Brief
The essential idea
Product value is not created by shortening every flow. It is created when a product helps a person make a better decision in a real context. Jobs To Be Done moves the conversation from a feature backlog to the job the user is trying to complete, and from output metrics to outcomes such as lower error rates, safer choices, and durable usefulness.
Friction is therefore a design variable rather than an unconditional defect. Frequent, safe, reversible actions should usually be streamlined, while expensive or irreversible actions may need confirmation, consequence previews, clear warnings, or meaningful alternatives. Accelerating the wrong action only scales the wrong choice faster.
Product managers, analysts, designers, and engineers jointly own decision quality. For a Tech Lead or Staff+ engineer, that means building feature flags, cohort rollouts, rollback paths, experiment infrastructure, and observability for cancellations, reversals, repeated attempts, time-to-decide, and other signals beyond conversion.
Decision lens
Key takeaways
JTBD starts with the user's job and context, not with a proposed feature.
Outcome metrics must include decision quality and the cost of mistakes, not only clicks or conversion.
Deliberate friction is appropriate for costly, risky, or irreversible decisions.
Useful friction explains consequences; harmful friction adds steps without meaning.
Cross-functional teams need one experiment plan and a shared decision review based on evidence.
Engineering enables safe product learning through observability, staged rollout, and fast rollback.
Workplace experiment
Apply it at work
- 1
Choose one high-risk user journey and write a JTBD brief with context, success criteria, and the cost of error.
- 2
Mark where the journey should be faster and where an explicit confirmation or consequence preview is justified.
- 3
Instrument decision-quality signals and run a cohort or A/B experiment rather than measuring completion speed alone.
- 4
Review short-term conversion together with cancellations, reversals, repeat attempts, and downstream effects before scaling.
Choose one action, define the observable effect, and keep the first test small enough to reverse.
Evidence
Sources and further reading
Additional sources
Channel, aggregator, and commentary links confirm the work; they are not the primary source.