This playbook introduces the original TRACE method: Trigger, Reality of the workflow, Alternatives, Constraints and Evidence. TRACE turns an ICP from a demographic portrait into a testable account-selection rule. It then connects that rule to a comparative position buyers can understand. Qualitative research can reveal mechanisms and improve hypotheses; a deliberately recruited interview sample does not establish market prevalence or statistically validate a segment. Begin with a decision that could change, and finish with a claim you can support, qualify or remove.
1. Frame the research decision
Research should resolve a choice rather than fill a persona template. Distinguish segment selection, buyer-role selection and message selection; they require different evidence.
Execution and evidence
Write the current recommendation, strongest alternative and consequence of being wrong. Define the least expensive research that could change the recommendation. Set a recruitment budget and two-week review.
Diagnostic and decision rule
If no plausible finding would change the plan, the exercise is advocacy rather than discovery.
Pause recruitment until the decision owner agrees what evidence would alter investment.
Applied working example
Use a decision contrast such as operations-led versus finance-led buyers. Specify which workflow evidence would favor each route. If the answer could change only a headline but not qualification or resource allocation, narrow the claimed research purpose accordingly. Small wording studies are legitimate, but they should not be presented as validation of the entire customer profile.
Working decision
Which selection rule or claim is genuinely open to change?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Research lead and executive sponsor.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
2. Create a sampling matrix
Recruit for variation in buying situation, not convenient enthusiasm. Recent wins, recent losses, stalled evaluations and noncustomers using alternatives can expose different mechanisms.
Execution and evidence
Build cells by situation and decision status. Recruit beyond the sales champion, including users, implementers and budget owners. Track refusals and recruitment channel to expose selection bias.
Diagnostic and decision rule
A sample consisting only of successful customers can explain satisfaction while missing reasons others never bought. A single company with many interviewees remains one buying context.
Keep missing cells visible. Do not call a convenient sample representative of the market.
Applied working example
Maintain a recruitment disposition log. A refusal because someone lacks time differs from a refusal because the workflow never occurs. Record both without inventing a reason when none is given. Review whether the final sample overrepresents accounts reached through a founder’s network, and seek a contrasting recruitment route before making the selection rule operational.
Working decision
Which important buying situation is absent from our sample?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Customer research.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
3. Recruit with neutral language
The invitation can bias the conversation before it begins. Recruiting people who love innovation invites self-selection that may have little to do with the job under study.
Execution and evidence
Describe the workflow and time commitment without promising a product demonstration. State any incentive consistently, clarify recording preferences and explain how notes will be handled under your approved research process.
Diagnostic and decision rule
If participants expect a sales pitch, they may offer polite future interest instead of a detailed account of past behavior. Reset expectations before continuing.
Do not pressure participation or condition compensation on positive feedback. Replace an unsuitable participant rather than forcing the planned narrative.
Applied working example
Make the invitation concrete: a conversation about the last time the team coordinated a month-end process, not a discussion about digital transformation. Offer a separate sales conversation only if the participant asks for it. This protects the research context and helps the interviewer resist converting every expression of interest into a product pitch.
Working decision
Can a skeptical nonbuyer reasonably accept this invitation?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Research operations.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
4. Interview the last concrete event
Past episodes reveal sequence, dependencies and trade-offs. Strategyzer recommends grounding customer jobs in concrete situations. Questions about hypothetical purchase interest support weaker conclusions.
Execution and evidence
Ask: When did this last happen? What started it? Show the old process. Who became involved? What did you try first? What happened when nothing changed? Follow the timeline before asking for evaluation.
Diagnostic and decision rule
When answers become abstract, return to a date, document or person. Do not infer frequency from vividness: a memorable incident might be rare.
Exclude unsupported episode details from the evidence record. Keep your interpretation in a separate field.
Applied working example
Use a critical-incident reconstruction: ask what document opened the process, who edited it, what was sent and which response followed. The artifact need not leave the customer’s environment. A verbal walkthrough can still reveal sequence and constraints. Respect confidentiality and record when the artifact was described rather than directly observed.
Working decision
What actually happened before a search for a solution began?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Interviewer.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
5. Map the buying group
A user experiences friction, a manager authorizes process change and another function may control budget or implementation. Titles alone do not establish decision rights.
Execution and evidence
For each event map initiator, user, champion, economic approver, implementation owner and blocker. Ask participants which person could stop the purchase and what evidence that person required.
Diagnostic and decision rule
Contradictory accounts may reflect different responsibilities rather than dishonesty. Interview the missing role instead of averaging incompatible answers.
Do not position an operational benefit as a budget-owner outcome without checking the link between them.
Applied working example
Ask each role what would make the project unacceptable. A user may fear extra data entry while an approver fears unpredictable cost. These constraints can coexist with strong interest in the outcome. Positioning should connect their concerns through the actual implementation mechanism, not imply that one enthusiastic champion speaks for the whole buying group.
