Title: How to Conduct Customer Interviews, with Freckle’s Nathan Merzvinskis

Session context

  • Format: Talk + live Q&A + doc walkthrough (pre-mortem, interview guide, pricing survey, design partner program)
  • Host: Michael Batko (Startmate); Speaker: Nathan Merzvinskis (Freckle; ex-Everproof → acquired by Go1; ex-Schedulo)
  • https://www.linkedin.com/in/nathan-merzvinskis/
  • Focus: End-to-end customer discovery system from idea → MVP → pricing → ongoing feedback loops

Why customer interviews (beyond “talk to users”)

  • Reduce time-to-wrong: validate need, ICP, and monetization before building
  • Build champions: discovery participants become design partners, then first paying customers
  • Fundraising edge: evidence-rich discovery (transcripts, tagged insights, quotes) answers investor risks upfront

Core workflow overview

StagePurposeOutput
Pre-mortemIdentify top risks/assumptionsClustered assumptions/failure reasons; prioritized risk quadrant
Interview guideNon-leading, specific questions30–45 min script centered on real tasks, events, and pains
SynthesisTurn qual into decisionsTagged insights; persona fit/no-fit; next tests
Design partner programRepeatable feedback mechanism5–10 partners; biweekly sessions; clear give/get; path to paid
Pricing discoveryTest value metric + willingness to payVan Westendorp range + pricing hypotheses
Ongoing loopsPost-MVP prioritization and feature tradeoffsSlack channels, rapid sprints, continual interviews

Pre-mortem (start here)

StepHowWhy it matters
Solo brainstorm (20 min)Green: “I assume that …”; Pink: “This failed because …”Surfaces implicit beliefs and failure modes
Cluster + labelCombine into themes (pricing, ICP, regulation, competition, unit economics, etc.)Reveals blind spots across co-founders
Prioritize on 2×2Likelihood (low→high) vs Impact (low→high)Focus interviews on top-right risks first

Common top-right risks:

  • Real need/pain priority for target user
  • Willingness to pay / viable value metric
  • ICP clarity (who buys, who uses)
  • Regulatory or switching constraints (market dependent)

Interview fundamentals

PrinciplePractice
Avoid biasNo pitching up front; disclose only the problem space; ask non-leading, specific, time-bound questions
Specific over general“Walk me through yesterday” vs “usually”; “last time you …” vs “would you …”
Persona consistencyRun 6–7 interviews per persona; switch personas if you see weak pull
Record and tagGet consent; use Fathom/Granola for transcripts; tag in Dovetail (or equivalent)
Build rapport, then digReflect back terms; ask “Can you tell me more?”; chase throwaway comments

Favorite question patterns:

  • Context: “Was yesterday a typical workday? Walk me through it from start to finish.”
  • Pain: “Top 3 biggest challenges last week? Which cost you time/money/stress?”
  • Trigger: “Tell me about the last time you tried to solve this. What did you do first?”
  • Substitutes: “What tools/spreadsheets/processes are you using now? What’s broken?”
  • Constraints: “What would stop you switching? Who else needs to approve?”
  • Value: “If this were solved, what improves? Time saved, revenue, risk, compliance?”
  • Price framing (later): “At what point would this feel too cheap/too expensive?”

What not to do:

  • Don’t pitch at the start
  • Don’t ask hypotheticals (“Would you use/pay?”) without anchoring to past behavior
  • Don’t skip qualitative before surveys or AI-simulated “users”

Designing and running the guide

SectionExamples
PrimerExplain discovery vs pitch; request permission to defer solution until end; ask to record
KickoffRole, responsibilities, tools, org structure
Day-in-the-lifeTimeboxed walk-through; locate pain moments
Problem deep diveLast incident, frequency, workaround, cost-of-pain
AlternativesCurrent stack (incl. spreadsheets/do-nothing); switching frictions
Buying dynamicsWho decides; budget; procurement constraints
WrapOffer design partner slot; ask for 1–2 intros to similar peers

Timebox: 30–45 minutes. Be explicit that you’ll read from a guide to stay on track.

Synthesis and decision-making

ToolingUse
Dovetail (or Notion tags)Upload transcripts; tag by theme (pricing, ICP, workflow, compliance, etc.)
Persona boardsSPICE framework: Situation, Pain, Impact, Critical event; include representative quotes
Risk trackerMap new evidence back to pre-mortem clusters; mark validated/invalidated/unknown

Stopping rule:

  • After 6–7 consistent interviews per persona, you can usually predict answers → move to next risk or next persona.

