Something recently occurred to me.  I’ve been researching chart and dashboard

design standards for an article I’m working on, and the eventual inclusion in an upcoming

book.  Funnel charts are commonly used in the sales and customer relationship

management (CRM) industry to visualize the volume of leads at various stages of the

sales qualification process.  Most of the leading reporting and chart tools include

funnel charts – which are little more than a stacked colored band chart arranged in

a funnel-shaped cut-out graphic.  The metaphor makes some sense but I’m questioning

the application.  Every funnel chart I’ve ever seen is the same shape - a funnel

with a linear taper down to a small diameter opening.  The idea is that we start

with a certain number of sales leads and then as they’re cold called, schmoozed and

nagged by sales people, a certain number eventually will buy something.  There

are a certain number of stages in this process and at each stage, a number of prospects

will fall away.

In some implementations of the funnel chart, the relative number of leads at a particular

stage is represented by the height of the band within the funnel.  I suppose

this makes a certain amount of sense.  One analogy would be that if you poured

different volumes of liquids having different viscosities into a funnel, they would

stack-up with different heights (remember JellO-1-2-3? – maybe not, I’m getting old).

Here’s my problem with the funnel chart: The funnel shape suggests that the volume

of leads is somewhat proportional to the diameter of the funnel at some point on the

vertical axis point representing a sales qualification stage.  At the top, it’s

wide and at the bottom it’s narrow to suggest that most of the leads don’t make it

to the actual sales stage.  Shouldn’t the diameter of each stage be consistent

with the actual aggregate of leads for that stage?  So if you were to sell stuff

to every lead in the first stage and had no attrition at all, the sales funnel would

actually be a cylinder.  Is the industry using this funnel shape because the

first person to design it was too lazy to do the math and draw a shape to represent

the actual data?  Sure, it would be a funny-looking funnel but it would look

like that data rather than a linear funnel.

Just thinking

By the way, if you’re the guy who originally designed it and I get a comment that

says “hey, butthead, I designed the funnel chart way back when and it seems to work

just fine for a lot of people!”  I’m not arguing that point.  I just think

it can be improved.

Weblog by Paul Turley and SQL Server BI Blog.