Working decision
Who owns the cost, outcome and risk of changing?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product marketing and sales.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
6. Study alternatives as behavior
The competitive set includes what the customer does today. Spreadsheets, shared inboxes, internal developers and postponed work can be more relevant than named software vendors.
Execution and evidence
Capture each alternative actually tried, its implementation effort, switching cost and reason it remained or was rejected. Separate a vendor mentioned in passing from one that reached serious evaluation.
Diagnostic and decision rule
A competitor frequently named by your sales team may be absent from real buying histories. Do not build differentiation around an alternative buyers do not recognize.
Retain alternatives supported by observed evaluation or workflow behavior; label untested competitive beliefs as hypotheses.
Applied working example
Create an alternative history for each account: used, considered, rejected and never considered. Those categories prevent a competitor mentioned for completeness from acquiring the same significance as a purchased incumbent. Include the cost of staying with the current process, but separate measurable spend from your estimate of lost opportunity or inconvenience.
Working decision
What does the customer give up by switching from the current approach?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Competitive research.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
7. Code without erasing context
Coding helps compare cases, but categories can silently import the researcher’s conclusion. Keep source excerpts short and preserve the context that changes their meaning.
Execution and evidence
Start with descriptive codes for trigger, workaround, consequence, constraint and action. Have a second reviewer independently code several records, discuss disagreements and revise definitions before applying them broadly.
Diagnostic and decision rule
If every quote fits a desired benefit, inspect the codebook for leading categories. Count independent accounts, not quotation fragments.
Do not report percentages from a purposive sample as population estimates. Report coverage and contradictions instead.
Applied working example
Keep a codebook revision history. If “urgency” initially means any complaint and later means an active deadline with an owner, recode earlier cases before comparing patterns. Preserve a short explanation for the change. Otherwise the same label may contain incompatible evidence and create the appearance of a trend that comes entirely from researcher behavior.
Working decision
Can another reviewer reproduce the meaning of this code?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Research analyst.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
8. Construct the ICP as a rule
A useful ICP predicts suitability for a specific offer and route. It includes problem conditions, implementation capacity, observable trigger and disqualifiers. Firmographics support these conditions rather than replace them.
Execution and evidence
Draft: include an account when [situation], [trigger] and [ability to implement] are present; exclude when [constraint]. Add observable indicators a seller can verify without privileged access.
Diagnostic and decision rule
If the rule accepts every account in a broad industry, it cannot focus scarce resources. If it accepts only existing customers, it may overfit historical relationships.
Test the rule on fresh named accounts before publishing it as the default qualification standard.
Applied working example
Test qualification in a blind exercise. Hide whether accounts became customers, then apply the proposed rule using only information available before selling. Reveal outcomes afterwards and inspect false positives and false negatives. This does not statistically validate prediction, but it exposes hindsight-dependent criteria that a seller could never have used at the relevant moment.
Working decision
What can a seller observe before investing a discovery call?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Sales operations.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
9. Build the value evidence chain
Connect task, operational consequence and business implication without skipping steps. The Value Proposition Canvas is a mapping tool; completing it does not prove demand.
Execution and evidence
For each proposed benefit write the current workflow, observable friction, mechanism your product changes and evidence of the resulting effect. Mark estimated time savings as estimates until measured.
Diagnostic and decision rule
A feature can be real while the claimed business outcome is unproven. Separate product capability evidence from customer outcome evidence.
Remove absolute outcome promises when the link depends on customer implementation or unmeasured downstream factors.
Applied working example
Build a claim ladder with three levels: capability exists, task effect observed, business impact supported. Link each message to its highest defensible level. A task measurement in one configuration can support a bounded example, while a broader financial claim needs additional assumptions. Publish those assumptions rather than silently stretching the original observation.
Working decision
Which step in our benefit chain is still an assumption?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product marketing and product.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
10. Write a comparative position
Positioning selects the situation where a particular mechanism matters. It does not require inventing a new category or calling every capability unique.
Execution and evidence
Write a full internal statement: for [situation], unlike [alternative], this offer enables [task] through [mechanism], supported by [evidence], with [limit]. Translate it into a short external explanation without removing the limit.
Diagnostic and decision rule
If the difference consists only of adjectives such as seamless or innovative, ask for an observable behavior the alternative cannot deliver equally in this situation.
Publish only differences that can survive a product demonstration and an honest alternative comparison.
Applied working example
Maintain an internal full statement even when the external headline is short. It is the reference for sales answers, product-page details and demos. When an editor removes a qualification for brevity, check whether the resulting sentence changes meaning. If so, retain the limit nearby rather than allowing a clean headline to create an unsupported promise.
