A company can improve every channel metric and still lose more revenue.
Paid clicks rise. Signups rise. The CRM fills. Product usage looks healthy enough. Sales still complains about lead quality, new users disappear after setup, and the team cannot explain which activity created the customers who stayed.
Each dashboard may be technically correct. The journey between them can still be broken.
That broken transition is a growth leak.
The job of a Growth Leak Diagnostic is to find the constraint that matters before the team buys another tactic, rebuilds another page, or spends another quarter improving a part of the system that was not limiting growth.
Sometimes demand really is the constraint. The wrong people arrive, there is not enough qualified volume, or the offer has not earned attention yet. A useful diagnostic has to be allowed to reach that conclusion.
It also has to be allowed to say the opposite.
More traffic can make a weak handoff more expensive. More leads can feed a response process that already loses them. More signups can inflate acquisition numbers while activation quietly gets worse.
The point is to establish what is true before deciding what to sell, build, automate, or scale.
This is the method behind Encanta's Growth Leak Diagnostic. It follows one live route from demand to commercial outcome, separates the visible symptom from the limiting constraint, and turns the evidence into a ranked first fix.
A low number is not yet a diagnosis
Teams are surrounded by symptoms.
Conversion is down. Churn is up. Paid media has become expensive. Organic traffic has flattened. Demo requests are arriving but not closing. Wallet connects look healthy while product usage does not.
Those numbers matter. They still do not tell you what to change.
A proper diagnosis separates five things:
- Symptom: the visible outcome or metric that created concern.
- Leak: the transition where qualified intent stops progressing.
- Cause hypothesis: the most plausible explanation supported by the available evidence.
- Constraint: the break currently limiting movement through the wider route.
- Intervention: the smallest useful change that can improve movement or test the diagnosis.
Consider an illustrative SaaS example.
The symptom is weak trial-to-paid conversion. The leak appears earlier: new workspaces are created, but users do not connect the data source required to produce the first useful report. The cause hypothesis is that setup arrives before the product has demonstrated enough value to justify the effort. The constraint is the missing first-value path.
The intervention might be a sample-data mode, a guided connection route, or a narrower onboarding path for the highest-intent segment.
These are different statements. Collapsing them is how a dashboard turns into the wrong scope of work.
A low conversion rate can be real and still sit downstream from the constraint. If acquisition is bringing the wrong intent, changing the form will not repair lead quality. If activation is weak, a win-back campaign arrives months too late. If enquiries wait in an inbox, generating more of them simply feeds the queue.
The method borrows a useful principle from the Theory of Constraints: system performance is shaped by its limiting constraint. Improving a non-constraint can make one local metric look better while overall movement barely changes.
Find the break with the greatest causal influence. Do not automatically choose the lowest number or the loudest stakeholder.
Map one revenue route, not the whole company
A diagnostic becomes vague when the scope becomes heroic.
Every audience. Every channel. Every product. Every journey. Every commercial outcome.
That produces a large audit. It rarely produces a clear decision.
Start with one audience, one source of qualified demand, one meaningful action and one commercial outcome. Map the route between them.
For a SaaS product, that might be:
High-intent comparison page → signup → workspace created → data connected → first report produced → paid plan → retained account
For a Web3 product:
Community or partner attention → proof and risk understanding → account or wallet action → first meaningful product action → second session → retained usage
For a service business:
Local search or referral → service proof → enquiry → useful response → booked appointment or consultation → paid work → review or referral
The route should describe what actually happens. Internal team boundaries are irrelevant to the customer.
Marketing may own the first touch. Product may own activation. Sales may own the conversation. Operations may own delivery. The buyer experiences one chain of promises and handoffs.
Use this sentence to force the route into plain language:
For [segment], qualified demand enters through [source], gains confidence from [proof], commits through [action], reaches [first value or progression], continues because [reason or follow-up], and becomes [commercial outcome] recorded in [system].
Struggling to complete that sentence is useful. It usually means the team has not agreed what first value is, the journey relies on an informal handoff nobody owns, or revenue cannot be connected back to the path that created it.
Those are findings. They are not admin problems to tidy up later.
