3 tips for data analysts on how to identify KPIs for a dashboard

In the world of business, key metrics (or Key Performance Indicators) are business signals that can help the person in charge to quickly assess the state of things and act upon them in priority driven and timely manner. Performance measurement is a well studied and documented practice in the business arena.

Here's my top 3 tips for data analysts to help identify the most appropriate KPIs for your dashboard or report or visualization.

You can apply these same three tips in either of cases - i) if you are starting from a blank slate and identifying new KPIs to implement in a new dashboard, or ii) if you already have an existing dashboard (perhaps in Excel) with some existing KPIs in use that you want to validate and port to a different dashboard tool (like Tableau); or anything in between.   

1) Audience driven: A KPI must match the needs and experience level of its intended audience. Make sure that in the process of identifying the key metric, there's sufficient attention paid to identifying the business needs addressed as well as granularity level required based on intended audience's experience level. Most new dashboard or reporting tools allow you to have drill thrus so that you can maintain multiple levels of abstraction in multiple layers of dashboards. 

Don't be shy to interview the audience and ask questions about who will be the most frequent user of these metrics, how frequently will these KPI be required a refresh/rerun, how much summarization and aggregation will be beneficial and so on. There's usually no single correct answer, but both your audience as well as yourself will most likely gain invaluable additional thought clarity by mini-workshopping on these points. 

2) Context sensitive: KPIs make a lot more sense when put in context of other related KPIs. For example, a metric about Region X's growth rate may become much more relevant when compared with growth rates of Regions Y, Z and overall growth rate. It will help to always think about the context of the KPI that is needed to provide the proper business value to the KPI.  

There's not many rules of thumb, so getting the right context around each of the metrics may take some trial and error, but rest assured the output is much more likely to provide the impact and swiftness in decisioning than a poorly contextualized set of metrics. 

3) Actionable insight: Perhaps the most important aspect of a metrics - it must be actionable. The person using the metric should be able to spend least amount of time possible glancing at the metric and know the priority/severity and next course of action required.  

Going back to interviewing the audience, you should ask questions about the intended actions or take-aways that users expect from the set of metrics . It will not only help yourself in finessing the KPIs by getting visibility into how downstream processes are kicked off based on the insights provided by the metric.

As you can sense, identifying and defining metrics is meant to be a very iterative, interview-driven and collaborative process. The biggest mistake you can make is to create a dashboard in isolation. You should be ready to reach out to various stakeholders and understand the downstream needs and expectations, along with upstream constraints like data availability (topic for another day). 

So now that we've talked about how to identify and define key metrics needed for your business, in next post I'll share tips on how to implement these metrics in a dashboard. 

Until then, happy reporting!