scoring. If you're still choosing between whole segments, use the ICP scoring matrix first, then come back here.

Why does evidence per row matter?

Most lead scores are a number with no trail. Three weeks later nobody remembers why Acme got an 8. Worse, the founder fills gaps with optimism, so every company looks like a fit.

Writing the evidence and the link next to each score fixes both problems. You can audit any row in ten seconds. You can see which criteria you're guessing on. And when a company turns out to be a bad fit, you can find the row that misled you and fix the criterion.

How to build the match-score sheet

1. Turn your ICP into checkable criteria

Each criterion must be something you can verify from public information. "Values efficiency" fails. "Has 3 or more open roles in support" passes.

Good criteria for B2B SaaS:

CriterionHow you check itTypical source
Headcount in rangeEmployee countLinkedIn company page
Team you sell to existsJob titles at the companyLinkedIn people search, team page
Trigger eventHiring, funding, launch, new leaderCareers page, press release, news
Tech fitUses a tool you integrate withTheir docs, job posts, BuiltWith or Wappalyzer
Business model fitSells B2B, has a self-serve plan, etc.Their pricing page
Pain signalComplaints, reviews, job post wordingG2 reviews, job descriptions
DisqualifierUses a competitor on a long contract, too bigCase studies, partner pages

2. Weight the criteria

Give each criterion a weight from 1 to 3. The trigger event and the pain signal usually deserve 3, because they decide timing. Headcount usually deserves 1, because it's a filter, not a reason to buy.

Mark one or two criteria as must-have. Fail a must-have and the company is out, whatever the total.

3. Score 0 to 3, with a strict rule for each number

ScoreMeaning
0No evidence found, or evidence against
1Indirect evidence (e.g. a related role, an old post)
2Direct evidence, older than 6 months
3Direct evidence, recent

The zero rule is the whole point. If you can't find it, you don't know it.

4. Fill one row per criterion, per company

This is the sheet. One block per company:

CriterionWeightScoreEvidenceSource
Headcount 50 to 30013140 employeesLinkedIn company page URL
Support team exists239 people titled "Support"LinkedIn search URL
Trigger: hiring support332 open support roles posted this monthCareers page URL
Uses Zendesk22Help center on Zendesk domainHelp center URL
Pain: slow replies32Reviews from last year mention slow support repliesG2 review URL
Weighted total28 / 33

Paste the actual URL in the source column. "LinkedIn" isn't a source; a link is.

5. Tier the list

  • Tier A (80% of max or more, all must-haves met): contact this week, with a message built on the trigger.
  • Tier B (55 to 79%): contact after Tier A, or wait for a trigger.
  • Tier C (below 55%): park. Re-check quarterly.

6. Feed results back into the weights

After 30 conversations, look at who said yes. If every yes had a recent trigger and headcount didn't matter, raise the trigger weight and drop headcount. The sheet gets more accurate every month you use it.

Symptom, check, fix

SymptomCheckFix
Every company scores highHow many scores have no source link?Apply the zero rule strictly
Tier A doesn't replyIs the trigger recent?Raise the weight on recency, re-score
Scoring takes 30 min per companyToo many criteria?Cut to 5; drop ones that never change the tier
Good customers came from Tier CWhich criterion marked them down?Lower its weight or remove it

Worked example

For example, say you sell a support-analytics tool and you score two companies.

Company one has 140 staff, a 9-person support team, two support roles posted this month and a Zendesk help center. It scores 28 out of 33. Tier A. Your first line writes itself: you're hiring two support people, here's how teams that size keep reply times down while they ramp.

Company two has 400 staff and a large support team, but you found no trigger, no tool evidence and no pain signal. Those rows score zero. It lands at 12 out of 33, Tier C, even though it "looks like" your ICP on size. Without the evidence column, you'd probably have put it first.

How to test it with RightAudience

The sheet ranks companies against your criteria. It can't tell you whether the criteria are right. That's where RightAudience comes in.

  1. Turn your two or three strongest criteria into segment descriptions. For example: "Support lead, B2B SaaS, 50 to 300 staff, currently hiring support" versus "Support lead, B2B SaaS, 300 to 1,000 staff, not hiring".
  2. Describe your offer and price, and run those segments.
  3. Read the ranked segment table, with purchase intent, objections and the focus recommendation per segment.

If the "currently hiring" segment ranks well above the other, your trigger deserves its weight of 3. If size matters more than you thought, adjust the weights before you score the full list. RightAudience ranks segments, not individual companies, so use it to set the rules and use the sheet to apply them.

Test your ICP criteria in RightAudience

FAQ

How do I rank companies against an ICP?

List 5 to 7 criteria you can verify from public sources, weight them by how much they predict a purchase, score each company 0 to 3 per criterion with the source linked, and sort by weighted total. Remove anyone who fails a must-have.

What's the difference between ICP fit and intent?

Fit is whether the company matches who you sell to: size, team, tools. Intent is whether they need it now: a trigger, a complaint, a new hire. Score both, and weight intent higher, because fit without timing produces polite "not right now" replies.

Can AI tools score companies against my ICP automatically?

They can gather evidence fast, from careers pages, pricing pages and reviews. Keep the evidence-and-link column anyway, and check a sample by hand. An unsourced AI score has the same problem as an unsourced human one.

How often should I re-score my list?

Triggers go stale, so re-check Tier B and C monthly or quarterly for new hiring, funding or leadership changes. Re-weight the criteria after every 30 conversations using who actually bought.

Next step

Pick your top 20 companies and fill in the sheet with sources for each row. Before you scale to the full list, check your criteria in RightAudience. Then write the first message off the trigger; how to personalize cold outreach at scale shows how to use that evidence in the opening line.