Competitive Decision Toolkit · Decision Definition Canvas
Complete the canvas before you collect competitor evidence
A Decision Definition Canvas is a one-page contract for the research job. It says who owns the decision, what action may change, which buyer and offer are in scope, which market and time boundary count, which alternatives are legitimate, which criteria matter, what evidence is good enough, and what future change should reopen the decision.
This is the decision-framing layer before a competitor research brief, competitor-analysis scope, competitive-analysis framework or evidence collection: those downstream activities inherit this frozen decision object rather than redefining it.
If those elements are vague, research can become impressively detailed while answering the wrong question. A strong canvas prevents that drift.
Memorable explanation: before comparing routes, pin the destination, name who is driving, mark the roads that are out of bounds, and decide what evidence would actually make you turn.
Create a usable Decision Contract in two minutes
Finish this sentence without using vague nouns such as “growth,” “SMBs,” “better competitors,” “global” or “strategy” on their own:
Decision ID [ID]. By [decision date], [decision owner] will decide whether to [specific action] for [buyer + material job], within [offer / use-case boundary] and [market / geography / evidence window]. The permitted alternative universe is [alternative types]. Explicit exclusions: [exclusions]. Material criteria: [three to five criteria]. Hard constraints: [constraints]. Evidence threshold: [threshold + custom rule if used]. Evidence may change [actions that may change]. Reopen if [material trigger].
If you cannot complete that sentence: the next task is definition, not competitor research.
One canonical Decision Contract object drives every product surface
The article, browser builder, workbook, PDF and handoff now use the same substantive fields. Different layouts are allowed; dropping a governed field is not.
Controlled-token rule: machine values are exact and case-sensitive across browser JSON, Excel and PDF. Alternative tokens are DIRECT, ADJACENT, SUBSTITUTE / SERVICE, INTERNAL BUILD and STATUS QUO. Human-friendly labels may differ; stored tokens may not.
Identity
Decision ID · owner · decision date
Decision
Specific action under decision
Commercial scope
Buyer + material job · offer/use-case boundary · market/geography/evidence window · explicit exclusions
Choice architecture
Permitted alternative universe · three to five material criteria · hard constraints
Evidence governance
Evidence threshold · custom evidence rule when CUSTOM is selected
Decision consequence
Actions evidence may change · reopen trigger
State
READY TO RESEARCH / REPAIR CANVAS / STOP - RESEARCH NOT JUSTIFIED · governed next action
The canvas has eight blocks because each one prevents a different research failure
Owner + date
Who can commit to action, and by when?
Decision
What specific choice can change?
Buyer + job
Whose problem and buying situation count?
Market boundary
Which geography, segment, channel and evidence window count?
Alternatives
Which direct, adjacent, substitute, internal or status-quo routes are legitimate?
Criteria + constraints
What must an option achieve, and what cannot be violated?
Evidence threshold
What quality is required before a material claim may change the decision?
Action + reopen trigger
What changes after research, and what later change invalidates the current frame?
The canvas does not prove the answer. It defines the question tightly enough that the Evidence Audit, competitor qualification, opportunity analysis and executive memo can all inherit the same decision boundary.
