customer-research

name	customer-research
description	Use when the user wants to conduct, plan, or synthesize customer research. Also use when the user mentions 'customer research,' 'ICP research,' 'talk to customers,' 'user interviews,' 'win/loss analysis,' 'why do customers churn,' or 'understand our customers better.'

Customer Research

Expert knowledge for planning, conducting, and synthesizing customer research that informs positioning, messaging, and product direction.

Why Customer Research is Marketing's Most Underused Lever

The best marketing copy is written in the customer's own words. Customer research reveals:

• The exact language customers use to describe their problem (use it verbatim in headlines)
• The trigger event that made them look for a solution (use it in email subjects)
• The fears they had before buying (address them in the sales process)
• Why they chose you over alternatives (your real differentiator, not the one you invented)

Research Methods by Goal

Goal	Best Method
Understand why customers buy	Jobs-to-be-Done interviews
Improve onboarding and activation	User interviews + session recordings
Find product-market fit signals	Survey (NPS + open-ended)
Understand why deals are lost	Win/loss interviews
Define ICP characteristics	Customer data analysis + interviews
Validate messaging	5-second test, preference tests

Jobs-to-be-Done (JTBD) Framework

Customers don't buy products — they hire them to do a job. The job is the progress they're trying to make in their life or work.

Three layers of the job:

1. Functional job: The practical task ("I need to generate monthly reports faster")
2. Social job: How they want to be perceived ("I need to look organized to my manager")
3. Emotional job: How they want to feel ("I need to feel in control of our data")

Marketing that only addresses the functional job leaves 2/3 of the motivation untouched.

JTBD Interview Script

Session setup: 45-60 minutes, recorded (with permission), one interviewer.

Opening: "I want to understand the process you went through when you decided to look for a solution like [product]. I'm not going to pitch anything — I want to understand your experience."

The timeline interview (work backwards from the purchase):

1. "Take me back to when you first realized you needed something different. What was happening?"
2. "What triggered that realization? Was there a specific event?"
3. "What did you do first when you decided to look for solutions?"
4. "What other options did you consider? Why did you look at those?"
5. "What made you choose [product] over those alternatives?"
6. "What were you worried about before you bought?"
7. "What would have made you not buy?"
8. "How has your life/work changed since using [product]?"

The magic question (ask this in every interview): "If [product] disappeared tomorrow and you couldn't use it, what would you do?"

The answer reveals your true competitive set and what the customer values most.

User Interview Best Practices

Recruiting:

• Aim for 5-8 interviews per segment (more interviews don't add proportional value after 5)
• Interview recent buyers (within 90 days) — memory is fresh
• Interview churned customers separately — they reveal unmet expectations

Interview principles:

• Ask about the past, not the future ("What did you do?" not "What would you do?")
• Follow curiosity, not the script — if something interesting surfaces, dig in
• Silence is a tool — pause after answers and let them fill it
• Never ask leading questions ("Don't you think X would help?")
• Never ask hypothetical questions ("If we built X, would you use it?")

Forbidden interview mistakes:

• Pitching your product during the interview
• Asking yes/no questions ("Do you like the dashboard?")
• Combining multiple questions into one

Transcript Synthesis

Raw transcripts are not insights. Synthesis turns observations into patterns.

Synthesis process:

1. Highlight quotes — every quote that: describes a pain in the customer's own words, reveals a trigger event, reveals a fear or objection, describes the benefit they wanted, describes how they measure success.

2. Extract to observations — one observation per quote: verbatim quote + customer type + category (pain / trigger / fear / benefit / success metric).

3. Group into themes — cluster similar observations. Name each cluster.

4. Count frequency — themes appearing in 3+ of 5 interviews are insights. Themes in only 1 are data points, not patterns.

5. Write insight statements — "[N] of [total] customers said [theme] in their own words. Example quotes: [quote 1], [quote 2]."

Win/Loss Analysis

Who to interview:

• Recent wins (within 30 days): understand what closed them
• Recent losses (within 30 days, especially competitive losses): understand what you're missing

Win/loss interview questions:

1. "Can you walk me through how you evaluated your options?"
2. "What were the top 3 things you considered?"
3. "What almost made you choose a different vendor?"
4. "What were you worried about with us?"
5. [For losses] "What would have changed your decision?"

Win/loss reporting: Track across 20+ interviews: win rate by competitor, top win reasons by frequency, top loss reasons by frequency, feature gaps mentioned in losses.

ICP Definition from Research

ICP dimensions to capture:

• Firmographic: Company size, industry, geography, funding stage
• Technographic: Tech stack, tools they currently use
• Psychographic: Attitudes, priorities, decision-making style
• Behavioral: How they evaluate software, buying process, team involved
• Situational: Trigger events that make them ready to buy right now

ICP validation test: Plot your best customers (highest LTV, lowest churn, most referrals) on each firmographic dimension. Where do they cluster? That cluster is your confirmed ICP.

The One Question That Predicts PMF

"On a scale of 0-10, how disappointed would you be if you could no longer use [product]?"

• If > 40% say "very disappointed" → product-market fit signal
• If < 40% → more work needed

Common Rationalizations

Rationalization	Reality
"We already know our customers — we built the product"	Founders' mental model of customers diverges from reality within months of launch. Research is a reset.
"5 interviews isn't enough data"	5-8 interviews reveal 80% of the patterns. 50 interviews don't add proportional insight after the first 8.
"Surveys are faster and cheaper than interviews"	Surveys quantify; they don't explain. You need interviews to know what to survey.
"We'll do research after we fix the obvious product issues"	Research tells you what the real issues are. "Obvious" issues are often symptoms of deeper causes.

Verification

• Research goal defined before recruiting begins
• Interview guide written and reviewed
• 5-8 interviews completed per segment
• Interviews recorded and transcribed
• Synthesis completed: highlight → extract → group → count → write insight statements
• Insights expressed as patterns (N of M), not as individual anecdotes
• ICP definition updated with research findings (firmographic + trigger events)