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
What a “bullseye customer” is (and why startups shouldn’t start with “everybody”)
Lenny and Michael open with the core definition: a bullseye customer is the specific subset of a broader market most likely to adopt first. Michael uses classic examples (Amazon starting with books, Facebook starting with college students) to frame focus as the path to scale, not the opposite.
- •Bullseye customer = the first, most-likely-to-adopt slice of the target market
- •Ambitious founders want “everyone,” but successful products start narrower
- •Early focus helps you learn faster and build the right initial thing
- •Examples of focused starts: Amazon (books), Facebook (college students)
Michael Margolis’s 30-year path to fast, team-based research (GV, Google, Walmart, ethnography)
Michael shares the career arc that shaped this method: anthropology and deep ethnography, then compressing techniques at Walmart.com, scaling research at Google, and finally productizing the approach at GV across hundreds of startups. He introduces “watch parties” as a key mechanism for alignment and speed.
- •Background: anthropology → usability/testing → innovation consulting → Walmart.com → Google (Gmail) → GV
- •Core theme: compress deep research into faster cycles without losing insight
- •Watch parties emerged from Google’s early investment in streaming research
- •Scale of experience: 300+ hands-on research sprints across many industries
Bullseye customers vs. ICPs: specificity as a competitive advantage for learning
Michael distinguishes bullseye customers from typical ICPs/personas: bullseyes are usually even more specific and are chosen to accelerate learning and alignment. The emphasis is not the ultimate market, but the best initial segment to reduce noise and prioritize decisions.
- •Bullseye customers are typically narrower than teams’ existing ICPs/personas
- •Focus helps prioritize roadmap choices and interpret feedback
- •The goal is alignment: what we’re doing first, and for whom
- •Early segment focus is a proven pattern across breakout companies
The bullseye customer sprint in one formula: “five and three and one”
Michael gives the high-level sprint structure: interview five bullseye customers, show three simple prototypes, and run it all in one day with the whole team observing and debriefing. He explains why qualitative interviews, batching, prototype comparison, and team participation create speed and clarity.
- •Sprint formula: 5 customers + 3 prototypes + 1 day
- •Qualitative interviews provide the highest signal-to-noise for early learning
- •Batching interviews makes patterns obvious (data saturation around 4–5)
- •Comparative prototypes reduce fixation and yield clearer preference signals
- •Team watch + debrief replaces the need for long research reports
When to run the sprint: before building, during expansion, or when traction feels “polite”
Michael outlines ideal timing: before heavy investment in building, when entering new markets or customer tiers, or when something feels off post-launch. The sprint is positioned as repeatable—run it whenever major assumptions or motions change.
- •Best before you’ve sunk major time/money into building
- •Useful when expanding to new geographies or segments
- •Great for changing go-to-market motion (enterprise → self-serve, etc.)
- •Helpful when feedback is encouraging but not converting (“polite no”)
Step 1 — Align on goals: the 45-minute meeting to surface assumptions and “what keeps you up at night”
The process begins with a short but structured team meeting to clarify what you’re trying to learn. Michael uses prompts to elicit hypotheses, unknowns, and recurring debates so the sprint answers the questions that matter most.
- •Start with a focused meeting to define the key questions
- •Prompts: assumptions, hypotheses, team debates, and “what must be true to succeed?”
- •The questions determine who you should interview and what to prototype
- •The goal is clarity on learning objectives—not a broad wishlist
Step 2 — Define the bullseye customer: why “comically narrow” targets produce clearer learning
Michael explains how teams define bullseye customers by interrogating their shorthand definitions until they become concrete and recruitable. He emphasizes that if the target is too broad, results become ‘mushy’ and hard to interpret; narrowness reduces variables and increases learning confidence.
- •Bullseye definition depends on what you’re trying to learn (e.g., onboarding → new users)
- •Teams often resist narrowing; Michael pushes until the group is unmistakable
- •Narrow targets reduce variables and make insights easier to interpret
- •If the team agrees “these are our people,” they can’t dismiss the results later
Bullseye attributes: inclusion criteria, exclusion criteria, and triggers that make customers ‘ripe’
Michael provides a practical way to specify bullseye customers using three buckets: inclusion, exclusion, and trigger events. Lenny reads an example attribute list, and they discuss ‘VIP’ customers and how triggers (like a life event) change readiness to buy.
