Notes — Gibson Biddle on Product Strategy

Four questions [Adler frame]

Q1 — What is this about? A masterclass in product strategy and prioritisation, grounded in case studies from Netflix (2005–2010). The central claim: product strategy is a portfolio of high-level hypotheses about how to delight customers in hard-to-copy, margin-enhancing ways (the DHM model). Most hypotheses fail. The job is to run them at sufficient volume and isolate the ones that work.

Q2 — How is it argued? Primarily through interactive mini case studies — Gibson poses a Netflix decision (Perfect New Release, auto-cancel inactive members, ad-supported plan, account sharing), walks through the DHM analysis with Lenny as a Socratic proxy, and shows where the framework leads. Secondary frameworks (JAM, SWAG, two-way doors) are illustrated similarly. The teaching is inductive: cases first, abstraction second.

Q3 — Is it true? The DHM model is well-grounded in strategy theory (competitive moats, Porter). The empirical cases are credible — Netflix data supports them (e.g., Perfect New Release only improved retention by 0.05%, making $1M against $5M inventory cost). The JAM force-ranking is practical but understates the complexity of actually measuring engagement, which Gibson acknowledges. SWAG is sound as a behaviour nudge but not a methodology. The “career as product” framing is an analogy, not a predictive model.

Q4 — What of it? The DHM model is the most practically actionable strategic framing encountered so far in this wiki. Its strength is the tri-partite test: every proposed strategy can be assessed across delight, defensibility, and business impact simultaneously. The JAM force-ranking resolves the most common leadership misalignment (growth vs. product quality vs. revenue) in a concrete, meeting-ready format. Both frameworks should be added to the wiki’s concept pages.

Glossary

DHM model — Delight customers in Hard-to-copy, Margin-enhancing ways. Gibson’s tri-partite product strategy test, learned from Reed Hastings at Netflix. Any proposed strategy is a hypothesis about how to satisfy all three simultaneously. [§ DHM model]

Hard-to-copy advantage — A structural benefit that competitors cannot replicate quickly. Netflix examples: personalisation (taste graph of 1B+ profiles), brand trust, original content, economies of scale. Contrasted with easy-to-copy tactics like “put a happy family on the non-member sign-up page.” [§ DHM model]

Margin-enhancing — A business impact test: does this make the unit economics better? At Netflix, personalisation enables right-sizing content investment (more people watching = willing to invest more). [§ DHM model]

Word-of-mouth factor — A multiplier on retained-customer value to estimate total benefit of an improvement. Netflix used 2X (each retained customer brings one more); Amazon used 10X. The disagreement over this multiplier was a real argument between Gibson and CFO Barry McCarthy. [§ Perfect New Release]

JAM — Growth, Jobs-to-be-done (engagement), and Monetisation. Gibson’s term for the three competing prioritisation drivers in any company; force-ranking these is the primary tool for alignment. Actually the letters stand for the three words but the J is colloquially used for the engagement/quality axis. [§ JAM model]

SWAG — Stupid Wild-Ass Guess. Gibson’s term for a rapid provisional hypothesis, intended to break the paralysis of “waiting until you have the perfect answer.” The SWAG is shared one-to-one for rapid refinement before the first group presentation. [§ SWAG methodology]

Two-way door decision — Amazon’s term (Gibson applies it): a reversible decision. Low stakes when reversible + small relative to revenue. High stakes when irreversible or large. Most product decisions are two-way doors, but PMs treat them as one-way doors. [§ Decision calibration]

Personal board of directors — Peers + senior mentors deliberately cultivated over a career; used to sanity-check major decisions (company joining, strategy, career moves). Rule: do not ask someone to be your mentor — find ways to be useful to them first. [§ Personal board of directors]

Engagement metric — Gibson uses retention rate as the primary product quality proxy at Netflix (monthly churn went from ~10% at launch → 4.5% in 2005 → 2% at time of episode). A more engaging product reduces churn. The hardest part of the JAM model is choosing and agreeing on this metric. [§ JAM model]

DHM model

The core strategic framework. Gibson learned the phrase from Reed Hastings in a reference check Hastings ran before hiring Gibson: “Can Gib delight customers?”

Three legs:

Why hard-to-copy is load-bearing: the “happy family on the non-member page” example. It delighted customers and improved sign-up conversion. But it was easy to copy — Blockbuster could do it in a week. Easy-to-copy advantages are temporary; hard-to-copy advantages compound.

The delight–margin trade-off: if Netflix dropped its price from $20 to $5/month with identical service, customers would be delighted but the business fails. DHM demands both simultaneously. The word-of-mouth factor is one lever for resolving tension — higher WOM means more delight investment is financially justified.

The portfolio nature of strategy: Gibson expected ~10 strategic hypotheses with ~6 failures. Netflix strategies that worked: personalisation, streaming, original content. Failed: social movie recommendations from friends, Xbox Party watch-together. The model does not guarantee success; it frames the hypothesis correctly so failures are informative.

Perfect New Release Test

A/B test at Netflix ~2005: 10,000 members received their new-release DVD the next day (test); the rest waited ~2 weeks (control). Hypothesis: faster delivery will dramatically reduce churn.

Result: churn dropped from 4.5% to 4.45%. Saved ~5,000 customers/month if rolled out to 1M subscribers.

