Why is my end-to-end rate so much worse than each step?

Because they multiply. Four steps at eighty percent each is 0.8 to the fourth, which is forty-one percent, not eighty.

Businessmarketing

Funnel Conversion Calculator

Step-by-step and end-to-end funnel conversion, with the weakest step identified. Rates multiply rather than average, so one bad step caps everything downstream.

Also called: funnel drop off calculator, stage conversion.

End-to-end conversion

2.5%

2.5% end to end. The weakest step is step 3 to 4, converting 27.78%. Fixing that one step to match the next best would deliver 20 more at the end.

Rate at each step

1 to 2: 30%, 2 to 3: 30%, 3 to 4: 27.8%

Weakest step

step 3 to 4

Its rate

27.78%

Extra conversions from fixing it

Total lost through the funnel

9,750

How this is calculated

Method and background

How this is calculated

Each step converts a fraction of the one before, and the end-to-end rate is the product. That is why a funnel with four steps at eighty percent each converts at forty percent overall rather than eighty. It also means effort belongs at the weakest step: improving a step that already converts at ninety percent has almost no headroom, while a thirty percent step has most of the loss sitting in it.

end-to-end conversion is the product of the step rates, not their average

r_i: Rate at each step

Worked examples

Each of these is asserted on every build. If a change to the engine ever moved one of these answers, the build would fail before the page could print it.

a four stage funnel

Stage 1 count: 10,000
Stage 2 count: 3,000
Stage 3 count: 900
Stage 4 count: 250
Stage 5 count: 0

End-to-end conversion2.5%

250 of 10,000 end to end

Open this example

a perfect funnel loses nothing

Stage 1 count: 1,000
Stage 2 count: 1,000
Stage 3 count: 1,000
Stage 4 count: 0
Stage 5 count: 0

End-to-end conversion100%

boundary

Open this example

Method and limits

What it assumes

Each stage is a strict subset of the one before.

What it deliberately does not model

Users who skip a stage or re-enter break the subset assumption.
Time lag between stages means a snapshot mixes cohorts.

Formula version 1.0.0 · definition 1.0.0 · India · Report a problem with this calculator

Frequently asked questions

Why is my end-to-end rate so much worse than each step?: Because they multiply. Four steps at eighty percent each is 0.8 to the fourth, which is forty-one percent, not eighty.