Product Management Foundations · Module 4 of 15
Don't turn the first idea into “the problem.”
Strategy has narrowed where the product intends to create value. This chapter teaches how to move from messy requests, behavior, research, complaints and market signals toward a bounded opportunity worth exploring, while keeping alternative interpretations, solution directions and important assumptions visible.
The direct answer
Product opportunities usually begin as signals: requests, behaviors, complaints, research findings, metrics, market changes or ideas. Discovery turns those signals into evidence-backed opportunity hypotheses, keeps alternative interpretations visible, explores more than one solution direction when the decision warrants it, and hands explicit assumptions into validation rather than presenting a preferred solution as proven.
Discovery and delivery are complementary modes of work that can overlap. Teams may learn through research, prototypes, technical spikes, partial implementations, data or shipped behavior. The purpose is not to finish all learning before building; it is to avoid expensive commitment while important assumptions remain invisible.
The chapter question
How do I know whether I have found an opportunity worth exploring rather than merely renamed a feature idea as a problem?
Need the broader operating system for ongoing discovery cadence, recruiting, synthesis, opportunity mapping and governance? Use the Product Discovery course.
You came from
Where will the product focus and what choices constrain the search?
You are here
What opportunity and solution possibilities deserve exploration?
Next
Which material assumptions should be tested, and how?
Preview context
The real first lesson uses “discovery phase” language. In this chapter, read that wording as a distinction between learning and execution modes, not as a rule that discovery must finish before delivery starts or that delivery begins only after certainty.
Real CraftUp lesson
Try the quiz questions taken directly from the published CraftUp lesson for this module.
Quick practiceWant another challenge? Practice another round.
Real CraftUp curriculum · no sign-up required
Core mental model
The transitions matter more than the labels. Each step should make clearer what was observed, what is being inferred, what is being proposed and what must still be true.
Something observed, reported or proposed: a request, complaint, behavior, metric, research finding, competitor move, stakeholder idea or technical possibility.
A signal is an input to investigate. It is not automatically a supported opportunity or roadmap requirement.
A bounded, provisional statement about a person or context, desired progress, friction or constraint, and the evidence that makes it worth exploring.
Keep the word hypothesis mentally attached. The framing may change when new evidence distinguishes competing explanations.
One possible mechanism for addressing the opportunity. Different mechanisms may solve the same underlying problem in materially different ways.
Do not confuse the first plausible solution with the opportunity itself or force an arbitrary number of alternatives.
Something that must be true for the opportunity framing or solution direction to hold: desirability, usability, feasibility, viability, trust, safety or another material constraint.
Assumptions become the handoff to Validation and MVPs. They are not facts because a discovery artifact looks polished.
10× learning asset
Work through one fully hypothetical field-service scheduling platform. Open each reveal only after deciding what the current evidence can justify. This is a learning lab, not a production workspace or a scoring tool.
Dispatchers ask for one-click rescheduling. Technicians say changes arrive late. Support sees questions about overlapping jobs. Some teams export schedules to spreadsheets. Sales asks for WhatsApp notifications. A competitor launches an AI scheduler.
Reveal the reasoning
Observed in this hypothetical case: dispatchers manually reconcile last-minute changes across calls, messages and spreadsheets.
Reveal the reasoning
“Users need better scheduling” sounds reasonable but is too broad to guide exploration. “Users need an AI scheduler” goes further and embeds a preferred solution.
Reveal the reasoning
Frame A: dispatchers cannot quickly identify conflicts when jobs change. Frame B: technicians lack a reliable current view when schedules change. Frame C: customers receive change confirmations too late.
Reveal the reasoning
Bad: “Dispatchers need an AI conflict detector.” Better, but still provisional: “Dispatchers struggle to identify schedule conflicts quickly after same-day changes.”
Reveal the reasoning
Ask whether the opportunity is coherent enough to explore options, or whether a competing interpretation could still materially change what you would build.
