Interactive funnel analysis
Funnel Growth Lab
A worked example of finding where applicants drop off, estimating the value of one focused change, and setting clear checks before launch.
TL;DR
Compare customer groups, find the biggest drop-off, test a realistic improvement, and connect added activations to an editable revenue estimate.
All applicants
- Started10,000
- Eligible7,600
- Offer viewed5,900
- Verification started4,100
- Verification complete2,850
- Activated2,420
Sample monthly group · 24.2% activation
01 · Find the problem
Find the biggest drop-off before choosing a fix.
Switch customer groups to see where people leave. Conversion and drop-off come directly from the sample counts.
Check the finding.
Change the inputs and see whether your choice still holds.
- 01Switch customer groups and see where the largest percentage drop-off moves.
- 02Change the improvement and revenue assumptions, then check the estimates.
- 03Review the test idea and safety checks before choosing what to do.
People at each step
- Started10,000100.0% of selected group
- Eligible7,60076.0% from prior stage
- Offer viewed5,90077.6% from prior stage
- Verification started4,10069.5% from prior stage
- Verification complete2,85069.5% from prior stage
- Activated2,42084.9% from prior stage
Step-by-step rates
| Step | Conversion | Drop-off | Applicants lost |
|---|---|---|---|
| Started → Eligible | 76.0% | 24.0% | 2,400 |
| Eligible → Offer viewed | 77.6% | 22.4% | 1,700 |
| Offer viewed → Verification started | 69.5% | 30.5% | 1,800 |
| Verification started → Verification complete | 69.5% | 30.5% | 1,250 |
| Verification complete → Activated | 84.9% | 15.1% | 430 |
Biggest drop-offs
Start with the largest leak
The list ranks each step by drop-off rate and keeps the number of people lost in view.
- 130.5% drop-off
Offer viewed → Verification started
1,800 applicants lost
- 230.5% drop-off
Verification started → Verification complete
1,250 applicants lost
- 324.0% drop-off
Started → Eligible
2,400 applicants lost
- 422.4% drop-off
Eligible → Offer viewed
1,700 applicants lost
- 515.1% drop-off
Verification complete → Activated
430 applicants lost
02 · Estimate the value
Test a realistic improvement, not a promised result.
Change one step and one revenue assumption. Later-step behavior stays the same so the math is easy to check.
Percentage points change the rate directly. A move from 40% to 45% is a five-point gain.
Change this input to connect conversion to money. It is an assumption, not measured customer value.
Estimated additional activations
+174per sample monthly group- Estimated step conversion
- 74.5%
- More people moving ahead
- 295
- Later-step completion held at
- 59.0%
- Illustrative revenue estimate
- $17,400
174 × $100 = $17,400. Excludes repeat use, refunds, costs, and profit.
See the math
Improvement used: +5.0 percentage points
5,900 × 0.050 × (2,420 ÷ 4,100) = 174People entering the step × the point increase × the share who finish later steps = estimated added activations.
- Assumes the group size and mix stay the same.
- Assumes added applicants finish later steps at the current rate.
- Does not include capacity limits, season changes, or test uncertainty.
03 · Decide what to test
Turn the finding into a test that can prove you wrong.
The funnel shows where people leave, not why. Customer research must support the idea before launch.
Make the hardest all applicants step easier
30.5% drop-off · 1,800 applicants lost
- What may be happening
- Applicants put off verification because they do not know why it matters, how long it takes, or which documents they need.
- Smallest useful test
- Test an upfront verification checklist with a time estimate and an explanation of why each item is needed.
- Primary metric
- Offer viewed → Verification started conversion
- Decision rule
- Keep the change only if more people start without more support questions or quick cancellations.
- What this test will not do
- Do not remove required verification or alter the offer itself.
04 · Protect customers and the business
A conversion gain only matters if quality stays strong.
The main metric shows possible growth. Safety checks show whether the change is worth keeping.
Safety checks
| Metric | Threshold | Why it matters |
|---|---|---|
| Verification quality | No meaningful drop | Growth cannot come from weaker required checks. |
| Support contact rate | No more than a 5% increase | A faster funnel should not create more confusion for support. |
| Seven-day cancellation rate | No more than +1 percentage point | More activations only help if people still want the product. |
| Largest customer-group conversion gap | Must not widen | Overall growth should not hide a worse result for one group. |
Checks before launch
- 01
Research
Is this the real problem?
Customer sessions confirm why people stop, not only where the numbers fall.
- 02
Design
Can people understand it?
Applicants can explain the requirement, next step, and result during testing.
- 03
Engineering + Data
Can we trust the tracking?
Views, steps, errors, and customer-group events match before the test begins.
- 04
Legal / Compliance
Does it follow the rules?
The copy and flow keep all required notices, consent, eligibility, and checks.
- 05
Operations
Can support handle problems?
Support steps, owners, escalation paths, and rollback rules are clear.
What comes next
Data shows where to look. Customer research shows whether the problem is real.
Before the roadmap changes, validate this analysis through customer sessions, a prototype test, and a measured experiment.