Customer Interviews for Real Decision Criteria
A step‑by‑step guide to interview for the real rules customers use to choose, then turn them into a one‑page Decision Criteria Card.

This guide is for working PMs who must decide what to build when users say many things. You will interview for the rules customers use to choose. At the end you will have a one‑page Decision Criteria Card with ranked drivers, hard thresholds, and likely winners for your next build choice.
Elicit decision heuristics: rankings, thresholds, triggers, and costs avoided.
Ask for choices under constraint, not opinions or wishlists.
A one‑page Decision Criteria Card you can share and test.
Why choices, not stories
Stories are context. Choices are rules. In interviews, people recall anecdotes and pains. Useful, but not enough. Your job is to pin down what they choose when trade‑offs are real. That means fixed price buckets, explicit costs, and forced rank.
| Collect | Skip | Why |
|---|---|---|
| Choices between concrete options | Open opinions on "what do you want" | Choices reveal decision rules |
| Thresholds (max price, max delay, min accuracy) | Vague ranges that drift | Thresholds constrain viable designs |
| Costs avoided and triggers to switch | General pain statements | Costs and triggers show timing and urgency |
The steps
-
Frame the decision
State the decision the customer makes. Name the options as they see them. Fix the context: job to do, who clicks, and when. Define the price buckets you will use and hold them steady.
Tip: If you need a ready structure, see the Interview guide for prompts and flow.
-
Map the current workflow
Walk through a recent case end to end. Capture actors, tools, touchpoints, wait times, and handoffs. Note where the decision happens and what information is available at that point.
-
Run choice and cost probes
Present concrete options. Change one variable at a time. Ask what they would pick now, given that change, and why. Log the cost they avoid with each pick.
-
Force trade‑offs
Use constrained scenarios. Limit budget, time, or staff. Ask for a rank order of options. Ask what they drop first and what they never drop.
-
Validate with a micro‑experiment
Turn the top rule into a small test. Use a fake door, a prototype, or a manual concierge run. Measure revealed choice, not clicks alone.
-
Synthesize the Decision Criteria Card
Write ranked drivers, hard thresholds, and the likely winner under common contexts. Add the decision rules and one testable hypothesis per rule.
Three question patterns
- Thresholds: "At what point would you not proceed?" then tighten: "So if it takes 24 hours, is that still okay? 48?"
- Anchors: "Last time you chose X. What would have made you switch?" anchor in a real case, then nudge one variable.
- Avoided costs: "If you pick option A, what cost or risk does it remove today?" push for specific time or money saved.
Turn verbatims into rules
Right after each interview, write rules from the words you heard. Use a simple form: if X and cost < Y, then choose A; else choose B. Keep variables concrete. Keep price buckets fixed. Note any trigger that flips the rule.
From there, write testable hypotheses. For example: "When setup time is under 10 minutes and per‑invoice fee is under $2, 70% choose built‑in over export." Then design a micro‑experiment to check the split.
Checklist and traps
- Fix the context before asking choices.
- Use stable price buckets. Do not let them float mid‑interview.
- Ask for rank orders and drops, not feature wants.
- Record thresholds in exact numbers or clear bounds.
- Write a rule after each answer while it is fresh.
- Do not treat a story as a rule. Seek the choice, then the reason.
- Do not equate frequency with importance. A rare blocker can still be a hard stop.
- Quick win: run three 15‑minute interviews this week to draft your first Card, then expand.
Worked example: Add an in‑app invoice generator?
Context: You manage a B2B tool used by small agencies. Today, users export project data to a spreadsheet or an accounting app to invoice clients. You are deciding whether to add an in‑app invoice generator.
Invite: "I want to learn how you create and send invoices after a project. We will walk through your last one, then make a few choices with fixed budgets and times. 15 minutes."
Map the workflow: User closes a project on Friday, gathers time entries, applies rates, drafts an invoice, sends a PDF, and logs payment. Tools: your app, a spreadsheet, and an accounting app. Decision point: where to assemble and send the invoice.
Choice probes (fixed buckets):
- Option A: In‑app invoice, setup 10 min, per‑invoice fee $1.50, PDF + email send.
- Option B: Export CSV to accounting app, zero fee per invoice, 20 min manual mapping.
- Option C: Spreadsheet template, 15 min, no fee, manual numbering.
Ask: "Which would you choose for your next invoice? Why?" Then vary one input per round:
- Round 2: In‑app fee $3.00, setup 5 min. Choose again.
- Round 3: Client needs same‑day invoice. Choose again.
- Round 4: Team is short‑staffed Friday. Choose again.
Forced trade‑offs: "Rank A, B, C for next Friday with two invoices due and one junior staff." "What do you drop first if you must cut time by 10 minutes?" "What do you never drop?"
Verbatims into rules (examples):
- If same‑day send is required and setup ≤ 10 min and fee ≤ $2, choose in‑app; else export.
- If mapping takes ≥ 15 min and staff is junior, choose in‑app even if fee is $3.
- If client requires accounting app branding, export wins regardless of time.
Hard thresholds surfaced:
- Max acceptable per‑invoice fee: $2 for most routine work.
- Max total time to send: 30 minutes on Fridays; 10 minutes for rush jobs.
- Non‑negotiable: Correct tax lines and sequential numbering.
Likely winners by context:
- Rush jobs: In‑app wins if setup ≤ 10 min and fee ≤ $2.
- Routine monthly batches: Export to accounting app wins due to reports.
- One‑off small invoices: Spreadsheet template is acceptable if under 3 lines.
Micro‑experiment: Add a "Create Invoice" in‑app entry point after project close (fake door). Show a short setup flow with the promised 10‑minute time and a fee preview of $1.50. Offer a fallback link to export. Measure the split between in‑app start vs export, and the drop when fee toggles to $3.00 for half the traffic. Success metric: share of users who proceed past fee confirmation under each price.
Decision Criteria Card (draft):
Decision Criteria Card: In‑app Invoice Generator
Ranked drivers: 1) Time to send, 2) Accuracy/compliance, 3) Per‑invoice fee, 4) Staff skill, 5) Reporting needs.
Hard thresholds: Time ≤ 10 min for rush; ≤ 30 min Friday batch. Fee ≤ $2 for routine. Must include correct tax and numbering.
Likely winners: In‑app for rush and junior staff; Export for batch and reporting; Spreadsheet for tiny jobs.
Rules: If rush and fee ≤ $2 and setup ≤ 10 min, choose in‑app; else if reporting needed, export; else choose spreadsheet if ≤ 3 lines.
Next test: Fake door + fee toggle to validate fee threshold and time driver.
Run it fast, then scale
Start with three 15‑minute interviews this week. Use fixed options, collect thresholds, and write rules. Build the first Card. Share it with design, engineering, and sales. Then run a wider study with the same probes, in the same order. Keep the buckets and the script stable so the rules compare.