- •Inclusion criteria: who must be in the group
- •Exclusion criteria: who to set aside (experts, locked-in competitors, atypical edge cases)
- •Triggers: events that create urgency/readiness (new leader, incident, life milestone)
- •Heuristic: ~7 concrete, measurable attributes often gets you recruitable specificity
Step 2 example in action: specialty medication delivery and discovering the true ‘heat’ segment
Michael walks through a real case: designing delivery for expensive specialty prescriptions. The sprint reveals that predictability (narrow delivery window) matters more than speed, especially for refrigerated/cold-chain meds—leading to a refined bullseye and a second round to validate.
- •Initial question: ASAP delivery vs. precise delivery windows
- •Prototypes tested multiple delivery promises and mechanisms
- •Finding: predictable window > speed for many; critical for cold-chain/refrigerated meds
- •Outcome: refine bullseye to a sharper segment and rerun to build confidence
- •Signal of true fit: noticeable energy—“Can I sign up?”—vs. polite encouragement
Step 3 — Recruiting bullseye participants: screeners, panels (User Interviews), and paying enough to ensure show-up
Michael explains how he converts bullseye criteria into a screener questionnaire without ‘telegraphing’ answers. He shares recruiting via panel tools like UserInterviews.com, notes response volume variability, and emphasizes compensation and reliability practices (NDAs, reminders) to prevent no-shows.
- •Translate criteria into measurable screener questions; avoid obvious ‘right answers’
- •Common source: UserInterviews.com (plus alternatives like Respondent)
- •Response volume can range from dozens to hundreds depending on niche
- •Pay enough to ensure attendance (e.g., $125/hr consumer; higher for professionals)
- •For hard-to-reach roles (e.g., oncologists): snowball recruiting, conferences, associations
Step 4 — Three effective prototypes: compare ‘recipes,’ borrow competitors, keep it simple, and proofread
Michael describes building three distinct, lightweight prototypes (often simple PDFs) that stand on their own without narration. He encourages using competitor products as ‘free prototypes,’ focusing on crisp value propositions over high-fidelity interactions, and avoiding credibility-killing typos.
- •Create three distinct value-prop ‘recipes’ to enable comparison
- •Use competitor products when possible to shortcut prototyping
- •Keep prototypes simple (often static) to reduce overcommitment and distraction
- •Make each prototype visually distinct so observers can reference them easily
- •Proofread—small errors can derail trust and feedback quality
Step 5 — The interview guide: two-part interviews, rapport-building, and weighting past behavior over predictions
Michael lays out the interview structure: discovery first (stories, prior behavior, pain), then prototype comparison. He shares interviewing craft tips—smile, build rapport, adopt a ‘listener’ mode—and reinforces that past behavior is more reliable than what people claim they’d do.
- •Two-part interview: discovery (first half) → prototype compare/contrast (second half)
- •Discovery provides context that explains prototype reactions and objections
- •Interview craft: build rapport quickly; make the participant the ‘expert’
- •Practice matters; founders must shift from pitching to ‘humble inquiry’
- •Interpretation rule: anchor on past experiences, not stated future intentions
Step 6 — The watch party system: live-streaming, structured debriefs, predictions, and turning insight into momentum
Michael explains the operational setup that makes the sprint fast: the team watches live via stream, takes manual notes with assigned roles, and debriefs between interviews in a structured spreadsheet. They capture pre-interview predictions to combat hindsight bias, then end with a takeaways form to align on what changed and what’s next.
- •Separate observer stream keeps the participant experience 1:1 and less intimidating
- •Manual notes keep the team engaged; roles rotate to manage intensity
- •Between-interview debriefs map observations back to the key learning questions
- •Pre-study predictions create accountability and reduce hindsight bias
- •End-of-day takeaways form produces alignment and clear next steps without a report
- •Common learning: teams realize they need more research—even if they ‘talk to customers’ अक्सर
Common pitfalls, plus closing: avoiding ‘mushy’ recruiting, spotting bias, and applying the method to biotech
Michael highlights failure modes: recruiting too broadly, letting experts skew feedback, and falling for confirmation bias or overly trusting stated intentions. He closes with excitement about adapting these methods to biotech contexts (TPPs, clinical trials as products) and shares resources for listeners to try the sprint themselves.
- •Biggest failure mode: bullseye isn’t specific enough → conclusions feel ambiguous
- •Avoid stacking interviews with known experts or atypical edge cases
- •Watch for confirmation bias; teams should ‘police’ interpretations gently
- •Biotech application: productizing therapies, fitting physician workflows, improving trial accrual
- •Resources: learnmorefaster.com (free book, templates, demo videos); contact via email/LinkedIn
