A go-to-market plan is a sequence of commitments under uncertainty. This guide introduces the original EVIDENCE loop: Establish the decision, Verify the situation, Identify alternatives, Design a test, Evaluate behavior, Narrow investment, Commit owners, Examine retention. It is a working framework, not a validated scoring model. Ninety days may expose a viable route or disprove a tempting one; they cannot prove long-term retention in an annual enterprise contract. Use this guide for one product, one primary segment and one accountable cross-functional team.
1. Write the commitment before the campaign
A revenue goal alone cannot tell a team what to do next. Define the decision that will change: selecting a buyer, authorizing a pilot budget or choosing between a sales-led and assisted product route. Steve Blank’s original customer-development writing places external learning before a repeatable execution model.
Execution and evidence
Create a one-page charter with product scope, decision owner, available cash, available engineering time and a review date. Specify what the team will stop doing to create this capacity. Include existing contractual delivery obligations.
Diagnostic and decision rule
If every department interprets success differently, the charter is unfinished. Ask each lead to describe a successful day 90 without using a marketing activity count.
Do not approve campaign spend until the owner can name the decision, evidence required and maximum affordable loss.
Applied working example
For example, a founder may ask whether to hire a second seller. That decision requires evidence that the first seller’s motion can be repeated, not merely that the website attracts interest. Write the hiring trigger as a combination of eligible pipeline, delivery capacity and an affordable downside case. A busy calendar alone does not meet it.
Working decision
Which commitment can this project responsibly support?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: CEO with GTM 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].
2. Build an evidence ledger
Separate market descriptions from customer observations. A large category may contain few reachable accounts with a current trigger. Interview notes, support tickets and signed purchase records answer different questions. Confidence depends on relevance, independence and recency, not the number of slides.
Execution and evidence
Create records with claim, evidence type, account, date, uncertainty and next test. Tag opinions, observed behavior, commercial commitments and measured outcomes separately. Add negative evidence from stalled or lost opportunities.
Diagnostic and decision rule
Ten comments from one enthusiastic champion are one account perspective. A signed pilot proves a limited commitment, not adoption by the whole segment.
Retain a claim only at the strength its source permits. Reopen it when product scope, buying conditions or segment changes.
Applied working example
Use a ledger identifier in every planning claim. If the statement is “teams change after an audit,” link the actual episodes and the counterexample that did not trigger change. Record whether the event was independently described or suggested by the interviewer. This lets the next reviewer challenge a specific inference instead of debating confidence in general.
Working decision
What could falsify our most expensive assumption?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product marketing with research 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].
3. Choose a reachable beachhead
An ideal customer profile should help reject accounts as well as select them. Combine operational situation, triggering event, workflow complexity and ability to implement. Firm size is a useful filter only when it changes the problem or route to purchase.
Execution and evidence
Map three candidate segments and identify ten named reachable accounts in each. Compare existing relationships, recurring problems and deployment requirements. Record disqualifiers such as unsupported integrations or an unavailable administrator.
Diagnostic and decision rule
A segment that looks attractive only after excluding implementation costs is not yet attractive. Distinguish market size from the number of accounts you can actually engage this quarter.
Select one primary segment for the first experiment. Keep another as a documented alternative; do not mix their response rates.
Applied working example
Test the beachhead against a fresh list before committing. Give a seller five previously unseen accounts and ask which qualify and why. If the answer requires a founder’s private knowledge in every case, turn that knowledge into observable criteria or accept that access is still relationship-dependent. Do not assume the same rule will scale automatically.
Working decision
Which accounts share both a problem and a feasible buying path?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Sales lead with product 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].
4. Reconstruct the buying situation
Customers buy within a sequence of events: a problem becomes visible, an alternative is tried, funding becomes available and risk is evaluated. Strategyzer’s customer profile separates jobs, pains and gains; this guide adds the buying timeline.
