Find the earliest meaningful transition that is failing. That is a better starting point than changing the channel with the most visible dashboard.
A funnel is a working model of customer progress. Its stages need to represent real changes in behavior. Otherwise the diagram looks organized while the business remains difficult to diagnose.
Put the journey on one page
List the stages from relevant audience to retained customer. Record the count, conversion rate, elapsed time, and owner at each transition. Keep the cohort and time period consistent.
For a sales-led startup, the path might include qualified visit, enquiry, meeting, opportunity, customer, and successful onboarding. A self-serve product will need different stages. Use the actual buying and value-delivery process.
Match the symptom to an investigation
| Symptom | Investigate first |
|---|---|
| Few relevant people reach the offer | Audience access, demand, distribution |
| Relevant visitors do not act | Offer clarity, trust, page friction |
| Enquiries rarely become useful meetings | Qualification, scheduling, follow-up |
| Good meetings do not become customers | Buying process, differentiation, price, risk |
| Customers buy but do not reach value | Onboarding, product fit, implementation |
| Customers leave after initial success | Continuing value, expectations, competitive alternatives |
Do not treat the table as a verdict. Use it to choose which records and conversations to inspect.
Check the denominator
A low conversion rate may reflect a broader audience entering the funnel. A high conversion rate may reflect aggressive filtering that excludes valuable buyers. Review both volume and quality before deciding a stage improved.
Also examine time. This month’s opportunities may come from last month’s enquiries. Comparing them as if they were one cohort can create a misleading diagnosis.
Change the constrained step
If qualified buyers reach sales but cannot complete procurement, more traffic may produce more stalled deals. If onboarding fails, a better ad cannot repair the delivered experience. Allocate work where it can change the customer’s progress.
Choose one intervention and define the evidence that would support it. Recheck the whole journey afterward because removing one constraint can reveal the next.
What if every stage looks weak?
Start with a narrower customer segment and inspect individual journeys. Weak aggregate results can hide a promising segment or a fundamental offer problem. Use the ICP template to make that segmentation operational rather than merely descriptive.
Co-founder and CEO of Stackmatix, startup advisor, and former Head of Sales at MightyHive. Connect with Matt on LinkedIn ↗ · More about Matt →