SaaS funnel analysis worksheet

Copy into your preferred editor. Define units, progression and cohort maturity before calculating. A scenario is a model with assumptions, not a forecast or proof of achievable lift.

Measurement definition

Decision this analysis should inform:
Unit: user / account / another defined unit:
Eligible starting population:
Cohort entry date/range:
Observation window and minimum cohort age:
Required order of events:
Identity matching approach and limitations:
Excluded internal/test records:
Data sources and known definition changes:
Validation checks performed:

Funnel

Step	Exact event / condition	Eligible count	Conversion from previous step	Quality / data concerns
Start			—
Step 2
Step 3
End

Step conversion = next-step count / eligible previous-step count. Use rates as decimal fractions in multiplication: 5% = 0.05.

Scenario comparison

Candidate change	Baseline rate	Scenario rate	Expected downstream outcome	Assumptions / evidence	Effort / risk

Relative lift = (new rate - baseline rate) / baseline rate. Percentage-point change = new percentage - baseline percentage. Relative lift is undefined when baseline is zero; investigate counts and measurement instead of dividing by zero.

Completed fictional example

Ordered mature cohort with resolved user identities: 2,000 visitors → 100 signups → 40 activated → 10 paid. The rates are 5%, 40%, and 25%; overall conversion is 0.5%.

Scenario	Calculation	Expected paid customers
Baseline	2,000 × 0.05 × 0.40 × 0.25	10
Signup rate reaches 6%	2,000 × 0.06 × 0.40 × 0.25	12
Activation reaches 50%	2,000 × 0.05 × 0.50 × 0.25	12.5
Activated-to-paid reaches 35%	2,000 × 0.05 × 0.40 × 0.35	14

Fractional outcomes are model expectations. Real customer counts are whole numbers. These scenarios hold other rates and traffic constant; changes can affect audience quality and downstream rates. Figures are not benchmarks.

Investigation and intervention brief

Where the experience appears to fail:
Evidence from data QA:
Evidence from successful and unsuccessful users:
Proposed explanation and alternatives:
Change to investigate or make:
Expected outcome and assumptions:
Quality/reliability guardrails:
Owner, dependencies and effort:
Review method and window:
Result, confidence and next decision:

Final checks

Same unit and comparable maturity across steps?
Counts reflect ordered progression rather than unrelated monthly totals?
Missing, duplicate and internal events checked?
Impact compared with effort, evidence and risk?
Small-sample or before/after evidence described honestly?