Execution and evidence
Interview recent buyers, non-buyers and users separately. Ask for the last concrete incident, the workaround, who noticed the impact, who paid and what delayed action. Request a redacted artifact where appropriate.
Diagnostic and decision rule
Statements about future enthusiasm do not establish urgency. Search for a recent action that cost time, money or reputation and compare it with the stated priority.
If the problem is real but no trigger exists, design nurturing or a different segment rather than escalating immediate demand generation.
Applied working example
Draw the buying event as a timeline on one page. Put the customer’s own action above the line and your interpretation below it. An auditor’s request might be a trigger, but the decision to assign an internal project owner is stronger evidence of mobilization. Ask what happened between those two events before designing urgency language.
Working decision
What changed from tolerable inconvenience to active evaluation?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Research 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].
5. Make positioning comparative
Positioning becomes useful when a buyer can recognize why this option fits their situation better than a familiar alternative. That alternative may be spreadsheets, an internal process or postponement. Avoid claiming a universal category leadership that the evidence cannot support.
Execution and evidence
Write a message with target situation, desired outcome, distinct mechanism, supporting evidence and limitations. Test it against the buyer’s current workaround. Ask participants to explain the message back in their own language.
Diagnostic and decision rule
High likeability with low comprehension means the wording may be pleasant but commercially weak. A request for a feature list often signals that the differentiated mechanism remains unclear.
Keep claims that participants can understand and that product evidence supports. Qualify or remove claims that require unbuilt functionality.
Applied working example
Create a side-by-side internal comparison with the existing process. Include one condition where the old process remains adequate. For a small team with rare tasks, a shared spreadsheet may be rational. Admitting that boundary helps sales select the situation where your mechanism matters and prevents a weak account from entering a costly pilot.
Working decision
What would the buyer compare us with at the moment of choice?
- 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].
6. Design a paid learning offer
An offer needs a bounded outcome and a believable path to value. It should state implementation effort, required customer access, price, exclusions and what happens at the end. A pilot that merely postpones procurement can consume scarce delivery capacity without reducing uncertainty.
Execution and evidence
Draft a pilot charter naming the workflow, sponsor, administrator, baseline and acceptance review. Agree what data may be used and who owns deployment tasks. Present the commercial continuation separately from evaluation criteria.
Diagnostic and decision rule
If the buyer accepts only because the scope is unlimited or free services are excessive, revise the offer before treating acceptance as market evidence.
Reject pilots without an available sponsor, a measurable workflow or an agreed end date. Treat this as a local operating rule, not an industry benchmark.
Applied working example
Separate pilot success from permission to expand. A completed workflow can validate feasibility while exposing an expensive implementation path. At the closing review, ask both whether the buyer received value and whether the supplier could deliver it again at the planned cost. Keep those answers separate when deciding the next offer.
Working decision
What will the buyer do or pay to investigate this problem?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Sales and customer success.
- 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. Model the funnel backwards
A target becomes operational when its dependencies are explicit. Meetings, qualified opportunities, proposals and wins need distinct definitions and timestamps. Assumed conversion rates are planning variables until an observed cohort is old enough to measure them.
Execution and evidence
Calculate required wins from new recurring revenue divided by average recurring revenue per win. Divide again by each stage conversion assumption to estimate upstream workload. Check that the sales team can handle that number of conversations.
Diagnostic and decision rule
A mathematically possible funnel can exceed seller or implementation capacity. Run low, central and high scenarios rather than hiding uncertainty in a single decimal.
Pause expansion when the required workload exceeds capacity or when the acquisition cost exceeds the cash commitment leadership approved.
Applied working example
Use one definition sheet for the funnel. A qualified opportunity should have explicit eligibility and a real next step; a friendly conversation is insufficient. Record entry dates so that a recent opportunity is not treated as a loss simply because it has not had time to close. Show open cases separately from won and lost cases.
Working decision
Which funnel assumption most strongly changes the resource requirement?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Revenue operations and finance.