How to complete each field without creating false precision
| Field | What to write | Weak entry | Decision-grade entry | Why it matters |
|---|---|---|---|---|
| Decision owner | One role or named accountable person who can authorize the action. | “Marketing team” | “VP Marketing; founder approves if annual spend exceeds the agreed threshold.” | Research without an owner becomes information without authority. |
| Decision date | The date the choice must be made, not when the research should be “done.” | “Q4” | “Decision required before the 15 October planning lock.” | Timing determines the evidence window and the cost of delay. |
| Action | A verb that can actually change: enter, reposition, prioritize, fund, defer, exclude, test. | “Understand competitors” | “Decide whether to fund a UK professional-services segment entry pilot.” | A topic is not a decision. |
| Buyer + job | The accountable buyer/user and material job or outcome. | “SMBs” | “Operations leaders at 50–250 employee professional-services firms choosing a workflow system for multi-team intake and approval.” | Competitors change when the buyer and job change. |
| Offer boundary | The product/service capability being evaluated, including excluded use cases. | “Automation” | “Packaged intake, approval and audit workflow; exclude generic integration-only tooling unless it can own the buyer outcome.” | Prevents feature overlap from becoming false competitive equivalence. |
| Market / geography | Commercial market, geography, channel and relevant time window. | “Global B2B” | “UK mid-market direct-sales motion; evidence checked for current 2026 offer and buying conditions.” | Market boundaries change pricing, regulation, alternatives and evidence. |
| Alternative universe | The categories that may legitimately win the same decision. | “Top 10 vendors” | “Direct software, adjacent workflow platforms, managed-service substitute, internal build and status quo.” | Prevents search visibility from deciding the competitor set. |
| Decision criteria | Three to five conditions that could change the action. | “Features, price, brand” | “Buyer-job fit; implementation burden; security/procurement fit; switching/exit; commercial model.” | Criteria should precede evidence, not be invented after seeing results. |
| Constraints | Hard limits such as budget, regulation, timing, architecture or policy. | “Must be affordable” | “Pilot must use existing identity architecture and fit the planning-cycle deadline.” | A constraint can exclude an otherwise attractive option. |
| Evidence threshold | The minimum source quality/freshness/state for material claims. | “Use reliable sources” | “Material product, security and contract claims require attributable current evidence; inference stays labelled and cannot silently become observed fact.” | Defines what is allowed to change the decision before research starts. |
| Action that may change | The actual downstream decision affected by the findings. | “Update our strategy” | “Fund pilot / defer pilot / narrow segment / change comparison set.” | If no action can change, the research may not be justified. |
| Reopen trigger | A future event that invalidates the frame. | “Review later” | “Reopen if buyer segment, offer, pricing model, regulatory constraint or implementation architecture changes materially.” | Stops an old canvas from becoming permanent truth. |
Define the alternative universe before naming competitors
A common failure is to list familiar vendors first and then reverse-engineer the scope around them. Instead, define the types of alternative that could win the same buyer decision.
Direct software DIRECT
Same buyer, same material job, similar buying stage and realistic replacement path.
Adjacent platform ADJACENT
Overlaps part of the job or architecture but is not a like-for-like commercial replacement.
Service / substitute SUBSTITUTE / SERVICE
Different delivery model that can still fulfil the material outcome.
Internal build INTERNAL BUILD
Spreadsheet, script, internal workflow or custom build that can absorb the job.
Status quo STATUS QUO
Delay, no purchase, existing process or incumbent arrangement.
Strategic ruleThe canvas defines who is allowed into research; later qualification decides which specific candidates actually belong in each role.
Set criteria before seeing the competitor evidence
Criteria are the bridge between the decision and the research. If a criterion could not change the action, it does not deserve equal space merely because it is easy to measure.
- Start from the buyer outcome. What must an alternative actually accomplish?
- Add adoption reality. What implementation, procurement, security or organizational burden matters?
- Add replacement reality. What switching, exit or status-quo condition changes feasibility?
- Add commercial consequence. Which price/contract or time-to-value condition affects the decision?
- Keep the set bounded. Three to five material criteria are usually more decision-useful than a 40-row feature matrix.
Do not assign arbitrary weights here merely to make the canvas look quantitative. Weighting belongs only where the downstream method genuinely needs it and the decision owner can explain the trade-off.
Set the evidence threshold now; conduct the Evidence Audit next
The Decision Definition Canvas does not audit evidence. It defines what evidence must be true of material claims before those claims may influence the decision.
Attributable
A reviewer can identify where the claim came from.
Current enough
The source is recent enough for the decision and has not been invalidated by a material change.
Role-appropriate
A pricing page can support published price; a directory can support discovery; neither automatically proves buyer substitution.
State-labelled
Observed, inferred and unresolved states stay visibly different.
Conflict-preserving
Contradictory evidence is recorded rather than averaged into false certainty.
Ethically obtainable
The collection method stays within legal, policy and professional boundaries.
