Lenny's PodcastMichael Margolis: Why bullseyes beat ICP personas in one day
Through five qualitative interviews and three prototypes in one day; team watch parties hit data saturation fast and reveal who's a real bullseye yes.
CHAPTERS
- 0:00 – 5:57
What a bullseye customer is—and why focus beats “build for everyone”
Michael defines a bullseye customer as the narrow subset of a market most likely to adopt first, using examples like Amazon (books) and Facebook (college students). He explains why selecting this initial wedge unlocks faster learning, clearer prioritization, and stronger team alignment.
- •Bullseye customer = most likely early adopters within the broader target market
- •Big markets still start with a focused entry point
- •Focus helps prioritize roadmap decisions and interpret feedback correctly
- •Creates shared clarity across founders, product, and investors
- 5:57 – 9:11
Michael Margolis’s research journey: from anthropology to GV and “watch parties”
Michael traces how decades of anthropology, usability testing, enterprise consulting, Walmart execution speed, and early Google research infrastructure shaped his approach. He shares how working with hundreds of startups at GV led to compressing research into repeatable, high-velocity sprints.
- •30+ years across anthropology, usability, ethnography, and product research
- •Learned to compress deep research into fast cycles at Walmart and Google
- •Introduced/standardized “watch party” methods for team-wide learning
- •Conducted 300+ hands-on research sprints across many industries
- 9:11 – 13:24
Bullseye customer vs. ICP: specificity, learning, and early adoption
The conversation distinguishes bullseye customers from typical ICPs and personas: bullseyes are usually even narrower and defined for rapid learning. The goal is not to exclude forever, but to choose the best first adopters so signals are strong and decisions are grounded.
- •Bullseye is often more specific than a typical ICP/persona
- •Designed to accelerate learning and reduce noise from mixed audiences
- •Helps filter external feedback (investors, friends, broad “target market”)
- •Clarifies sequencing: who to serve first vs. later
- 13:24 – 20:55
The bullseye customer sprint overview: “5 and 3 and 1”
Michael lays out the sprint structure: interview five bullseye customers, test three simple prototypes, and do it in one day with the whole team observing. He explains why qualitative interviews in tight batches plus prototype comparison create fast pattern recognition.
- •Formula: 5 customers + 3 prototypes + 1 day
- •Qualitative interviews deliver high insight density and motivation context
- •Clumping interviews makes patterns obvious (data saturation)
- •Multiple prototypes prevent over-attachment and enable richer comparisons
- 20:55 – 22:29
When to run the sprint: before big build, during expansion, or when traction feels “polite”
Michael explains the best moments to use the sprint: early before investing heavily, when entering new markets or segments, or when you’ve launched but results are underwhelming. It’s also a troubleshooting tool when feedback sounds positive but doesn’t convert into action.
- •Use before heavy build spend to avoid wasted effort
- •Use when changing segment, geography, or sales motion
- •Use when traction is unclear or feedback is “sounds good… but no”
- •Repeatable method—can be run multiple times as you refine
- 22:29 – 24:01
Step 1: Align on goals, hypotheses, and the questions that “keep you up at night”
Step one is a short team meeting to surface the real uncertainties, assumptions, and internal debates. The output becomes the learning agenda that drives customer selection, prototyping, and the interview guide.
- •45-minute team session to identify key questions and assumptions
- •Prompts: what must be true for success, what worries you most
- •Surfaces recurring debates to resolve with evidence
- •Creates a shared learning objective for the sprint
- 24:01 – 29:09
Step 2: Define the bullseye customer—and why “comically narrow” is the point
Michael describes working backward from your key questions to determine who you need to talk to (new users vs. existing, etc.). He argues for extreme narrowness to reduce variables—otherwise results feel “mushy” and hard to interpret.
- •Customer choice depends on what you’re trying to learn (onboarding vs. feature)
- •Teams tend to use shorthand; facilitator pushes specificity with ‘what do you mean?’