Reveal the reasoning
Possible directions include clearer conflict visualization, guided rescheduling, constraint warnings, a confirmation workflow, automated recommendations, or a workflow/process change.
Reveal the reasoning
Evidence, strategic relevance, user consequence, constraints, risk, feasibility, reversibility and learning potential can all matter. Which matter most depends on the decision.
Reveal the reasoning
Suppose the team narrows to proactive conflict warnings. Material assumptions include conflict frequency, whether earlier visibility changes dispatcher behavior, reliable warning generation, user trust and acceptable alert burden.
Reveal the reasoning
The signal is too thin to support a useful opportunity frame.
A plausible bounded opportunity is visible, but important interpretation remains provisional.
The opportunity is coherent enough to compare materially different solution mechanisms.
A solution direction and its material assumptions are explicit enough for the next chapter. Ready does not mean validated.
Worked discovery trace · explicitly hypothetical
The product direction is to help small field-service teams reduce coordination failures around changing jobs. This comes from the hypothetical strategy context, not claimed market research.
Dispatcher feature request: one-click reschedule. Technician report: changes arrive late. Support pattern: overlapping-job questions. Observed workaround: some teams export schedules to spreadsheets. Internal sales hypothesis: WhatsApp notifications. Market signal: a competitor launches an AI scheduler.
It does not prove that all signals share one root cause, that scheduling is the only opportunity, that AI is needed, or that the competitor feature creates customer value.
Conflict visibility may be weak. Current schedule visibility may be weak for technicians. Change confirmation may be late for customers. Approval workflow may be fragmented. Treat these as hypotheses until evidence distinguishes them.
Dispatchers struggle to identify schedule conflicts quickly after same-day changes. The frame is bounded enough to explore, but remains provisional.
Remove “AI conflict detector” from the opportunity statement. AI is one possible mechanism and carries its own reliability, trust and data assumptions.
Consider conflict visualization, guided rescheduling, constraint warnings, a confirmation workflow, automated recommendations, or process changes. The set is illustrative, not a required idea count.
Suppose the team carries proactive conflict warnings forward because the current evidence points to time-sensitive conflict visibility, the mechanism fits strategic direction, the implementation is plausible, and the team can learn from it without an irreversible commitment. This is a rationale, not a scored winner.
Conflicts occur often enough to matter; earlier visibility changes behavior; warnings can be generated reliably enough; dispatchers trust them; alert burden remains acceptable.
Validation now asks which of those assumptions is decision-critical enough to test and what is the least costly credible way to reduce that uncertainty. C08 deliberately stops before prescribing the experiment.
Observation
Keep the source and context attached: observed behavior, customer report, support pattern, stakeholder proposal or market move.
Interpretation
Generate plausible explanations without turning one into fact. Ask what evidence would distinguish interpretations that lead to different product choices.
Assumption
Make the belief explicit enough that the next chapter can decide whether and how to reduce the uncertainty.
Mini practice · signal, opportunity or solution?
Signal / solution request
The request is useful evidence about context, preference, constraint or desired progress, but it does not yet establish the underlying opportunity or what should be built.
Behavioral signal
You observed a pattern. You still need evidence that distinguishes why the drop happens before promoting one explanation to an opportunity.
Potential opportunity hypothesis
This can be an opportunity frame if the struggle is supported by real evidence and bounded to the relevant user/context. Without that provenance, it is still an interpretation.
Solution candidate
It names a mechanism. The team should be able to connect it to a supported opportunity and the assumptions that make the mechanism plausible.
Market / competitor signal
The move may change alternatives or expectations, but it is not proof that your users have the same need or that AI is the right response.
Signal that may support an opportunity hypothesis
With clear provenance, user/context and consequence, this repeated behavior can support a bounded opportunity. The wording still should not manufacture a cause.
Mini practice · too early to converge?
Scenario A
One stakeholder has proposed one feature and there is no relevant user evidence yet.
