HUNCHMULTI: several questions, one read of the text
=HUNCHMULTI(A2:A500, "Mentions price?|Mentions support?|Wants to cancel?") answers up to 10 yes/no questions about each text and returns one probability column per question. In Excel it is =HUNCH.MULTI. The text is read once for all the questions.
Syntax
=HUNCHMULTI(text, questions)=HUNCH.MULTI(text, questions)| Argument | Required | What it is |
|---|---|---|
text | Yes | A cell or one-column range with the text to judge. |
questions | Yes | 1 to 10 yes/no questions separated by |, or a row of header cells holding one question each, such as C1:F1. |
The result is a grid with one row per text and one column per question, each a probability from 0 to 1. Put the formula in the first cell under your headers and it fills across and down.
Examples with real outputs
=HUNCHMULTI(A2:A7, "Does it mention the price or cost?|Does it mention customer support?|Is the person thinking of leaving or canceling?")In Excel: =HUNCH.MULTI(A2:A7, "Does it mention the price or cost?|Does it mention customer support?|Is the person thinking of leaving or canceling?")
| Row | A | B | C | D |
|---|---|---|---|---|
| 1 | Review | Price? | Support? | Leaving? |
| 2 | Support answered in ten minutes and fixed it. Worth every cent. | 0.91 | 0.97 | 0.08 |
| 3 | Third price increase this year. We're evaluating alternatives. | 0.99 | 0.01 | 0.78 |
| 4 | The app is fine but nobody replies to tickets anymore. | 0.02 | 0.96 | 0.55 |
| 5 | Great for small teams, we just renewed for another year. | 0.02 | 0.02 | 0.05 |
| 6 | Too expensive for what it does and support is slow. Canceling next month. | 0.99 | 0.99 | 0.98 |
| 7 | Would love a dark mode. | 0.01 | 0.02 | 0.09 |
Each question gets its own column, so one review can be yes on several. The last review is yes on all three. "Nobody replies to tickets anymore" is clearly about support (0.96) but only 0.55 on leaving, because it hints at leaving without saying it. Rows like that are the ones to read. "Worth every cent" counts as mentioning price (0.91), which is right for the question as written.
Questions in header cells
Write each question as a column header in row 1, then point the formula at the header row. The headers document the sheet, and editing a header re-asks that column.
=HUNCHMULTI(A2:A500, C1:F1)In Excel: =HUNCH.MULTI(A2:A500, C1:F1)
Headers can be short labels if the full question sits in a note, but the formula reads the header text itself. "Price?" works worse than "Does it mention the price or cost?".
When to use it
- Tags that can overlap: a review can mention price and support and churn at once. HUNCHPICK would force one.
- Coding survey answers into themes, one column per theme, then COUNTIF on each column above a threshold.
- Several flags on the same ticket or email, such as urgent, angry, mentions a competitor.
Cost and limits
- One credit per question per row: three questions on 100 rows cost 300 credits, the same as three HUNCH columns. The text is sent once per row, so it is faster and lighter than three separate formulas.
- Up to 10 questions per formula. Split longer lists across two formulas.
- Blank cells and repeated texts are free. Answers are cached for six hours.
Questions
Can I mix yes/no questions with categories?
Do the questions influence each other?
What happens if a header cell is empty?
Run it on your own column
Copy the template (or use the Excel add-in), paste your key, point the formula at your data. Every formula here has its Excel spelling underneath. 100 rows free to start, then $29 for 5,000 rows. Credits never expire and there is no subscription.
More reference and guides
- HUNCH: the probability that the answer is yes HUNCH
- HUNCHPICK: one option from your list HUNCHPICK
- HUNCHSCORE: a position on your own scale HUNCHSCORE
- Analyze NPS comments in Google Sheets NPS comments
- Categorize support tickets in Excel Ticket triage
- Analyze open-ended survey answers in Excel Survey coding