How to score buying windows without a huge TAM list

You do not need a million-row TAM to find buyers.

You need a small cohort and a score that tracks real movement. Ehrenberg-Bass and the LinkedIn B2B Institute put in-market near 5%. Building a giant list of the other 95% does not create demand. It creates work and spam risk.

A million-row TAM with no signal score is a vanity list. About 95% are out of market anyway. Score the few who move.

What is a buying window score?

A buying window is a stretch when an account is actually in motion. See what is a buying window.

A score is a simple way to rank which accounts deserve first-touch this week. Not a black-box "AI lead score" that nobody can explain. A few weighted signals you can defend in a standup.

Fit still matters. Fit alone is not intent. Read firmographic fit vs buyer intent. Score windows on behavior after fit clears.

Which signals belong on the scorecard?

Start with three you can actually instrument.

1. Website de-anon. Named accounts on pricing, docs, or comparison pages. Tools like RB2B or Vector. See website de-anonymization.
2. G2 category or competitor traffic. Someone researching the shelf. See G2 category pages.
3. Competitor LinkedIn engagement. Comments and follows that show attention, not just a job change. See competitor LinkedIn engagement.

Optional fourth: product stuck states in PostHog, or pathing in Factors.ai, if you have product traffic.

Job change can add a point. It is not the motion by itself.

How do you score without a huge TAM?

Build a seed cohort, not a universe

Pick 50 to 200 fit accounts. Enough to learn. Not enough to burn a domain.

Weight recent, stacked signals higher

One de-anon hit last quarter is weak. De-anon plus G2 plus competitor engagement in the same two weeks is a window. Stacking beats list size.

Decay old signals

A visit from six months ago is history. Forrester's ~84-day cycle when a deal is needed is a reminder that windows have length. They also expire. Decay beats forever scores.

Cap the queue

Only the top N each week get a human first-touch. Everything else stays in out-of-market nurture or silence. Salesforce's ~27% quota reality does not improve because your TAM spreadsheet got longer.

What does a simple scoring model look like?

Example weights you can change:

• Fit hard-fail: score 0, never send
• De-anon on high-intent page this week: +3
• G2 category or competitor page: +2
• Competitor LinkedIn engagement this week: +2
• Second distinct signal inside 14 days: +2
• Job change alone: +1

Send threshold: 5+. Review weekly. Kill any rule that floods the queue with junk.

This isn't magic. This is a checklist with numbers.

What should a team of one do this week?

1. Export 100 fit accounts. Stop at 100.
2. Wire one sensor (de-anon or G2 or competitor engagement).
3. Score by hand for two weeks. Then automate the boring parts.
4. Measure meetings from scored sends, not list growth.

Would you rather 50,000 rows and zero windows, or 80 scored accounts and a calendar?

You need 7 sales, not 7,000 TAM rows.

Adapt or fail. Quality of signal beats size of spreadsheet.

Start Signals, Convert, Grow

FAQ

Is a small cohort enough for enterprise?

Enterprise still concentrates. You may need more named accounts, not more anonymous rows. Score the named set. Do not confuse coverage theater with windows.

Can I buy intent data instead of building this?

You can buy feeds. You still need fit filters, decay, and a send threshold. Bought intent without rails is another vanity list.

Where does enrichment fit?

After the score clears a threshold. Enrich survivors. Do not enrich the whole TAM first. That is how Clay becomes a spam press.

What if two accounts tie on score?

Prefer the one with the freshest stacked signal. Then prefer the one where you have a clear first line. Message quality still wins ties.

Do I need a data science team?

No. You need honesty and a spreadsheet that updates. Complexity that nobody trusts is worse than a three-signal score you can defend.