- •Narrowness reduces confounds and strengthens conclusions
- •Bullseye definition is a hypothesis you expect to adjust after interviews
- 29:09 – 38:23
Step 2 example: specialty medication delivery—and discovering the real “heat” attribute
A detailed case study shows how interviews revealed predictable delivery windows mattered more than speed. The team learned the highest-need subset involved refrigerated (cold-chain) meds and theft/weather concerns—leading to a refined bullseye and a second sprint.
- •Initial question: ASAP vs. predictable delivery window
- •Prototypes explored different delivery promises and methods (including drones)
- •Insight: predictability beats speed for many; cold-chain meds create acute need
- •Outcome: refine bullseye and re-run sprint to build confidence
- 38:23 – 43:32
How to specify attributes: inclusion, exclusion, and trigger events
Michael provides a practical rubric for defining bullseyes: inclusion criteria, exclusion criteria, and triggers that make someone “ready now.” Triggers (e.g., new exec, incident, new baby) often separate ‘should’ from ‘will.’
- •~7 attributes is a common level of specificity
- •Inclusion criteria: who must be in the sample
- •Exclusion criteria: who would distort learning (experts, locked-in competitors)
- •Triggers: events that create urgency/readiness (new job, breach, marriage/baby)
- 43:32 – 56:13
Step 3: Recruit and pay participants—screeners, non-telegraphed questions, and show-up reliability
Recruiting starts by translating criteria into a screener questionnaire that avoids giving away “correct” answers. Michael shares tools (User Interviews, Respondent), typical response volumes, compensation norms, and how to handle hard-to-reach audiences.
- •Turn bullseye criteria into measurable screener questions
- •Avoid telegraphing the right answers; prefer indirect/open-ended prompts
- •Userinterviews.com enables fast recruiting (often hundreds of responses)
- •Pay enough to ensure attendance; use alternative methods for niche pros (e.g., oncologists)
- 56:13 – 1:01:17
Step 4: Build three effective prototypes—simple PDFs, strong positioning, and comparison-ready variants
Prototypes are intentionally lightweight: often single-page ‘homepage’ PDFs or competitor products used as “free prototypes.” The goal is clear, blunt value propositions that stand alone without narration, enabling shoppers-style comparison across distinct recipes.
- •Create 3 distinct “recipes,” not small variations A/B/C
- •Use competitor products when possible to learn faster
- •Keep prototypes flat and simple; avoid over-commitment to build artifacts
- •Prioritize crisp copy/positioning; ensure prototypes look distinct for easy discussion
- 1:01:17 – 1:08:54
Step 5: Draft the interview guide—two-part structure and interviewer mindset
Each interview has a discovery half (past behavior, stories, failures) and a prototype comparison half. Michael emphasizes rapport-building, smiling, genuine curiosity, and practicing a ‘listener character’ distinct from founder selling mode.
- •Two-part interviews: discovery first, prototypes second
- •Use past experiences as truth—stories provide context for prototype reactions
- •Build rapport quickly; make participant the expert (humble inquiry)
- •Practice interviewing; adopt a deliberate “listener mode” to avoid pitching
- 1:08:54 – 1:19:40
Step 6: The watch party method—team note-taking, debrief spreadsheet, and prediction snapshots
The sprint culminates in a structured watch party: the interviewer speaks 1:1 with participants while the team observes via livestream, takes manual notes, and debriefs between sessions. The day ends with a takeaways form and a comparison to pre-interview predictions to combat hindsight bias.
- •Separate interview Zoom from team livestream to avoid intimidating participants
- •Manual note-taking keeps the team engaged; roles rotate across interviews
- •Between-interview debriefs fill a structured spreadsheet tied to key questions
- •Collect predictions before and takeaways after; compare to reduce hindsight bias
- 1:19:40 – 1:29:38
Common pitfalls, signals of “no,” and closing reflections (including biotech applications)
Michael highlights failure modes: weak bullseye specificity, over-weighting what people say they’ll do, and confirmation bias in interpretation. He closes with thoughts on applying these methods to biotech and shares where to find the free book and resources.
- •Biggest pitfall: recruiting too broadly leads to mushy conclusions
- •Trust past behavior over stated intent; people are bad predictors
- •Actively counter confirmation bias during debriefs
- •Resources: learnmorefaster.com, contact via email/LinkedIn; exploring biotech/TPP parallels