The five handoffs to inspect
A revenue route can be described through six states:
Demand → Confidence → Action → First value → Continued movement → Commercial outcome
Diagnostic instrument / 01
Growth route and evidence rail
One live route. One illustrative primary leak. Evidence under every handoff.
- 01Demand
- 02Confidence
- 03Action
- 04First value
- 05Continued movement
- 06Commercial outcome
Illustrative primary leak
Action → First value
Continuous evidence rail
Runs beneath every transition, not after it.
- 01Commercial evidence
- 02Behavioural evidence
- 03Human evidence
- 04Operational evidence
The names adapt to the business. The transitions remain useful.
Demand to confidence
The first question is whether the right people arrive and find enough relevance, clarity, and proof to keep moving.
A team may call this a traffic problem because volume is low. The real issue could be poor demand capture, a mismatch between campaign promise and landing page, unclear positioning, weak search visibility, or traffic from audiences that were never likely to buy.
Look at the queries and sources bringing people in. Compare the promise that earned the click with the page they reach. Listen to the questions sales still has to answer. Separate high-intent behaviour from the blended site average.
Traffic volume is a weak diagnosis. The useful question is whether qualified demand reaches the point where a buyer understands the offer and believes it deserves consideration.
Confidence to action
People can understand an offer and still hesitate.
The risk may feel too high. Proof arrives too late. Price framing is unclear. The CTA asks for more commitment than the page has earned. A form collects information for the team's convenience before giving the buyer a reason to provide it.
For Web3, this can be the moment a wallet prompt appears before permissions, risk, and utility are understood. For SaaS, it may be a trial or demo route that hides what happens next. For a service business, it can be a mobile booking form, weak review proof, or no clear indication of availability and response time.
Inspect CTA movement, form starts and completion, recordings around the decision point, sales objections, and the position of pricing, proof, process, and risk information in relation to the ask.
A page can be attractive, informative, and commercially weak. The decision path is what matters.
Action to first value
A signup is not activation. A wallet connect is not product adoption. An enquiry is not a booked customer.
The action creates an obligation for the business to carry momentum forward.
This handoff often breaks because the organisation measures the easiest event rather than the useful one. Account created. Form submitted. Call booked. Community joined.
The harder question is what must happen next for the user or buyer to receive enough value to continue.
Inspect the event path after signup, time to first useful action, lead status after enquiry, response speed, onboarding completion, support demand and the difference between people who progress and those who stop.
This is where acquisition and delivery meet. The promise has been accepted. The system now has to prove it.
First value to continued movement
One useful moment does not create a durable customer relationship.
SaaS users may complete the first task and never return. A Web3 user may claim, stake, vote, or configure something once, then disappear. A service lead may receive a competent response but no follow-up when they are not ready to book that day.
Continued movement might mean a second meaningful action, a return session, progression through a buying stage, a reply to follow-up, expansion, renewal, a review, or a referral.
Cohorts, lifecycle states, CRM history, repeat use, support themes and reasons deals go quiet all help here.
Retention problems rarely begin on the cancellation screen. The decline usually becomes visible in behaviour first.
Continued movement to commercial outcome
The final handoff asks whether useful movement becomes something the business can recognise and learn from.
That might be paid conversion, qualified pipeline, retained revenue, booked treatment, repeat work, expansion, product usage tied to protocol utility, or another agreed outcome.
Perfect attribution is not always available. Inspectable evidence is still possible.
A service business may connect search source, call or form, response status, booking, and paid work. A SaaS team may connect acquisition source, activation depth, plan conversion, and retained account value. A Web3 team may connect campaign or community source, account action, product usage, return behaviour, and retained participation.
The exact stack matters less than the decision it supports.
Evidence runs underneath every handoff
Measurement sits underneath the route rather than waiting as a final stage.
Google Analytics funnel exploration and product analytics tools can show where users succeed or stop between defined events. That helps locate a transition worth inspecting.
It does not automatically explain the cause.
Amplitude's guidance on conversion drivers warns that correlation does not establish causation. Its Session Replay documentation shows how teams can add qualitative context to quantitative trends by inspecting individual sessions.
A useful evidence case combines commercial outcomes, observed behaviour, customer language, operational records, and comparison across segments or periods.
Each finding should also carry an evidence status:
- Observed: directly visible in source data or behaviour.
