The Product Ecosystem Trade-off Map
Map every actor, incentive, and interaction before choosing a product trade-off.
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 99%
Replace the isolated-user lens with a map of the whole product ecosystem. Define every relevant actor: user cohorts, the company, suppliers, competitors, complementors, and infrastructure owners. For each, record incentives, constraints, and the behavior the product is trying to encourage. Then trace interactions. A recommendation shown to one Facebook user may slightly worsen that person's experience while helping a new user reach enough connections to retain; algorithmic ranking may frustrate power users while rescuing the newcomer experience. Compare options by who gains, who loses, and which strategic or second-order effects follow. The output is not a frictionless answer but a rational, explicit trade-off. This systems view is especially useful at scale, where Jobs to Be Done can become too idealized because no single person's job captures the incentives of everyone the product affects.
Origin
Krishnan offered systems thinking, learned from early Facebook, as a practical alternative to relying on Jobs to Be Done for complex products with multiple actors.
Core principles
- 01A product serves interacting actors, not one isolated customer equation.
- 02Improving one participant's outcome may rationally worsen another's.
- 03Company, competitor, supplier, and platform incentives belong in the product decision.
- 04A useful framework makes trade-offs discussable rather than pretending they disappear.
How to run it
- 1
Bound the ecosystem
Define the decision and the surrounding system that can materially affect or be affected by it. Include actors outside the immediate product screen.
Pro tip Start wider than the end user, then remove actors only when their influence is demonstrably negligible.
Watch out A boundary drawn around one user can hide the exact trade-off the team needs to make.
- 2
Enumerate the actors
List user cohorts, the company, partners, competitors, suppliers, and platforms with a stake in the outcome. Split groups whose incentives differ.
Pro tip Separate power users from newcomers when their product needs and switching costs are different.
Watch out Treating all users as one actor averages away meaningful conflict.
- 3
Record incentives and constraints
For each actor, state what it wants, what behavior the product seeks, and what limits the actor's choices. Include economic and strategic incentives, not only user utility.
Pro tip Write incentives in observable terms such as retain, protect data, earn margin, or avoid switching cost.
Watch out Do not assume that stated preferences reveal the full incentive.
- 4
Trace the interactions
Map how changing one actor's experience alters another's behavior or outcome. Include feedback loops, competitive responses, and second-order effects.
Pro tip Use arrows to make who helps or harms whom visible.
Watch out A locally positive feature can create a negative system-level response.
- 5
Compare explicit trade-offs
For each option, name the winners, losers, strategic benefit, and cost. Debate whether the total system effect justifies making one experience worse to improve another.
Pro tip State the harm plainly instead of hiding it inside an aggregate metric.
Watch out The map supports judgment; it does not make value choices automatically.
- 6
Choose and monitor
Select the trade-off, document the reasoning, and define signals for the predicted interactions. Revisit the map when incentives or market structure change.
Pro tip Track affected cohorts separately so the aggregate outcome does not conceal damage.
Watch out A system map becomes stale when competitors, supply chains, or user composition shift.
In the wild
Facebook believed a new user who reached ten friends in fourteen days was likely to keep using the product. People You May Know therefore showed existing users that newcomer, slightly worsening some recommendations for established users to improve the newcomer's chance of retaining. No single-user job explains the choice; the interaction between user cohorts does.
→ The decision becomes an explicit exchange between established-user relevance and newcomer network formation.
Twitter's algorithmic ranking angered power users who already knew how to control their timelines. Krishnan said the change was built for regular new users who otherwise struggled to get a good experience, while power users were already retained.
→ The company accepted dissatisfaction in an established cohort to improve the product for a strategically important one.
Common mistakes
Optimizing one user in isolation
A choice that fulfills one user's job can damage another cohort or ignore the company and ecosystem incentives needed for the product to survive.
Hiding the harmed cohort
Aggregate improvement can obscure that the decision deliberately worsens a specific group's experience.
Treating the map as permanent
Supply chains, competitors, interfaces, and margins change, so a previously rational interaction may no longer hold.
Is it for you?
Best for
Social networks, marketplaces, platforms, and mature products whose choices affect different user cohorts, partners, suppliers, or competitors.
Not ideal for
Simple early prototypes with one dominant user and few meaningful externalities, where a narrow user-need hypothesis may be sufficient.
From the episode
Hot takes and techno-optimism from tech’s top power couple
Sriram and Aarthi