Many of my clients use KPI's. They often want an overview of all their KPI's together, but that gives some challenges: how to fit them all on one page? Which goes where? How do they fit together? Does the end user know what all these numbers mean?
In this blog post I will share some of my learnings
KPI stands for Key Performance Indicator. It is a numerical value that is measured over time, to indicate change in performance. Conversion for example; how many visitors actually buy something? Or NPS; how likely are you to recommend us? Or number of complaints; how many complaints are logged as a percentage of items sold?
A KPI can have multiple drivers. Drivers are the elements that influence a KPI. For example, if there are less Returns, your Profit is higher. Return percentage is therefore one of the drivers of Profit. Sales is another driver of Profit. Sales can be split into sub-drivers Quantity and Price. All these drivers are KPI's themselves: numerical values that can be measured over time.
When you visualise all these drivers/KPIs in their hierarchy, a tree starts to emerge. It is important to visualise all these seperate elements together, because this can show the relationship between the drivers and KPIs. "Margin comes down, Quantity goes up, as does Profit, but NPS drops. Why?" Maybe because shipment time went up too, causing unhappy customers.
When every element in your KPI tree remains stable, but still your lead driver is dropping, it is a signal that you are either missing a driver, or there are external factors that are impacting it. Not all drivers are within your control; when the economy is going through a recession, your sales might go down. Or when everyone is working from home and ordering online, your shipment time goes up.
A single value KPI should always be reported with a target, % change or Forecast. I prefer to have all: a single value, a vs Previous Year and a trendline with a forecast or target line. Idealy the single value can be adjusted to show Current Year, Last month or Last week and the trendlines are shown per month or per week.
Another nice addition is to hightlight the weeks in all trendlines when you hover over them. This makes it easy to compare different drivers in the same week. "I see a spike in quantity in week 39, what happened with other drivers?" Sales remained stable, but discount spiked too.
This flexibility of the dashboard is important because it can support multiple users; Management might only be interested in current year and quarter overviews, but Accountmanagers who are responsible for perfomance have need for a weekly overview.
Yes! I've created an example dashboard in Tableau which uses all these learnings. View it in the portfolio section!