Need evidence before solution convergence.
The team has a proposal, not yet a supported opportunity. Investigate the context and consequence before treating implementation choice as the main decision.
Scenario B
Repeated supported problem evidence exists, but the team has discussed only one implementation because it was the first idea.
Explore alternatives.
The opportunity may be coherent while the solution space is artificially narrow. Divergence can reveal mechanisms with different cost, risk, usability or learning value.
Scenario C
A reversible, low-risk direction is strongly constrained by technical or legal reality and alternatives add little learning value.
More brainstorming is not automatically valuable.
Divergence is a means, not a quota. Narrowing can be reasonable when the options are understood well enough and the next uncertainty is better resolved through evidence or testing.
Divergent vs convergent thinking
Divergence expands plausible interpretations or mechanisms when one solution dominates too early. Convergence narrows when options are sufficiently distinct, decision criteria are clear enough and further ideation adds little learning. New evidence can reopen divergence later.
Cross-functional discovery benefits from product, design and technical perspectives being represented early, but team shape varies. If you use Teresa Torres' “product trio” concept, treat it as a model for cross-functional discovery rather than a universal org-chart requirement.
Opportunity comparison
Not every lens belongs in every decision, and the lenses should not be automatically averaged into one total score.
Framework boundary
The curriculum introduces scoring, Design Thinking, the Business Model Canvas, Value Proposition Canvas, Kano and Opportunity Solution Trees. Use each only when its structure helps the current uncertainty. Do not treat a completed canvas, workshop or highest score as customer truth.
Teresa Torres developed the Opportunity Solution Tree to make the path from a desired outcome through opportunity and solution spaces to assumption tests visible. C08 teaches why that separation matters; the Opportunity Solution Tree Builder owns the actual editable mapping workflow.
Want to map customer jobs, pains and gains against a value proposition? Use the dedicated Value Proposition Canvas rather than recreating that workspace here.
AI in discovery
AI can suggest clusters, summarize material, generate alternative opportunity framings or solution mechanisms, and surface possible contradictions for review. It can also invent patterns, flatten context, over-weight repeated wording and hide minority signals. AI-generated themes and opportunities remain hypotheses until they are checked against source evidence.
Do not upload identifiable interview transcripts, confidential customer data, internal strategy or sensitive support content to external AI services without appropriate company policy, privacy controls and data-handling safeguards.
Full curriculum
The real lesson order remains unchanged. The phase labels add orientation without pretending that discovery is a rigid sequence or that each lesson title is a universal operating rule.
Are we actually exploring an opportunity, or have we already decided the feature?
What evidence and perspectives should deepen the frame?
When should I expand options, and what evidence or constraints justify narrowing?
Which structure helps the current thinking job, and what can it not decide for me?
What must be true for this opportunity or solution direction to matter?
When should exploration stop and validation begin?
Use the specialist owner when the job changes
Use the Problem Statement Generator to turn actual evidence, provenance, assumptions and unknowns into a bounded problem hypothesis. A supported frame still does not mean roadmap priority.
Use the Opportunity Solution Tree Builder when a credible opportunity should be connected to a desired outcome, alternative opportunities, solution directions and assumptions.
Use Feature Request Triage when the operational job is preserving request provenance, avoiding false merges and moving from raw requests toward canonical problem records.
Use Prioritization when the decision becomes choosing among competing investments under constraints. C08 owns discovery and framing, not resource allocation.
Ongoing feedback cadence belongs to Customer Feedback and Continuous Discovery. Deeper strategic choice under constraints belongs to Strategic Decision-Making for PMs.
Next learning step
You now have a bounded opportunity, plausible solution directions and explicit assumptions. Validation asks which assumptions are risky enough to test and what is the smallest credible way to reduce decision-relevant uncertainty. It does not promise certainty.
Continue to Validation and MVPs →Build your product skills with short, practical lessons you can fit around your day.
Start with our free courses and upgrade anytime.