The DHM maths:

• Delight value: 5,000 customers × $100 LTV × 2X word-of-mouth = $1M
• Marginal inventory cost: $5M

Decision: do not roll out. The delight is real but insufficient to justify the margin cost at Netflix’s word-of-mouth factor.

Key lesson: customer research (surveys, focus groups) said “we want our new releases faster.” The A/B test revealed the retention impact was negligible. Never substitute stated preference for observed behaviour.

The word-of-mouth factor argument: Amazon would use a 10X multiplier, making the maths flip to $10M value vs. $5M cost. The choice of this number is both empirically hard to isolate and politically loaded. Gibson’s CFO was “apoplectic” about using even 2X, fearing it would justify excessive spending on product.

JAM model: force-ranking Growth, Engagement, Monetisation

The #1 source of misalignment Gibson observes in startups: the CEO wants growth; the CFO wants monetisation; neither has made an explicit trade-off between the two with the product leader.

The three axes:

• Growth: year-over-year customer growth rate. The clearest metric, easiest to agree on.
• Engagement: product quality, measured as retention/churn. The hardest one — choosing the right engagement metric requires genuine internal agreement.
• Monetisation: revenue per customer; price; expansion revenue. Often de-prioritised early, then urgently re-prioritised.

How to use it:

1. Take a SWAG on the right order (don’t wait for perfect data).
2. Get leaders into a room; ask them to force-rank the three.
3. If they disagree, resolve before proceeding. Unresolved disagreement is a structural risk.
4. Return to it when external context changes (COVID, earnings miss, funding).

Chegg case: CEO Dan said growth > engagement > monetisation. CFO Greg said monetisation > engagement > growth. After failed negotiation, Greg left. Not a personality conflict — a fundamental misalignment on what kind of company they were building.

Note: at mature companies, the order typically stabilises. At early-stage companies, it legitimately fluctuates with strategic context.

SWAG methodology

Stupid Wild-Ass Guess. The behaviour it targets: product leaders waiting for complete information before sharing a point of view, then consulting firms filling the vacuum.

The SWAG practice:

1. Take a rapid personal position on the question (strategy, prioritisation, metric definition).
2. Share it one-to-one with the smartest person who has more context. “This is my best thinking — what am I missing?”
3. Iterate through 4-6 one-to-one conversations over 1-2 weeks.
4. Present the refined version to a broader group.

This produces better group sessions (all stakeholders have already been consulted), faster decisions, and better SWAGs (the originator learns quickly what they missed). The approach respects that most people in the room have more context on their own area than the SWAG author does.

Gibson’s practice at Chegg: spent first 2 weeks producing a product strategy SWAG, then iterated one-to-one with key stakeholders before presenting to the company.

Decision calibration: magnitude + reversibility

Two simple filters for categorising decision stakes:

1. Magnitude: how large is the downside relative to total revenue? The auto-cancel decision lost Netflix $100M against $30B in revenue (0.3%). Not a high-stakes financial decision.
2. Reversibility: can this be undone? Amazon’s framing: two-way door (reversible) vs. one-way door (irreversible). Most product decisions are two-way doors but are treated as one-way doors.

Why this matters: PMs habitually over-index on stakes, turning two-way doors into endless debates. Explicitly asking “what’s the magnitude?” and “is this reversible?” often allows a team to take action they were paralysed on.

The auto-cancel case: delight ✓ (feels honest and fair), hard-to-copy ✓ (brand trust differentiator), margin ✗ ($100M loss). Low magnitude (0.3% of revenue) + reversible → run it.

Personal board of directors

Gibson’s most-valued career practice, discovered accidentally after a CEO told him to “build a community of peers” when he was promoted to VP.

Structure:

• Peers: former colleagues, LinkedIn connections at the same level. Easy to keep up. Good for “comparing notes” on specific tactical questions (mobile vs. desktop investment; pricing; team structure).
• Mentors (senior): people who can “see around corners.” Harder to cultivate. Rule: never ask someone to be your mentor — find ways to be useful to them first.

How to be useful: everyone needs something. Find what that is and provide it without expectation of return. Gibson’s example: a data person who wanted to break into product built Gibson’s baby website (Squarespace) in exchange for career introductions.

What the board is used for: company evaluation (“should I join this startup?”), strategy sanity checks, career direction questions. Gibson credits his personal board with correctly identifying EA, Netflix, and Chegg as good companies to join at the right time.

Career as product

Gibson’s meta-framework for career management, applied to his own arc:

1. Hypothesis: I might enjoy teaching outside formal employment.
2. Experiment 1: Stanford engineering entrepreneurship course. Liked but not loved — location constraint.
3. Experiment 2: Talks and workshops globally, virtual format. Loved it — flexibility, immediate feedback loops.
4. Iterate: continued refining via SWAG + feedback (Survey Monkey on every talk).

The parallel to product: define a hypothesis → run a minimum-viable experiment → measure engagement (retention, NPS, survey scores) → decide whether to invest or pivot.

Deb Liu’s framing [§ see also Deb Liu on Career and Product]: “PM your career — spec with milestones and success metrics.” Gibson’s version is more experimentation-oriented: hypotheses and pivots rather than a fixed spec. The two framings are complementary — Liu provides the planning frame, Biddle provides the learning loop.

See also

• Gibson Biddle on Product Strategy
• DHM Model
• JAM Model
• Deb Liu on Career and Product
• April Dunford on Product Positioning
• Bob Moesta on Jobs to Be Done