- Inferred: plausible and supported, but not yet proven.
- Unknown: missing, conflicting, or too weak to call.
Unknown is a valid output.
Pretending an unknown is a finding makes the report look more confident. It makes the work less useful.
Turn the evidence into a constraint call
Most journeys contain several problems. They do not all deserve equal weight.
A good diagnostic separates the primary leak, the amplifiers, and the unknowns.
The primary leak is the handoff where a fix is most likely to improve commercial movement, remove a dependency, or create the signal needed for the next decision. It needs a causal argument. "This page looks weak" is not enough.
Amplifiers make the primary leak worse without being the main constraint. A long form may amplify a trust problem. Generic lifecycle emails may amplify weak activation. Slow pages may amplify an already confusing decision route.
Unknowns are questions the team cannot answer yet. Which source creates retained customers? Why do leads disappear after a proposal? Which onboarding action predicts a second session? Unknowns often require instrumentation, interviews, or a small test before implementation.
An 80-item audit hides these distinctions. Everything becomes important. Nothing becomes owned.
The constraint should be written so another person can disagree with it:
For [segment], movement from [state A] to [state B] is weaker than expected because [cause hypothesis]. We see this in [evidence]. The likely commercial consequence is [effect]. Confidence is [high, medium, or low] because [reason].
Here is an illustrative service-business finding:
Mobile visitors arriving on high-intent treatment pages reach the booking route but fail to submit because the form asks for administrative and clinical detail before confirming availability. We see this in form starts, field-level abandonment, recordings, and repeated phone questions. The likely consequence is lost appointments from people who had already decided to enquire. Confidence is high because the same friction appears across behavioural and human evidence.
Compare that with "improve the website".
One can be tested, assigned, and disproved. The other is a label for future work.
Rank the first fix without pretending the score is science
Priority models help until fake precision takes over.
Scoring a finding 7.8 instead of 7.4 does not make the judgement objective. It can hide weak evidence behind a neat spreadsheet.
Encanta ranks interventions against five practical questions:
- Commercial impact: Which meaningful outcome could move?
- Confidence: How strong is the evidence that this addresses the constraint?
- Effort: What does it cost in time, money, complexity, and team attention?
- Dependency: What has to happen first?
- Time to signal: How quickly will the team know whether it worked?
The priority call should explain why this fix goes first, why it goes now, and what waits.
The intervention also needs a mechanism:
Change [specific part of the journey] for [segment] so movement from [state A] to [state B] improves. Measure [primary signal] and [commercial or guardrail signal]. Revisit the diagnosis if [falsifying condition].
That final sentence matters.
A diagnosis that cannot be proved wrong is just a confident opinion.
The finished diagnostic should make this judgement usable. It needs a route map, leak matrix, evidence notes, constraint call, annotated finding, first-fix brief, owner, measurement plan and a short operating sequence.
The Sample Growth Leak Diagnostic shows the structure. The team should be able to ship internally, scope a focused sprint, gather missing evidence, or stop spending until the route is clearer.
An audit folder does not do that.
A worked example (illustrative): weak paid conversion in a SaaS product
Consider a self-serve SaaS product with stable signup volume and weak paid conversion.
The first reaction inside the team is familiar. Add onboarding emails. Redesign the checklist. Offer a discount. Increase retargeting.
The diagnostic starts earlier.
The route
High-intent content or comparison page → signup → workspace created → data source connected → first report produced → teammate invited → paid plan
What the evidence shows
Signup and workspace creation are happening. Many new accounts stop before connecting a data source. Recordings show hesitation around integration choices and permissions. Support questions focus on which connection to choose and what happens after access is granted. Accounts that produce a first report are more likely to continue using the product.
That last point is correlation. It supports the investigation. It does not prove the fix.
The constraint call
The likely primary leak is workspace creation to first useful report.
The onboarding checklist and lifecycle emails are amplifiers. They repeat the setup process but do not resolve the missing value preview.
Pricing remains an unknown. It may become the next constraint after activation improves.
The first intervention
Create a first-value route that lets a new account see a realistic sample report, choose one recommended connection path, understand the permission request, and reach the first live report with fewer configuration decisions.