- 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. Choose channels through experiments
A channel is a repeatable access mechanism, not a list of platforms. Outbound, partner referrals and educational content create different response delays and selection biases. One good introduction does not establish repeatability; one silent week does not invalidate a long-cycle route.
Execution and evidence
For each candidate channel define audience, message, offer, delivery capacity, time window and decision metric. Use separate account lists to reduce overlap. Log who actually received the exposure, not just the number of sends.
Diagnostic and decision rule
Compare qualified conversations and progression by account cohort. If a channel produces engagement but no eligible accounts, investigate targeting before rewriting every asset.
Cap spend and labor in advance. Continue only when evidence supports the next bounded test; do not interpret an inconclusive small sample as a winner.
Applied working example
Charge labor to channel experiments even when no invoice appears. Five partner meetings, list preparation and seller follow-up consume capacity. Compare the next incremental test with the next best use of that capacity. A low media bill can still conceal an expensive route, especially if success depends on a senior executive’s repeated personal intervention.
Working decision
Which route reaches eligible buyers at a manageable effort?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Demand generation with 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].
9. Instrument the value journey
Acquisition metrics and product behavior must join at a suitable account identifier. Amplitude documents distinct charts for conversion and retention questions. Selecting a tool does not define your activation event or make an event a customer outcome.
Execution and evidence
Specify signup, setup completed, first meaningful workflow, repeated workflow and commercial status. Name event owners, properties, time zone and exclusions. Test events with synthetic internal accounts and remove those accounts from customer reporting.
Diagnostic and decision rule
If one action fires multiple events or several users represent one account, user-level conversion can mislead a B2B decision. Compare instrumented events with a small manual account audit.
Do not scale acquisition while identity, duplicate events or missing timestamps prevent the team from distinguishing failed onboarding from missing measurement.
Applied working example
Write the event specification in customer language before engineering names it. “First approved schedule published” is clearer than “activation complete.” Include the account identifier, actor role, completion condition and failure cases. Ask a success manager to inspect several event records against the actual workflow. This catches semantic errors that a technically successful event test misses.
Working decision
Which observable action indicates that a real workflow reached value?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Product analytics and engineering.
- 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. Deliver activation as part of GTM
The buying promise and first use should describe the same job. Fast signup does not guarantee activation when data imports, permissions or team coordination are prerequisites. Make customer effort visible before signing rather than hiding it in onboarding.
Execution and evidence
Map each dependency from agreement to the first completed workflow. Assign an owner on both sides, offer a narrow setup path and document the handover from seller to success manager. Track elapsed time and blocked time separately.
Diagnostic and decision rule
Repeated setup calls may indicate a product gap, unclear documentation or a mismatched segment. Tag the cause rather than calling all delays customer resistance.
If activation requires bespoke engineering outside the approved offer, pause new pilots until the scope or delivery model is corrected.
Applied working example
Track the waiting reason each day a pilot is blocked. Missing data, unavailable administrator, product defect and unclear scope require different owners. Removing one bottleneck can improve time to value without any new acquisition spending. Use this diagnosis before shortening the promised onboarding time, because a shorter promise does not remove the dependency.
Working decision
Which dependency prevents the promise from becoming a completed workflow?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Customer success 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].
11. Inspect economics and cash
Revenue quality matters alongside volume. CAC should include the acquisition costs allocated to the cohort, including relevant salaries and tools. Gross margin must use the same revenue scope and period as the direct delivery costs being subtracted.
Execution and evidence
Estimate CAC as attributable acquisition spending divided by new customers. Estimate gross-margin payback as CAC divided by monthly recurring revenue per customer multiplied by gross-margin rate in the denominator. Separate onboarding fees, recurring revenue and cash receipts.
Diagnostic and decision rule
Annual prepayment improves near-term cash but does not prove better unit economics. Heavy onboarding labor can make a pilot profitable on paper only if those costs are omitted.
