Reading Notes

Chandra Janakiraman on Product Strategy

Source: Chandra Janakiraman — An Operator's Guide to Product Strategy

Notes — Chandra Janakiraman on Product Strategy

Four questions [Adler frame]

Q1 — What is it about? An operator’s synthesis of strategy literature (Rumelt, Roger Martin, Porter, Sun Tzu) into a reliable 8–12-week procedural playbook that any product team can run — demystifying strategy as a learnable, repeatable process rather than an innate talent.

Q2 — How is it argued? Via a worked playbook with specific time-boxes, concrete deliverables, and illustrative examples drawn from Janakiraman’s own experience: Headspace (strategy birth story), Zynga (strategy as embedded culture), Meta Reality Labs (Oculus vs Portal divergence). The argument is partly retrospective (Zynga’s three pillars interpreted through the framework) and partly prescriptive (Meta’s privacy and growth strategy as deliberate applications).

Q3 — Is it true? The process is sound but highly PM-centric; it assumes a competent facilitator who can hold a cross-functional working group together. The “resonance” metaphor is evocative but not predictive. The four scoring criteria (impact, certainty, levers, differentiation) are the right variables but scoring them is still political and judgment-based. The Oculus/Portal divergence example shows that even good process cannot guarantee outcome — ultimately strategy is a hypothesis tested by execution.

Q4 — What of it? This is the most actionable product strategy playbook available in the wiki. The newspaper headline / winning aspiration technique and the “fruit story” for leadership interviews are immediately applicable. The small S / big S distinction is important: most PM work is small S, but dedicating design/UXR to big S in parallel is underused. The AI-assisted mock strategy point is forward-looking and worth revisiting as LLM tooling matures.

Glossary

Product strategy — sits between mission/vision and the roadmap; forces choice to deploy scarce resources for maximum impact; ideally three strategic pillars (focus areas) + explicit non-focus areas + why.

Resonance (strategy metaphor) — from physics: apply a frequency close to an object’s natural frequency and amplitude increases disproportionately. Analogy: the right strategy creates disproportionate product-market impact.

Small S strategy (present-forward) — problem-focused, 2-year horizon, led by PMs; 8–12-week process; the standard playbook described at length.

Big S strategy (future-backward) — aspirational, 5–10-year horizon, led by design/UXR; three distinct future scenarios → concept car prototypes → UXR testing → live product testing; ongoing up to 6 months.

Strategy working group — cross-functional team of PM + engineering + design + data (minimum); the group owns the strategy collaboratively, PM facilitates.

Preparation phase — 4 weeks; six inputs: (1) behavioural insights meta-analysis, (2) UXR insights meta-analysis, (3) leadership interviews, (4) competitive analysis, (5) adjacent roadmaps, (6) user observations. Output: comprehensive preparation readout deck.

Strategy sprint — 3–5 days; the heart of the process. Day 1: share-out and problem enumeration. Day 2: problem clustering → opportunity framing → scoring on 4 criteria → select 3 strategic pillars → generate how-might-we’s. Day 3: winning aspiration via newspaper headline technique.

Four scoring criteria — (1) expected impact on key metric; (2) certainty of impact (evidence quality); (3) clarity of levers (do you know how to solve it?); (4) differentiation (are you uniquely positioned to build it?). Score high/medium/low, debate, add, sort — top 3 become pillars.

How-might-we — positively framed question (e.g. “how might we improve discovery?”) derived from each strategic pillar; input to the design sprint. Psychologically opens ideation vs “how do we fix X?”

Winning aspiration — one aspirational statement synthesised from a newspaper-headline exercise: the working group writes headlines describing what success on all three pillars looks like in 2 years; merge into a single coherent statement.

Design sprint — 1 week; generates illustrative concept designs (not feature-ready) for each strategic pillar; led by design. Goal is to give stakeholders something concrete to visualise strategy — “a picture is worth a thousand words.”

Fruit story — leadership interview heuristic: if you don’t ask leaders what they want before presenting, you will bring the wrong fruit and be repeatedly rejected. Ask first: what does success look like? What are principles? What are your pet ideas?

Gatekeepers — 2–3 critical senior stakeholders who must align before the strategy is rolled out more broadly; met in individual 1-on-1s first.

Concept cars (big S) — prototypes of distinct futures built to inspire, not to ship; analogous to automotive concept cars (never commercialised; may yield a production technology or feature).

