Lead scoring in Excel without a CRM
=HUNCH.ASK reads a line of lead notes and returns the probability that this person can approve a purchase on their own. =HUNCH.SCORE next to it returns how well they match your ideal customer, so the two columns together give SORTBY something to rank.
Column B: can this person buy
=HUNCH.ASK(A2:A9, "Can this person approve a purchase for their team without asking someone else?")In Google Sheets: =HUNCH(A2:A9, "Can this person approve a purchase for their team without asking someone else?")
The answer is a probability from 0 to 1. The question carries the definition of "decision maker", which is what makes the number mean something specific rather than a vague sense of seniority.
Column C: does this person fit
=HUNCH.SCORE(A2:A9, "ICP fit: marketing agency with 5 to 50 employees that manages paid ad campaigns for other companies, based in North America", "no fit|weak|good|perfect")In Google Sheets: =HUNCHSCORE(A2:A9, "ICP fit: marketing agency with 5 to 50 employees that manages paid ad campaigns for other companies, based in North America", "no fit|weak|good|perfect")
Write your ideal customer as one sentence: company type, size, who they sell to, where they are. Put that sentence in a cell of its own and reference it in the formula, so changing the ICP re-scores the whole column without touching the formula itself:
=HUNCH.SCORE(A2:A9, $F$1, "no fit|weak|good|perfect")In Google Sheets: =HUNCHSCORE(A2:A9, $F$1, "no fit|weak|good|perfect")
Real outputs
Eight lead notes for a marketing-agency tool, with the model's answers as returned on 28 September 2026.
| Row | A | B | C |
|---|---|---|---|
| 1 | Lead notes | Can buy? (0–1) | ICP fit (0–3) |
| 2 | Growth lead at a 15-person paid social agency in Austin. Manages the media budget directly. | 0.71 | 2.99 |
| 3 | Freelancer asking about a single seat for a personal project. | 0.17 | 0.09 |
| 4 | VP Marketing at a 30-person performance agency in Toronto, signs off on all vendor contracts. | 0.86 | 2.89 |
| 5 | Intern researching tools for a school assignment. | 0.08 | 0.18 |
| 6 | Owner of a 6-person boutique agency in Miami running Facebook ads for local restaurants. | 0.81 | 2.97 |
| 7 | Analyst at a 2,000-person holding company, says budget decisions go through corporate. | 0.15 | 0.06 |
| 8 | Co-founder of an 80-person performance marketing shop in London, evaluating for the whole team. | 0.69 | 0.48 |
| 9 | Account manager, no budget authority, wants a demo to show her director. | 0.05 | 0.63 |
Row 7 is the one that shows the ICP sentence doing real work: an 80-person performance marketing shop is exactly the kind of company the sentence describes, and the co-founder can clearly buy (0.69), but fit lands at only 0.48 because the sentence says 5 to 50 employees and this company has 80. The size clause is not decorative; loosen it if a larger agency should still count as a fit. Row 8 shows the opposite problem: the note says nothing about the company at all, only that this person cannot approve a purchase (0.05), and fit still comes back at 0.63 rather than 0, because there is nothing in the text to rule it out either. Thin notes do not fail loudly, they drift toward the middle, which is why a one-line summary of company type and size is worth the extra ten seconds per lead.
Rank and cut
=SORTBY(FILTER(A2:C9, (B2:B9>=E1)*(C2:C9>=2)), C2:C9, -1)Put the probability threshold in E1, starting at 0.5, so you can tune it without editing the formula. Rows that pass both filters get a call this week; everyone else gets a nurture email or nothing at all.
A probability near 0.5 is the model saying the notes do not say enough to tell. Read that row yourself, or add one more line to the notes, rather than rounding it to yes or no.
Limits
- The model judges only the text in the cell. It does not look up the company website, LinkedIn, or firmographic data.
- A one-line note gives the model less to work with than three lines, and the score will sit closer to the middle. That is honest, not broken.
- Two formulas per row is two credits per row. A 2,000-lead list is 4,000 credits, about $23 on the Starter and Top-up packs combined.
- Cells calculating together travel in requests of up to 40 rows, and answers are cached for six hours inside Excel.
Setup in three steps
- Sideload the add-in from hunchsheet.app/excel.
- Save your key once:
=HUNCH.SETKEY("hunch_..."), then delete the cell. - Type both formulas in row 2, then fill both down together.
See also
- HUNCH.ASK reference and HUNCH.SCORE reference.
- The same two-column pattern in Google Sheets, and scoring ICP fit alone if you only need that column.
- AI formulas in Excel, an overview of all four formulas.
Questions
Does it check the company website or LinkedIn for me?
Why two columns instead of one combined lead score?
Can I combine the two columns into one number anyway?
=B2*0.6 + (C2/3)*0.4. Keep the raw columns too; changing the weights then costs nothing.What happens when I change the ICP sentence?
How many leads can one column handle?
Run it on your own column
Install the Excel add-in (two minutes, Windows, Mac and the web), save your key with =HUNCH.SETKEY, and point the formula at your data. 100 rows free to start, then $29 for 5,000 rows. Credits never expire and there is no subscription.
Reference: HUNCH · HUNCHSCORE