Working decision
Which mechanism explains why this offer fits this situation?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product marketing.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
11. Test comprehension before preference
A message cannot persuade the intended buyer if it is misunderstood. Preference tests alone can reward familiar language while hiding a wrong mental model.
Execution and evidence
Show one message in a realistic context, then hide it. Ask participants what it does, whom it serves, when they would use it and what evidence they would need. Rotate order when comparing variants.
Diagnostic and decision rule
Confusion about category, setup effort or intended user needs a different repair from weak urgency. Code these separately rather than combining them into a single likeability score.
Revise a message when the intended audience consistently infers a capability you do not provide. Test behavioral response separately.
Applied working example
Avoid teaching during the first comprehension response. Note the participant’s spontaneous interpretation before clarifying. Afterwards, explanations can reveal whether confusion is repairable through a small word change or requires a different category frame. Keep prompted and unprompted answers separate in the report so reviewers do not mistake a coached interpretation for immediate understanding.
Working decision
What did the buyer understand without prompting?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Research and content.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
12. Connect messaging with behavior
After comprehension improves, observe an appropriate next action. A booked qualified conversation, invited colleague or authorized pilot has a different meaning from a page view.
Execution and evidence
Select an action aligned with buying readiness and record the eligible audience, exposure period and response denominator. Keep the product, offer and follow-up consistent enough to interpret differences.
Diagnostic and decision rule
If a message increases clicks but attracts ineligible accounts, success at the surface has weakened sales efficiency. Check downstream fit before promoting the variant.
Do not declare causal lift from sequential cohorts with different audiences or periods. Record the result as directional when controls are weak.
Applied working example
Choose an action proportionate to the audience’s readiness. An early researcher may reasonably request a workflow example rather than a paid pilot. Compare equivalent stages, and avoid penalizing a message because it reaches people earlier in the journey. The commercial question is whether those eligible people advance appropriately, not whether every interaction becomes an immediate purchase.
Working decision
Does the message attract and advance the intended buying situation?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Demand generation and analytics.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
13. Estimate the cost of focus
Segment focus has an economic cost: accounts excluded, research effort, sales capacity and product commitments. It also avoids pursuing accounts that cannot become successful customers.
Execution and evidence
Use account-level assumptions for reachable accounts, eligible share, discovery capacity, potential contribution and implementation hours. Model a downside case with fewer qualified buyers rather than a large unsupported market total.
Diagnostic and decision rule
A high contract value may hide an implementation burden that makes a segment unsuitable for the current team. Track effort per successful activated account.
Do not select a segment on nominal revenue alone when cash, support or integration exposure breaches operating limits.
Applied working example
Treat excluded accounts as a capacity decision, not a moral judgment about their quality. A difficult integration might become viable after a product release. Keep a reason code and review trigger so the team can revisit relevant exclusions without reopening every rejected opportunity. This makes focus explicit and preserves learning for future product decisions.
Working decision
Can this team serve the selected situation profitably enough to continue?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Finance and operations.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
14. Operationalize and refresh the evidence
An ICP should be a shared working rule with an expiry review, not a permanent slide. Positioning must change when product scope, buyer constraints or alternatives change.
Execution and evidence
Days 1–30: recruit, interview and code. Days 31–60: test selection rules and comprehension. Days 61–90: observe behavior and deploy claims with evidence references. Assign sales enablement and ongoing review owners.
Diagnostic and decision rule
Compare newly rejected accounts and newly successful accounts with the rule. Investigate exceptions before broadening the profile to accommodate every request.
Reopen research when exceptions change the buying mechanism, when claims lose evidence or when the product no longer delivers the positioned workflow.
Applied working example
Store the ICP, message, evidence references and approved limits together. When the profile changes, name the lead routing, seller training and page copy that must change with it. A research conclusion has little operational value if the qualification form still sends the old account type to sales. Audit a few fresh handovers after rollout.
Working decision
Who can change the ICP and what triggers a fresh review?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product marketing lead.
- Expected observation before the next review: [fill in]; actual observation: [fill in].
- Decision taken and reason: [continue, change, pause]; smallest next action and its cost: [fill in].
Fictional case: LedgerFlow
LedgerFlow is fictional invoice-coordination SaaS. Its team assumes a research budget of EUR 6,000 and tests two proposed segments. These are invented planning inputs, not interview findings. Segment A has 80 reachable accounts, an assumed 30% eligibility rate and an assumed 20% conversion among eligible accounts: 80 × 0.30 × 0.20 = 4.8 expected customers in a scenario, not a forecast. At EUR 4,000 first-year contribution per customer before acquisition, expected contribution is EUR 19,200. Segment B has 40 reachable accounts, 50% assumed eligibility and 25% assumed conversion: five scenario customers. At EUR 2,500 contribution each, the comparable contribution is EUR 12,500.
