I have lost count of the number of times I have been handed team level benchmarking analysis and the data has been visualised in the form of a table. Why is this a problem I hear you ask? Well, in my opinion, the purpose of using data to benchmark, say teams, is to encourage improvement, identifying the teams performing well and encouraging lesser performing teams to learn from their peers and, as a result, leading to sustained excellence across an organisation. Surely the purpose of benchmarking data is not to concern ourselves with if performance is 0.1% higher in Team A when compared to Team B.
By visualising data as quantifiable numbers we are risking the data being used for judgement and individuals focusing on the minutia rather than the bigger picture. In the example to the right, if we are using the data to benchmark Teams A through E, do we really care if performance is 96.6% v a target of 95% for Team A, KPI 1? Is it not more important to understand that all of Team E's performance is consistently above target and exceeding that of many other Teams?
Never fear, there is an alternative! During Tableau Conference on Tour 2016 I was fortunate enough to watch a Bethany Lyons talk (I emphasise watch as I think at the time half of the content flew completely over my head as I was in awe of what she was able to achieve in Tableau!). During the talk Bethany shared a visualisation technique known as Lollipop Charts. A Lollipop Chart is a way of representing multiple measures on a single, vertical line, and as such understanding, visually, the spread of the measures. During the presentation I found myself thinking "Understand the spread of measures? Comparing variance? That sounds a lot like benchmarking to me!" And from that light bulb moment the concept of 'Performance Pops' was born, a visual technique of comparing performance across a range of measures for a dimension, whether that be Areas, Teams, or even Individuals.
How to read a Performance Pop?
I am not going to explain how to create a Lollipop Chart, Andy Kriebel has already recorded a brilliant step by step guide during a 'Tableau Tip Tuesday':
http://www.vizwiz.com/2015/02/lollipop-how-to.html
What I will instead focus on is how a business user would interpret a Lollipop Chart and why, within my organisation, individuals find them an incredibly intuitive way of understand team level performance variance.
Concept 1: Each 'pop' represents a single Key Performance Indicator (KPI)
Within each Team's vertical line, in my example, exists 5 'pops'. Each 'pop' represents an individual KPI. When designing the calculation for the 'pop' it is important to recognise that performance can have a wide range between Teams, for example a KPI could have a target of 90% and Team A could be performing at 95% whilst Team B is at 60%. If you attempted to visualise this actual variance it could result in the 'pops' losing their benefit as the range of the axis would be too great. As such I created a calculated field to standardise the variance. Any performance of that exceeded target by 5% or more was 'capped' at 5% whilst any performance that missed the target by 10% was also 'capped' at 10%. This allows the axis of the 'pops' to be fixed between -5% and 10% and 3 bands to be created (as described further in concept 2):
- Excellent performing metrics
- Average performing metrics
- Poorly performing metrics
Concept 2: The shaded area behind the Performance Pops allows a user to easily identify what level of performance a KPI is performing at
Visually representing performance, through coloured bands, supports a user to avoid focusing on the exact level of performance. This is critical as it is this technique that subtly moves users away from using the charts as a mechanism for judgement and towards using them for improvement and support.
Concept 3: The dotted line for each column represents the average performance of a team across all of the KPIs being analysed.
This allows Teams to be sorted, visually, identify those with the strongest levels of performance and those with the greatest opportunity for improvement.
Concept 4: Use tool tips to provide additional information
By using a Dual Axis chart you can actually have 2 different tool tips within the Lollipop Chart.
In my example if you hover over the vertical line you are able to see how performance changes over time, however when you hover over a 'Pop' you are able to see the actual performance v target performance.
Concept 5: Use highlighters to allow users to identify their Teams
When a Lollipop Chart has a large number of columns it can be important to provide users with the ability to highlight their Team or Teams on the chart. As such using Highlighter is an important element of functionality within the visualisation, in my example I have allowed users to highlight based on Team or Area.
Overall I think it is important to remember that judgemental behaviours are often caused by closed conversations, and presenting data quantifiably can increase the risk of it being used to critique or judge. I understand there is a time and place for when performance needs to be quoted at a set decimal place, for example when monitoring regulatory performance, however this is not the case when using data to benchmark. Benchmarking should be an open exercise, resulting in intrigue and conversation. It is important to visualise the data in such a way that encourages this behaviour, hence the lack of numbers in my chosen approach and looking at data thematically across a range of metrics.
My example 'Performance Pop' chart is as follows and can be viewed and downloaded via my Tableau Public profile (https://public.tableau.com/profile/simon.beaumont#!/vizhome/BusinessBoring-Performancebenchmarkinglollipopcharts/BusinessBoring-Performancebenchmarkinglollipopcharts?:SHOWVIZHOME=NO) :
Comments