Blog

BLOG

A focus on visualizations: Funnel charts

PUBLISHED ON 20 MAY, 2014

Today we’re going to take a break from the typical axis graphs to take a closer look at a rather unique visualization: the funnel chart. A funnel chart, very different from the previous charts observed in our visual series, shows a sequence of events in a proportional or static funnel. Each dimension value or metric represents a different stage of a process.

The name, funnel, is often misleading. When people think of funnels they think of an undisrupted flow. However, for the purpose of representing data, ignore your traditional ideas of a funnel. The name is solely there to describe the shape, as each funnel stage represents a definitive step or stage of a process.

So why do we use such a non-traditionally shaped graph? Tim Wilson stated the answer very clearly: Humans process relative lengths much more easily than they process raw numbers. We cannot disregard the funnel graph as a useful visualization tool, as it makes understanding data visually much easier for the reader, rather than trying to make sense of just numerical data.

A numerical table for example, hardly shows the same depth as a traditional funnel chart, making it difficult to pull insights.

 

When to use a funnel chart	When NOT to use a funnel chart
Funnel charts are best used to represent: Steps of a process Rate and location of loss Potential problems areas in a cycle	Funnel charts are generally not our best option for: Representing how individual parts contribute to a whole Comparing variables

Let’s consider for example where a funnel graph would be an ideal visualization: the representation of a Marketing and Sales cycle. An analyst can use a funnel graph to see how many leads complete each different key event leading to a sale. The funnel allows them to see how effectively the sales team’s strategies are working, as well as to observe where sales are lost and at what rate.

However, funnel charts can be controversial, no longer are scales or color choice the biggest issues. The graph itself, if done incorrectly provides little to no insight. So at Sweetspot our solution is to use a bar based funnel chart, a proportional funnel chart, or a pyramid.

Bar-based funnel charts represent each stage of a cycle with a different bar, as well as show the percentage of  progress between each stage, and a percentage from each step until the end (labeled below).

There are two options available for bar-based funnel: predefined bars or proportional bars.

The graph above has predefined bars, these show the basic flow of a funnel, but the numbers associated with each bar do not correspond proportionally. While less visual, this graph allows readers to focus on the numerical value of each stage.

The same data is captured below in bars which coordinate proportionally. As opposed to the predefined funnel chart above, this graph captures the same data but each bar is proportional to the variable or stage it represents. This allows to the reader to take in both visually and numerically the impact of the visualization.

 

The typical proportionate funnel chart represents data by magnitude, in order to offer a visual comparison from one stage to the next. In reality, any funnel chart which does not show a clear visual difference is ineffective to readers.

 

 

 

The Pyramid chart is an inverse funnel, which shows data in reverse order. This visual representation allows readers to find insights largely by comparing the magnitude of each variable.

 

Neither the pyramid nor the traditional funnel show a percentage explaining how many variables move from stage to stage, nor the percentage from each stage to the end.  The bar-based funnel shows a much clearer representation of success/fall out rate.

Alternatives:

Bar graph:

If you were to just center the bars, you would essentially have the same graph as the proportional bar-based funnel. However, you lose your percentages, which are very powerful in communicating key insights to stakeholders.

KPI Panel: 

 

A KPI panel allows users to define each variable separately as its own KPI. The problem is the visual element of how each variable is influencing the final sale is discarded.

In conclusion:

Although often criticized, the funnel chart can still be used as an effective visualization to identify the insights that matter most. The added detail of displaying a percentage between stages allows insights to be highlighted and acted upon quickly, which in the end is the most important quality of any data visualization.

Which funnel graph do you find the most effective at representing your processes, or do you have a preferred alternative?