Flow, process and networks

Funnel chart (stages and conversion)

Compares counts at ordered process stages to show continuation and drop-off.

Chart
Conversion
Drop-off

What it answers

Where are people or items lost between stages?

Shape: Progressively sized stage bars
What it shows: Bottlenecks and drop-off.

Example

The question

Where do we lose people between visiting and buying?

Online shop funnel, September

Bars shrink stage by stage; the share kept from the previous stage is shown. Visits 90,000; Product views 41,000; Added to basket 9,800; Checkout 5,600; Paid 4,300. Overall 4.8% of visits become paid orders.

Illustrative data. It belongs to no company.

When to use it

Use it when

A clearly defined sequential cohort.

Avoid it when

Independent stage populations presented as one journey.

How to read it

Read stage order and counts, then calculate stage-to-stage and overall conversion.

Note

Define cohort, stage order and conversion window.

The data you need

Layout: Long table
Fields: StageOrder: integer; Stage: text; Count: nonnegative number
One row is: One stage and its eligible cohort count. What is the grain?
Rules: Order stages explicitly. Sequential-cohort counts normally do not increase. Independent stage totals need a different interpretation.
Optional: Raw alternative: EntityID,Event,Timestamp with cohort and window rules.
Preparation: Define a sequential eligible cohort and time window; derive ordered stage counts and stage-specific conversion denominators.
Units: Step conversion =next-stage count/previous eligible-stage count; overall conversion =final/initial cohort.

A long table: one row per observation, few columns. New data adds rows.

Worked example

Visits 1000, trials 200, purchases 50 gives 20% visit-to-trial,25% trial-to-purchase and 5% overall conversion.

StageOrder	Stage	Count
1	Visits	1000
2	Trials	200
3	Purchases	50

All charts in the library