Design partner program (turn interviews into revenue)

ElementDetails
Who5–10 ideal users with strong pain and engagement
CadenceBiweekly 30–45 min sessions; Slack 1:1 channel for fast feedback
Give/getEarly access; 50% Y1 discount; input to roadmap; you get structured feedback and case study
ExitClear expectation to become paying customers post-beta if value delivered

Pricing discovery (when to run and how)

  • Run after need and scope are clear; before heavy build
  • Define value metric(s): seats, credits, usage unit most aligned with delivered value
  • Use Van Westendorp (4 Qs) to get price bands; add willingness-to-pay probes in interviews
  • Validate packaging and buyer economics in follow-ups; keep evidence for investors

Reference: Nathan’s Medium guide on pricing discovery (Van Westendorp) and value metrics.

Post-MVP: keep interviewing

  • Use interviews to prioritize features with equal build cost/time
  • Set up 1:1 Slack channels with top customers (e.g., 25 key accounts)
  • Continue short sprints: question set → calls → decision → ship

Benchmarks and pacing

AreaPractical bar
Interview volume10–20 across 1–3 personas; 6–7 per persona before switching
Guide disciplineNon-leading, time-bound, past-behavior anchored
Evidence qualityRecorded, transcribed, tagged; quotes curated into persona boards and investor-ready notes
Conversion20–50% of strong-need interviewees join design partner program

Selected Q&A highlights

TopicTakeaway
B2B vs B2C countsSimilar “6–7 per persona” rule; ensure persona consistency
Warm vs cold contactsWarm enables lighter disclosure; still avoid pitching upfront
CompetitionNot inherently bad; beware winner-takes-all markets; include substitutes and “do nothing”
Quant surveys/AI “fake users”Use only after qualitative rounds; AI simulators miss serendipitous insights
Recording toolsFathom (fast summaries), Granola (invisible presence), Dovetail for tagging
Fundraising without productPossible with high-quality discovery evidence, mockups, and design partners
Post-launch interviewsYes—use to choose between features with similar effort; maintain slack channels for continuous input
Feature whiplashKeep 80/20: founder vision 80%, feedback 20%; avoid reactive, incoherent product

8-week execution template (fill this in)

TrackBaseline (now)2-week goalWeekly targetSuccess criteria
Pre-mortemAssumptions not capturedRun session; cluster and prioritize risksTop-right risks defined for discovery focus
Interviews0–5 recent convos10 interviews across 1–2 personas5–10/wk; recorded + tagged6–7 consistent per persona; clear need signal
SynthesisUnstructured notesDovetail/Notion board; SPICE personas with quotesUpdate after each batchDecisions logged; next hypotheses queued
PartnersNone or ad hoc5+ design partners confirmedBiweekly sessions; Slack channelsOngoing feedback; 50%+ convert to paid later
PricingUnclear value metric/WTPDefine metric(s); run VWSM survey with partners5–10 responsesPrice band + packaging hypothesis validated

Practical metrics you can choose

  • Need strength: # interviews where pain ranks top 3; cost-of-pain evidence
  • Signal quality: % of interviews producing referrals; willingness to be a design partner
  • Pricing: VWSM acceptable range; early trial → paid conversion
  • Evidence readiness: transcripts tagged; investor-ready quote board (“wall of love”)

Workshop logistics and follow-ups

ItemDetails
ToolsFathom/Granola (record), Dovetail/Notion (tag), Miro (pre-mortem), Calendly (booking)
ArtifactsPre-mortem board; interview guide; SPICE persona boards; pricing survey; partner one-pager
Investor packCurated quotes, tagged insights, persona summaries, pricing results

Your 5 next actions

  1. Schedule a 90-minute pre-mortem with your co-founder; prioritize top-right risks
  2. Draft a 30–45 min non-leading interview guide; book 10 calls across 1–2 personas
  3. Record, transcribe, and tag; create SPICE persona boards with verbatim quotes
  4. Invite 5–10 strong-fit participants to a structured design partner program
  5. Define value metric(s); run a Van Westendorp pricing survey with partners and synthesize results