Introduction

This article will focus on collecting and defining metrics and key performance indicators for executive and operational dashboards. While the techniques discussed here can be used across many different business intelligence requirements gathering efforts, the focus will be collecting and organizing business data into a format for effective dashboard design.

With the explosion of dashboard tools and technologies in the business intelligence market, many people have different understandings of what a dashboard, metric, and key performance indicator (KPI) consist of. In an effort to create a common vocabulary for the scope of this article, we will define a set of terms that will form the basis of our discussion. While the definitions below might seem onerous and require a second pass to fully understand them, once you have grasped the concepts you will have a powerful set of tools for creating dashboards with effective and meaningful metrics and KPIs.

Metrics and Key Performance Indicators

Metrics and KPIs are the building blocks of many dashboard visualizations; as they are the most effective means of alerting users as to where they are in relationship to their objectives. The definitions below form the basic building blocks for dashboard information design and they build upon themselves so it is important that you fully understand each definition and the concepts discussed before moving on to the next definition.

Metrics: When we use the term metric we are referring to a direct numerical measure that represents a piece of business data in the relationship of one or more dimensions. An example would be: “gross sales by week.” In this case, the measure would be dollars (gross sales) and the dimension would be time (week.) For a given measure, you may also want to see the values across different hierarchies within a dimension. For instance, seeing gross sales by day, week, or month would show you the measure dollars (gross sales) by different hierarchies (day, week, and month) within the time dimension. Making the association of a measure with a specific hierarchal level within a dimension refers to the overall grain of the metric.

Looking at a measure across more than one dimension such as gross sales by territory and time is called multi-dimensional analysis. Most dashboards will only leverage multi-dimensional analysis in a limited and static way versus some of the more dynamic “slice-and-dice” tools that exist in the BI market. This is important to note, because if in your requirements gathering process you uncover a significant need for this type of analysis, you may consider supplementing your dashboards with some type of multi-dimensional analysis tool.

Key Performance Indicators (KPI): A KPI is simply a metric that is tied to a target. Most often a KPI represents how far a metric is above or below a pre-determined target. KPI’s usually are shown as a ratio of actual to target and are designed to instantly let a business user know if they are on or off their plan without the end user having to consciously focus on the metrics being represented. For instance, we might decide that in order to hit our quarterly sales target we need to be selling $10,000 of widgets per week. The metric would be widget sales per week; the target would be $10,000. If we used a percentage gauge visualization to represent this KPI and we had sold $8,000 in widgets by Wednesday, the user would instantly see that they were at 80% of their goal. When selecting targets for your KPI’s you need to remember that a target will have to exist for every grain you want to view within a metric. Having a dashboard that displays a KPI for gross sales by day, week, and month will require that you have identified targets for each of these associated grains.