Canvas output: a requirement. Next article/output: the Evidence Audit tests whether the evidence packet actually meets it.
Interactive · Nothing is submitted
Build your Decision Definition Canvas
Complete the fields below. The tool uses visible fail-closed rules and produces a portable Decision Contract. It does not score decision quality and does not send your entries to SERPIFAI.
What READY means: the required structure is complete and you have explicitly confirmed the visible quality tests. The tool does not verify that your business assumptions are true; the Evidence Audit and later review steps test the evidence.
Fail-closed parity: unsupported or wrong-case controlled tokens, duplicate criteria, fewer than two distinct alternative types, or a mismatch between canonical criteria and governed detail records cannot produce READY.
Fictional example: turn “analyze competitors” into a decision another person can research
Scenario: a fictional workflow-software company is considering a focused UK segment pilot. The names and conditions below are illustrative; they are not client results or market claims.
| Canvas block | Fictional completed entry | Why it is stronger |
|---|---|---|
| Owner + date | VP Marketing; decision before 15 October planning lock. | One accountable owner and a real decision deadline. |
| Decision | Fund, defer or narrow a UK professional-services segment pilot. | Research can change a finite action. |
| Buyer + job | Operations leaders at 50–250 employee professional-services firms choosing a system for multi-team intake and approval. | Buyer and material job are concrete. |
| Offer boundary | Packaged intake, approval and audit workflow; integration-only tools are adjacent unless they can own the outcome. | Prevents feature overlap from defining competition. |
| Market boundary | UK mid-market direct-sales motion; current 2026 offer and buying conditions. | Geography, segment, motion and evidence window are explicit. |
| Alternatives | Direct software; adjacent workflow platform; managed service; internal build; status quo. | The alternative universe exists before vendor names. |
| Criteria | Buyer-job fit; implementation burden; security/procurement fit. | Material decision criteria are fixed before evidence is seen. |
| Evidence threshold | Current attributable evidence for material product, implementation, security and contract claims; inference remains labelled. | Evidence cannot silently gain authority downstream. |
| Action + reopen | Fund / defer / narrow / change comparison set. Reopen on material buyer, offer, pricing, regulatory or architecture change. | Creates an action and lifecycle, not a static worksheet. |
Six quality tests separate a completed canvas from a filled-in form
1. Same-answer test
Would two reviewers name the same decision from the canvas?
2. Alternative test
Does the canvas contain at least two distinct governed alternative types, including no-action where relevant?
3. Action-change test
Can evidence change what the owner will do?
4. Boundary test
Can a reviewer tell what buyer, offer, market and time period do not count?
5. Evidence test
Is the minimum proof threshold stated before research begins?
6. Reopen test
Is there a material future change that would invalidate the frame?
Criteria distinctness rule: the 3-5 material criteria must be distinct after trimming whitespace and normalizing case. A duplicate criterion must return REPAIR CANVAS, not READY.
Fail closed: if the action is already fixed and evidence cannot change it, stop. Research would be documentation or persuasion, not decision support.
Complete the canvas with a team without turning it into a committee document
- The decision owner drafts first. They state the decision and action before analysts suggest fields.
- One researcher challenges ambiguity. They ask what evidence could actually change the action.
- One implementation owner challenges feasibility. They identify constraints the analysis could otherwise ignore.
- Specialists bound their own domains. Legal, security, finance or technical reviewers define material constraints without taking ownership of the whole decision.
- Freeze the canvas before collection. Changes after research begins require a change-log entry explaining why the frame changed.
If the team cannot agree on the decision itself, that is not a reason to collect more competitor data. Resolve the governance conflict or use the escalation guide first.
Portable handoff: what the next analyst should receive
A strong canvas allows someone else to start the Evidence Audit without reconstructing the assignment from meetings.
Identity: Decision ID, decision owner and decision date.
Decision: the specific action under decision.
Commercial scope: buyer + material job, offer/use-case boundary, market/geography/evidence window and explicit exclusions.
Choice architecture: permitted alternative universe, three to five material criteria and hard constraints.