Measure data-source connection, time to first report, return after the first useful session, and paid conversion for the affected segment.
If first-report completion improves but qualified paid conversion and retained use do not move, the activation diagnosis was incomplete. The next investigation may sit in pricing, packaging, account fit, or depth of value.
That is how the diagnostic earns the second decision. It does not declare the entire business fixed because one funnel step improved.
growth-leak-worked-example
IllustrativeDiagnostic artifact / 02
Worked example diagnostic ledger
Keep the constraint, supporting friction, uncertainty, first action and disproof test visibly separate.
- 01Primary leakLikely constraint
Workspace creation → first useful report
New accounts are stopping before a connected data source produces the first useful report.
- 02AmplifiersSupporting factors
Onboarding checklist and lifecycle emails
They repeat setup steps without resolving the missing value preview or permission uncertainty.
- 03UnknownsUnresolved
Pricing, packaging and account fit
Pricing may become the next constraint after activation improves; the current evidence cannot settle it.
- 04First interventionFirst action
Create a first-value route
Show a realistic sample report, recommend one connection path, explain the permission request and reduce configuration choices.
- 05Falsifying conditionDisproof test
First-report completion improves, but paid conversion and retained use do not
Treat the activation diagnosis as incomplete and investigate pricing, packaging, account fit or depth of value next.
This ledger restates the article’s illustrative SaaS example. It is diagnostic logic, not measured client data.
For a closer look at this route, read Why SaaS Signups Do Not Activate.
Where growth audits usually fail
Some audits inspect everything. Coverage feels reassuring, but the team still has to decide what matters after the report arrives.
Others recommend departments. "Invest in SEO." "Improve CRO." "Build lifecycle." Those are capability labels, not findings.
A single data source can mislead. Analytics may reveal a drop. A recording may reveal confusion. A sales call may reveal an objection. CRM history may reveal a delay. The stronger case comes from the pattern across them.
Correlation causes another problem. Users who complete an event may retain better, but forcing the event may not create retention. It could be a marker of stronger intent or better account fit.
Then there is the backlog dump. A long list transfers the hard decision back to the client.
The diagnostic should state what goes first, who owns it, how it will be measured, and what evidence would change the order.
It should also be able to conclude "stop".
Sometimes the next step is a small internal fix. Sometimes it is instrumentation, customer research, offer work, or product definition. A diagnostic that always leads to a large implementation scope is a sales mechanism wearing a lab coat.
Run a practical self-diagnostic
You can apply the core method without buying a report.
- Choose one route. Pick one audience, one source of demand, and one commercial outcome.
- Write the states. Demand, confidence, action, first value, continued movement, and outcome.
- Mark one handoff. Choose the transition creating the most credible commercial waste.
- Gather mixed evidence. Use at least one behavioural signal, one human signal, and one commercial or operational signal. Record what remains unknown.
- Write the constraint statement. Name the segment, transition, evidence, cause hypothesis, consequence, and confidence.
- Define the smallest useful intervention. State the mechanism, owner, dependency, success signal, and falsifying condition.
- Decide the next move. Ship internally, gather evidence, scope a sprint, or solve a more fundamental positioning, product, or offer problem first.
Do not choose every handoff. That is a backlog, not a diagnosis.
A Growth Leak Diagnostic is also premature when there is no clear audience, no live journey, no meaningful behaviour to inspect, or nobody with the capacity to act.
Waiting can be the correct commercial decision.
So can making one obvious fix without paying anyone to produce a framework around it.
Growth becomes manageable when the route is inspectable
Good growth work leaves a causal trail.
You can see who arrived. You can see what promise they encountered. You can see where confidence dropped, where commitment became difficult, whether first value happened, and whether the outcome reached revenue or retained use.
You can also see the limits of the evidence.
The standard is a visible route, a defensible constraint call, a ranked first fix, and enough signal to know whether the change mattered.
Sometimes the answer is more demand. Sometimes it is a clearer offer, a better product path, faster follow-up, stronger proof, or cleaner measurement.
The diagnostic earns its value by being allowed to reach any of those conclusions.
Inspect the Sample Growth Leak Diagnostic if you want to see the format. If your own route is active but the constraint is still unclear, request a Growth Leak Diagnostic.