The exercise shows why a smaller segment with a high qualification rate is not automatically economically superior. Neither scenario justifies investment until recruitment and behavior test its assumptions. If delivery capacity supports only three implementations, cap both scenarios accordingly. Expected fractional customers are arithmetic expectations, never a promise that a fraction of a customer can be won.
| Research cell | Recruitment intention | Evidence sought | Known bias |
|---|---|---|---|
| Recent buyer | Reconstruct choice | Trigger and alternative | Post-purchase rationalization |
| Lost evaluation | Inspect rejection | Constraint and trade-off | Limited access |
| Stalled account | Explain inaction | Missing trigger or sponsor | Incomplete story |
| Alternative user | Inspect current process | Satisfied and unmet tasks | No product familiarity |
| ICP field | Copyable prompt | Disqualifier example |
|---|---|---|
| Situation | When does the workflow occur? | Workflow absent |
| Trigger | What changed recently? | No active priority |
| Ability | Who can implement this? | Required administrator unavailable |
| Constraint | What prevents purchase or use? | Unsupported mandatory integration |
| Claim | Evidence needed | Permitted wording |
|---|---|---|
| Function exists | Current product demonstration | Supports this workflow |
| Time saving | Comparable task measurement | Observed in this bounded test |
| Business impact | Credible causal chain and baseline | Estimated contribution under stated assumptions |
| Market prevalence | Appropriate representative design | No prevalence claim from interviews |
| Review | Owner | Artifact | Stop condition |
|---|---|---|---|
| Day 30 | Research | Coded cases and gaps | Only friendly customers |
| Day 60 | Product marketing | ICP and comprehension notes | Unsupported differentiation |
| Day 90 | Sales operations | Fresh account behavior | Rule fails on new situations |
Research handoff worksheet
For each final ICP criterion, write the observation that supports it, the strongest contrary case and the sales action it changes. A criterion without an operational consequence may belong in background context rather than qualification. Assign an expiry trigger such as a new integration, a changed buyer role or repeated successful exceptions.
Give the handoff to a seller who did not participate in the study. Ask the seller to classify a fresh account and explain the evidence needed before proceeding. Record where the rule requires unavailable information. Repair that gap with a discovery question or narrower claim, then repeat the exercise. This is a usability check on the research output, not proof that the profile predicts the whole market.
Diagnosing contradictory buyer stories
When two companies evaluate the same capability differently, start with context. Compare task frequency, available alternative, involved roles and implementation requirements. The product may not have received contradictory evaluations at all; two different tasks may have been grouped under one label. Separate those cases before proposing a new segment boundary.
Then classify the action. A positive judgment without a next step means something different from authorizing time or budget. A negative judgment after poor setup may indicate implementation trouble rather than absent demand. Keep these as competing explanations and identify the smallest additional evidence that would distinguish them. It might be a conversation with the administrator or a walkthrough of the existing process.
Only then change the profile or message. Prematurely combining cases into an average removes the mechanism qualitative research should reveal. Preserve the case that still does not fit: it may show an offer boundary or define the next research question.
Primary sources and scope
Sources checked on 24 October 2026. Sources establish concepts or documented capabilities; all frameworks, worksheets, example figures and decision thresholds in this resource are original editorial tools rather than published benchmarks.
INTERACTIVE / FROM READING TO ACTION
Three checkpoints for your implementation
Choose a checkpoint and track your reading progress. Checks are neither saved nor sent and do not score your business.
CHECKPOINT 01
1. Frame the research decision
Research should resolve a choice rather than fill a persona template.
CHECKPOINT 02
2. Create a sampling matrix
Recruit for variation in buying situation, not convenient enthusiasm.
CHECKPOINT 03
3. Recruit with neutral language
The invitation can bias the conversation before it begins.
Frequently asked questions
Who is this resource for?
Product marketers, SaaS founders and customer research and sales leaders. Use the resource with the people accountable for implementation and approval.
What practical outputs can we create?
Recruitment plan with inclusion and exclusion criteria Coded register of concrete buying events ICP card with observed decision criteria Tested positioning with an evidence and claims register
What is included in the PDF?
14 chapters with explanations, decision rules, examples and working materials. 19 designed PDF pages including sources and a transfer canvas.
Are the examples real client results?
No. Cases labelled fictional are used to explain the method. Example figures are assumptions, not industry benchmarks or promised outcomes. Apply the approach to your own evidence.
Are the online edition and download free?
Yes. The complete resource is currently available to read and download without payment or a registration form.
Is there a German edition?
Yes. Use the language switch for the complete German online edition and German PDF.
Sources & context
Primary sources checked 24 October 2026. The working models are editorial decision aids. Fictional examples do not replace your own customer evidence. Verify current platform capabilities before implementation.