Use contribution and cash sensitivities before committing another acquisition tranche. Avoid a lifetime-value estimate when retention history is immature.
Applied working example
Keep a cash calendar beside the contribution model. Salaries and implementation may be paid before the first subscription receipt. If a buyer pays late, the model needs a delay assumption rather than a higher nominal contract value. Stress-test simultaneous onboarding of the planned cohort and identify the maximum cash trough before authorizing the next tranche.
Working decision
What cash and delivery exposure does another ten customers create?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: Finance.
- 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. Run a weekly decision review
Weekly reviews should change decisions, not reward presentation polish. Review promises made last week, observed evidence, contradictory cases and resource constraints. A record of a rejected hypothesis is valuable because it prevents the next team from buying the same uncertainty again.
Execution and evidence
Allocate time to one bottleneck and one upcoming commitment. Ask the owner for the source behind each changed assumption. Update a decision log with continue, revise, pause or stop plus consequences for adjacent teams.
Diagnostic and decision rule
If all indicators remain green while deals stall, inspect definitions and stage aging. If every meeting changes the segment, the test design may be too unstable to learn.
Do not change audience, offer, message and channel simultaneously unless the current test is explicitly closed as uninterpretable.
Applied working example
End each review with one decision sentence that another function can act on. For example: “Keep the segment, reduce pilot scope and stop inviting accounts without an administrator.” This is more operational than “improve conversion.” Name the collateral, routing rule and capacity plan that need updating so the decision reaches daily work.
Working decision
Which evidence warrants a change this week?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: GTM 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].
13. Assign the 30/60/90-day operating plan
The timeline organizes learning, not a promise that every enterprise buyer will sign within ninety days. Days one to thirty should reduce market and buyer uncertainty; days thirty-one to sixty should test offer and delivery; the final month should evaluate repeatability and the next commitment.
Execution and evidence
Create milestones with named owners and reviewable artifacts. Reserve engineering and customer-success capacity before inviting pilots. Include an explicit no-expansion option at each milestone and preserve procurement timing as an external dependency.
Diagnostic and decision rule
Missed interview recruitment is different from evidence that a segment lacks urgency. Diagnose whether the bottleneck is access, relevance, offer risk or operational execution.
Release the next budget tranche only after the previous artifact is reviewed. Unfinished measurement work is a reason to narrow scope, not to infer success.
Applied working example
Reserve review time in the calendar before the project starts. A milestone that depends on an unavailable finance or security reviewer can stall despite completed research. Identify deputies and the minimum material needed for each gate. The timeline should show dependencies and waiting time rather than pretending all work proceeds in a clean sequential line.
Working decision
What must be true before the next thirty days are funded?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: CEO and functional leads.
- 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. Close with a decision, not a victory deck
At day ninety, choose whether to deepen the same route, change one structural assumption or stop. Explain the strongest evidence, the remaining uncertainty and the operational cost of the next step. The original EVIDENCE loop ends by examining retention because early enthusiasm can mask weak repeat use.
Execution and evidence
Write a two-page investment memo with cohort dates, eligible accounts, observed conversion, time to value, contribution assumptions and open risks. Include counterexamples and a renewed maximum-loss limit. Retention calculations depend on cohort maturity and return definition.
Diagnostic and decision rule
Compare customers old enough to have had the relevant repeat-use opportunity. Do not compare a new cohort’s incomplete month with a mature cohort’s full month.
Stop or redesign when the problem lacks priority, delivery cannot support the promise or conservative economics breach the agreed cash limit.
Applied working example
Describe the next falsifiable commitment in the closing memo. A sensible decision may be to repeat a narrow offer with a fresh cohort rather than hire or expand geography. Identify which uncertainty this repetition resolves. Without that question, another ninety days can reproduce the same activity while leaving the strategic decision just as uncertain.
Working decision
What evidence justifies the next commitment and what remains unknown?
- Evidence reference and date: [fill in]; counterexample or missing perspective: [fill in].