Key notes

Small S — five-stage process

Stage 1: Preparation (4 weeks) Form strategy working group (PM, eng, design, data). Kickoff meeting sets purpose, phases, action items. Six parallel work streams:

  1. Behavioural insights meta-analysis — data person synthesises historical analytical work into macro themes.
  2. UXR insights meta-analysis — design/research synthesises qualitative signals (interviews, support, social).
  3. Leadership interviews — “fruit story” principle; ask what success/failure looks like, success metrics, principles, pet ideas. Divide leaders across working group members.
  4. Competitive analysis — feature release patterns → infer competitor investment angles; head-to-head heat map.
  5. Adjacent roadmaps — cross-team dependencies; avoid strategy collisions.
  6. User observation — each working group member watches a user session; write learnings. Purpose: build empathy, not action specific insights.

Output: comprehensive preparation readout deck (single master deck).

Stage 2: Strategy Sprint (3–5 days) Day 1: Share-out. Each working group member presents their preparation deliverable. Audience takes notes on problems/gaps/sub-optimal areas.

Day 2 (most important day):

  • Problem generation: free-flowing enumeration of all observed problems. Capture in Google Sheets.
  • Clustering: group related problems into 10–15 clusters.
  • Opportunity framing: flip each cluster from problem language (e.g. “difficulty finding content”) to opportunity language (e.g. “discovery”).
  • Scoring: for each of 10–15 opportunity areas, score on 4 criteria (T-shirt sizing: high/medium/low is fine). Debate scores.
  • Selection: add scores, sort; top 3 = strategic pillars. Remaining 7–12 = explicit non-focus areas.
  • How-might-we generation: 2–3 HMWs per pillar (1 hour).

Day 3: Winning aspiration via newspaper headline exercise (parallel generation → word cloud → converge → draft statement).

Stage 3: Design Sprint (1 week) Design lead runs a design sprint using strategic pillars + HMWs as inputs. Generates illustrative concept designs per pillar. Optional: concept testing with users via Google Ventures design sprint methodology. Output: mocks/concepts per pillar to embed in strategy doc.

Stage 4: Document Writing (1–2 weeks) Solo PM work. Sections: (1) broader context (leadership expectations), (2) key insights (user, behavioural, competitive), (3) strategic pillars with why and how, (4) winning aspiration, (5) illustrative concepts embedded per pillar, (6) alignment questions. Appendix: scoring table from day 2, additional concepts. Length: 3–4 pages + appendix. No roadmap in the strategy doc.

Stage 5: Rollout (2–3 weeks)

  • Pre-flight with gatekeepers (2–3 people, 1-on-1).
  • Key stakeholders (async or group review).
  • Team roadshows: 8–10 people per session for conversational Q&A.
  • Role of PM: defend strategic pillars using the scoring framework; update for clarifications; resist changing pillars unless the scoring logic is genuinely challenged.

Why the process works (three reasons)

  1. Alignment built in: the doc comes from the working group, not from the PM alone; leadership inputs are baked in from day 1.
  2. Better results: more minds produce better problem articulation and stronger pillars.
  3. Defensible criteria: if challenged, point to the scoring table in the appendix.

Small S vs Big S

DimensionSmall SBig S
OrientationPresent-forwardFuture-backward
Time horizon1–2 years5–10 years
Driven byProblems / gapsAspirations / trends
Led byPMDesign + UXR
Process length8–12 weeksUp to 6 months
Prototype typeIllustrative mocksConcept cars (never shipped)

Both feed a single roadmap simultaneously (“building a bridge from both sides of the river”). [?] Janakiraman claims VRChat runs both in parallel.

Zynga case study

Three de-facto strategic pillars visible across all games: (1) viral game loops (social graph tightly integrated into gameplay mechanics); (2) pay-to-complete (not pay-to-skip — pay to fill the last 20–30% of a progression task, creating elastic spend tiers); (3) cross-game network (all games cross-promoted at top; the network > any individual game). Explicit non-focus: high-fidelity graphics, complex game mechanics. Strategy was embedded in operations (Central Product Management function, data infrastructure). Worked until the mobile platform shift broke resonance.

Meta case study

Parallel strategies for Portal and Oculus Quest II using identical process → nearly identical pillars → completely different outcomes 18 months later. Oculus: highly successful, graduated to VR division. Portal: did not move needle, sunset. Lesson: strategy has no intrinsic business value — only as good as the results it produces.

AI and strategy formulation

Near-term: use AI for (1) competitive analysis (release note themes, competitor reviews analysis, head-to-head heat maps); (2) mock strategy generation (surprisingly comprehensive; weakness is lack of focus — still requires human down-selection). Medium-term: multi-agent workflows (strategy agent + roadmap agent + engineering agent) iterating on each other. [?] Onboarding optimisation via generative AI variations + advanced experimentation (contextual multi-armed bandits) as a worked example.

Connections