Product discovery
Customer interviews, problem framing, opportunity-solution trees. The work that prevents shipping the wrong thing fast.
interviews · OST · evidenceDevelopment · 04 of 04
A product team. Not a feature factory.
Discovery, design, engineering, data — one pod, one outcome. We take a thesis and turn it into a product in market, with the metrics to prove it landed. Long-running, dual-track, opinionated on the customer.
0→1To market in 4–6 months
1→NLong-running pods · 12–24 months
Dual-trackDiscovery & delivery in one cadence
What you get
A pod with everyone you need — in one room.
Six capabilities every product pod earns its keep with.
Design & UX
Flows, prototype, design system, accessibility. Design that ships, not design that lives in Figma forever.
flow · prototype · DSEngineering across the stack
Web, mobile, backend, infra under one tech-lead. No handoffs — the same humans own the experience and the database.
web · mobile · backend · infraAnalytics from day one
Activation, retention, NPS, North-Star metric instrumented before launch — not after. Telemetry as a feature.
north-star · activation · retentionGo-to-market enablement
Pricing experiments, onboarding instrumentation, sales-led / PLG split. We build the product — and the way it gets sold.
PLG · pricing · onboardingOwnership, not handover
The pod runs the product after launch — iteration, on-call, growth experiments — until you're ready to bring it in-house.
own · iterate · transitionHow a pod looks
Small, senior, cross-functional.
A typical 0→1 pod — six people, one outcome. We resize for 1→N when it's time to scale.
One pod, one outcome, one weekly demo.
The pod doesn't hand off design to engineering or engineering to QA. The same six humans own the product from interview to incident.
- 01
Dual-track from day one
Discovery runs in parallel with delivery — not as a phase. Evidence informs the backlog continuously.
- 02
Weekly demo & metric
Every Friday: working software, a customer call summary, one metric moved or measured.
- 03
Long enough to compound
Most pods run 6–18 months. Long enough to know the product, short enough to stay sharp.
- 04
Hand off when it's warm
When you're ready, we transition to your team — with the documentation and the customer relationships intact.
Tools & rituals
A pod's operating system.
Not a methodology religion. A small set of tools and rituals we've seen survive the messy middle.
Discovery
Customer interviewsOpportunity-solution treeContinuous evidence
Design
FigmaTokens StudioMazeUserTesting
Build
Next.jsSwiftKotlinGoPython
Analytics
AmplitudePostHogMixpanelSegment
Experimentation
StatsigLaunchDarklyGrowthBook
Collaboration
Linear · ShortcutNotionSlackLoom
Rituals
Weekly demo2-week iterationMonthly retroQuarterly bet
What we'll skip
SAFe ceremoniesStory-point theatreRACI matrices
From vision to victory
Thesis to market, in five movements.
The shape every successful 0→1 pod we've run has followed — with different details, same rhythm.
01
Wk 1–2Thesis & bets
Customer thesis, north-star metric, three biggest assumptions — written down, then tested.
02
Wk 3–5Validate & design
Interviews, prototype, design system, scoped MVP. The bet narrows; the system widens.
03
Wk 6–14Build to beta
Two-week iterations, weekly demo, telemetry live, real users by the end.
04
Wk 15–18GA & launch
Soft launch, billing, pricing experiments, public GA when activation hits target.
05
OngoingIterate or hand off
Pod stays for 1→N growth, or hands off to your team with docs and customer hand-off.
Where it lands
Three pods that shipped a product.
No internal tool with a coat of paint. Each one a product in market with paying customers.
From whiteboard to GA in 18 weeks.
B2B SaaS spinout. Pod ran discovery, design, build, launch, and the first three growth experiments. Hit $1M ARR in month nine.
18 wkTo GA
$1M ARRMonth 9
Next.jsPostgresStripe
The pod that doubled activation.
Marketplace founder hired the pod to fix onboarding. Three quarters in, activation 61% → 84%, retention up, NPS +24.
+38 ptsActivation
+24 NPSSame period
RemixPostgresStatsig
Consumer mobile product — both stores, both teams.
Lifestyle app for a consumer brand. Pod ran 22 weeks, shipped to both stores in lockstep, transitioned to client team in week 26.
22 wkTo both stores
4.7 ★Avg rating
SwiftUIComposeNode
Why ETY
A pod that owns the outcome.
18 wkMedian time from thesis to GA on 0→1 pods.
6 peopleTypical pod size — PM, design, tech lead, 2 engineers, fractional SRE.
Weekly demoEvery Friday. Customers in the room more often than not.
Long enoughMedian pod runs 14 months — we learn the product, then we improve it.
Turn a thesis into a product.
Send us the thesis, the customer you have in mind, and the one metric you'd defend. We'll come back with a pod shape, a six-month plan, and the bets we'd test first.