Responsible: 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].
Fictional worked example: ClearOps
ClearOps is a fictional workflow SaaS for maintenance teams. Every number below is an assumption for arithmetic practice, not a client result or market benchmark. The team has EUR 30,000 of acquisition capacity and expects six new customers at EUR 600 monthly recurring revenue each. Assumed monthly direct delivery cost is EUR 120 per customer. CAC = 30,000 / 6 = EUR 5,000. Gross margin = (600 − 120) / 600 = 80%. Gross-margin payback = 5,000 / (600 × 0.80) = 10.42 months, excluding one-time onboarding contribution, churn, expansion and financing costs.
If only three customers close, CAC becomes EUR 10,000 and payback becomes 20.83 months with otherwise unchanged assumptions. The difference is not a rounding error: leadership must decide whether the cash runway supports it. For six wins at an assumed 25% qualified-opportunity win rate, ClearOps needs 24 qualified opportunities. At an assumed 40% meeting-to-opportunity rate, it needs 60 eligible meetings. Neither conversion rate is observed evidence.
| Artifact | Owner | Acceptance question | Pause condition |
|---|---|---|---|
| Charter | CEO | One segment and loss limit? | No shared decision |
| Evidence ledger | Research | Independent dated cases? | Only opinions |
| Pilot offer | Sales + success | Deliverable with capacity? | Unbounded customization |
| Cohort memo | Finance + analytics | Comparable exposure and costs? | Incomplete denominator |
| Period | Primary work | Review artifact | Budget decision |
|---|---|---|---|
| Days 1–30 | Recruit and reconstruct buying events | Segment evidence and alternatives | Fund only a narrow offer test |
| Days 31–60 | Run offer and activation tests | Pilot workflow and cost log | Fund one route experiment |
| Days 61–90 | Repeat with a new eligible cohort | Economics and investment memo | Deepen, redesign or stop |
| Signal | Possible diagnosis | Check before changing | Next bounded action |
|---|---|---|---|
| Replies without meetings | Weak urgency or difficult scheduling | Read reply reasons | Test a lower-effort conversation |
| Meetings without offers | Wrong buyer or unsupported requirements | Review rejected accounts | Narrow qualification |
| Offers without activation | Implementation risk | Map prerequisites | Repair one dependency |
| Use without renewal evidence | Cohort immature | Inspect contract dates | Continue observation |
| Evidence class | Permitted inference | Not established |
|---|---|---|
| Interview statement | A person described a situation | Market prevalence |
| Observed workflow | A task occurred in context | A purchase budget |
| Paid pilot | A limited commercial commitment | Long-term retention |
| Repeat paid use | Continued value in that cohort | Universal product-market fit |
Capacity worksheet: the next customer cohort
Before authorizing the next tranche, list implementation hours for each new account, available delivery hours during the same period and work already promised to existing customers. Count specialist review and rework as well as the first setup call. Divide available hours by the conservative hours required per account to estimate a capacity ceiling. This is an operating constraint, not a sales forecast.
If the pipeline could exceed that ceiling, choose a smaller invitation cohort or a later start date. Do not solve the discrepancy by deleting assumed delivery work. Capture which dependency the team can improve and what that improvement would cost. Recalculate only when the process actually changes or new evidence justifies a different estimate.
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. Write the commitment before the campaign
A revenue goal alone cannot tell a team what to do next.
CHECKPOINT 02
2. Build an evidence ledger
Separate market descriptions from customer observations.
CHECKPOINT 03
3. Choose a reachable beachhead
An ideal customer profile should help reject accounts as well as select them.
Frequently asked questions
Who is this resource for?
SaaS founders, executives and marketing, sales and product leaders. Use the resource with the people accountable for implementation and approval.
What practical outputs can we create?
Evidence register of prioritized GTM assumptions Focused segment and tested value narrative Channel experiment with cost and stop rules Owned 30/60/90-day execution plan
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.