Validating Strategic Assumptions

Strategic intent: Before committing to detailed platform design, pursue Ecosystem-Potential-Platform Fit by identifying the assumptions that — if wrong — would invalidate everything, and designing the cheapest possible experiments to test them.

#Overview

The Boundaryless framework names the goal of this phase: Ecosystem-Potential-Platform Fit. The PSM hypotheses produced in Platform Value Propositions all rest on assumptions. Some are safe (reasonable inferences from existing evidence). Others are bets — beliefs without evidence that the entire strategy depends on. This technique surfaces the leap-of-faith assumptions, prioritizes them by risk, and designs experiments to validate or invalidate them.

The output is not just data but decisions: which hypotheses survive, which need more evidence, which are killed.

#When to use it

• After Platform Value Propositions has produced a set of testable hypotheses
• Before investing in MVP development
• Anytime a hypothesis cannot be confidently grounded in existing evidence
• After any major pivot or scope change

#Composition

1. #1. Surface assumptions

   For each value-proposition hypothesis from the PSM, list every assumption it depends on. Use the form: "For this hypothesis to be true, X must be true."

   Duration: 2 hours

   Probe across these categories:

   • Demand assumptions — does this entity actually want this outcome? (Jobs to be Done framing helps here)
   • Supply assumptions — can the other entity actually deliver?
   • Behavioral assumptions — will entities switch from current alternatives?
   • Economic assumptions — at what price/cost is this viable?
   • Technical assumptions — can the platform actually mediate this?

2. #2. Prioritize by risk × impact

   Plot each assumption on a 2×2: evidence we have (low → high) vs impact if wrong (low → high). Focus on the low evidence + high impact quadrant.

   Duration: 1 hour

   The top 3–5 assumptions are your leap-of-faith assumptions. Everything else can wait.

3. #3. Design Minimum Viable Platform experiments

   For each leap-of-faith assumption, design the cheapest experiment that could falsify it.

   Canvas: MVP Canvas · Duration: 2–3 hours

   Common experiment shapes Boundaryless practitioners use:

   • Existing-Experience Scan — analyze what users already do (no intervention) using the Existing Experience Ecosystem Scan
   • Concierge MVP — manual delivery to test demand without building tech
   • Smoke test — landing page or pre-order to gauge interest
   • Wizard of Oz — fake automation while you observe behavior
   • Painted-door test — placeholder feature to measure click-through

   Choose based on which assumption you're testing.

4. #4. Run, learn, decide

   Execute the experiments. Document evidence. For each hypothesis, decide: validated (proceed), invalidated (kill or pivot), or inconclusive (more evidence needed).

   Duration: variable (days to weeks for field validation)

   The result is a refined strategic brief: hypotheses validated/invalidated/refined, ready for the next pipeline.

#Inputs

• Required: value-proposition hypotheses (the PSM) from the previous technique
• Required: access to potential users for interviews / smoke tests
• Recommended: budget and timeline for experiments (typically 2–6 weeks)

#Outputs

• Assumption map — every assumption underlying each hypothesis, classified by evidence and impact
• List of leap-of-faith assumptions — top 3–5 to test
• Experiment designs — one per leap-of-faith assumption, with hypothesis, method, success criteria, disconfirming evidence
• Strategic brief update — hypotheses validated / invalidated / refined
• A clear decision to proceed to platform design, pivot, or kill

#Process heuristics

• Cheaper is better — if the experiment costs more than building the feature, it's not an experiment
• Time-box explicitly — "we'll know in 2 weeks" prevents indefinite drifting
• Define disconfirming evidence in advance — "if fewer than X% of users show interest, we kill it"
• Run experiments in parallel when assumptions are independent
• Don't validate everything — only the leap-of-faith ones
• Existing-Experience Scans are underutilized — observing current behavior often gives more validation than asking users hypothetical questions

#Validation criteria

• Each value-proposition hypothesis has its assumptions surfaced
• Top 3–5 leap-of-faith assumptions identified
• Each leap-of-faith has a designed experiment with clear success/failure criteria
• Experiments are time-boxed
• Disconfirming evidence is specified before running
• At the end, a clear decision per hypothesis (validated / invalidated / inconclusive)

#Common mistakes

• Validating safe assumptions — wasted effort on things you already know
• Hypotheses that can't be falsified — "users will love it" isn't testable
• Sunk-cost commitment — running experiments to confirm what you've already built
• Over-elaborate experiments — a Concierge MVP for 5 users beats a fully built MVP for 0
• Skipping this technique — leap-of-faith assumptions don't go away because you ignored them

#Used in pipelines

• Understanding Ecosystems — as Phase 4 (the validation gate)

#Connections

• Requires: Platform Value Propositions
• Feeds: Designing Platform Experience — only validated hypotheses move forward
• Uses outputs of: Existing Experience Ecosystem Scan when running observational experiments

#Related reading

• The PDT Platform Opportunity Exploration Guide — section on validation in Legacy PDT Exploration
• Discover Platform Opportunities with the Exploration Guide — the broader framing of the exploration → validation flow