Evidence governance: evidence threshold plus the exact custom evidence rule when CUSTOM is used.
Decision consequence: actions the evidence may change and the reopen trigger.
State: READY TO RESEARCH, REPAIR CANVAS or STOP - RESEARCH NOT JUSTIFIED, plus the governed next action.
Next step when READY: conduct the Evidence Audit against this frozen object.
The workbook and PDF use this same canonical object. A handoff is share-ready only when the governed state is READY TO RESEARCH and its snapshot controls pass.
Nine mistakes that make a canvas look complete while leaving the research undefined
“Analyze competitors” names an activity, not a choice.
If everyone owns the decision, nobody has final authority.
“SMBs” is not enough to define commercial substitution.
Familiar logos are listed before the alternative universe is defined.
The team chooses dimensions that make a preferred option look stronger.
“Reliable” and “credible” mean nothing unless source role, freshness and state are governed.
Status quo or internal build can be the real alternative even when they do not appear in search results.
The research is being used to justify a predetermined choice.
An old frame quietly survives a new buyer, market, offer or constraint.
Where the Decision Definition Canvas fits in the Toolkit
Questions about completing a Decision Definition Canvas
What is the Decision Definition Canvas for?
It converts a broad competitor-research request into one accountable decision with explicit owner, buyer/job, offer, market, alternatives, criteria, evidence threshold, action and reopen boundary.
How detailed should the canvas be?
Detailed enough that another analyst can start research without asking what the decision means, but short enough to stay reviewable. If a field needs several pages, keep the decision rule in the canvas and link the supporting specification separately.
Should I list competitor names in the canvas?
Usually not first. Define the alternative universe and inclusion logic before familiar names shape the frame. Specific candidates can then be discovered and qualified against that definition.
How many decision criteria should I use?
Use the smallest set that can materially change the action. Three to five criteria are often more decision-useful than a large feature inventory, but there is no universal number.
What if the team cannot agree on the buyer or market?
Do not hide the disagreement inside a broader phrase. Resolve the definition, split the decision into distinct canvases, or use the escalation guide if the disagreement changes authority or strategic direction.
What if new evidence cannot change the decision?
Stop and state the real purpose of the work. If the action is already fixed, competitor research is not serving a decision and should not be presented as if it were.
Does the canvas replace the Evidence Audit?
No. The canvas sets the evidence requirement. The Evidence Audit tests whether the actual evidence packet is attributable, current enough, role-appropriate and sufficiently governed.
When should I update or rerun the canvas?
Reopen it when a material buyer, offer, market, geography, constraint, alternative type, decision criterion or implementation condition changes enough to alter the research question.
Evidence base, competitor-content research and methodological limits
The canvas combines SERPIFAI's competitor-decision workflow with broader decision-framing, market-research and evidence-governance principles. External sources support individual principles; they do not certify the proprietary canvas or prove a commercial outcome.
| Source | Role in this guide | What this guide does not infer |
|---|---|---|
| HM Treasury - The Green Book (2026) | Objectives, constraints, options, business-as-usual, documented reasoning, risk/uncertainty and keeping the decision problem explicit before detailed appraisal. | That government appraisal rules govern private competitor analysis. |
| U.S. SBA - Market research and competitive analysis | Target-market, location, demand, pricing and indirect/secondary competitor context. | That the SBA defines SERPIFAI competitor roles or decision gates. |
| CEBMa - Evidence-Based Decision Making | Using critically evaluated evidence from multiple sources rather than relying on one evidence stream. | That evidence mechanically dictates the business decision. |
| SCIP - Ethical Intelligence | Legal, transparent and honest competitive-intelligence practice and explicit collection boundaries. | Legal advice, endorsement of this guide, or permission to use prohibited collection methods. |
Competitor-content gap: current competitor-analysis templates commonly emphasize competitor lists, features, pricing, positioning, monitoring and generic scope. This guide's distinct task is to freeze the decision contract before those activities so the research cannot quietly redefine the buyer, market, criteria or permitted action.