Design an Effective Performance Measurement Dashboard
Design an Effective Performance Measurement Dashboard
1. Gather and document requirements for what you expect to include in your dashboard.
Keep in mind that a performance measurement dashboard is rarely the best mechanism for answering all possible questions. It will be most effective if it shows: Te results of 1-5 key performance indicators (KPIs). Some standardized context for each of those KPIs, like performance versus targets or benchmarks. No more than one level down of supporting metrics. For example, a breakout of the KPIs by marketing channel. At minimum, your dashboard plan needs to include the following: The metrics to be included. Any necessary organization or grouping of those metrics. Which metrics are KPIs vs. which metrics are simply supporting metrics. The context that needs to be included for each metric (such as summary totals, comparisons to targets, trendlines, etc.) If warranted, the data source for each metric. Review this plan with the target audience for the dashboard and incorporate their feedback before proceeding with the design.
2. Create a rough wireframe of the dashboard that incorporates the components from your plan.
Create a true offline sketch to prevent getting prematurely locked into any perceived limitations of the platform that will ultimately be used to create the dashboard. Work for a design that fits on a single screen without becoming unduly crowded. It’s OK to need a second screen for truly supplemental data, but you may be trying to do too much with a single dashboard if multiple screens are needed to capture everything in your plan.
3. Create a standardized widget that displays relevant context for each KPI to ensure that they're clearly represented.
Your widget should include at least the following context: The name of the metric. An assessment of the metric (comparison to a target, a benchmark, or a historical value). Additional context, such as how the metric trended within the reporting period. Example dashboard widget for a single metric using Google Data Studio You could also sketch the widget offline before building it. Either way, confirm that the widget provides a true at-a-glance assessment of both the value of the metric and the context so that the viewer can immediately understand how the metric is performing.
4. Build your dashboard using a business intelligence platform like Power BI, Tableau, or Qlik - or within a spreadsheet.
Stick to the plan and design you developed in the previous steps where possible, even if it requires pushing the capabilities of the platform. This is an automated performance measurement dashboard built using Microsoft Excel: Example of an automated performance measurement dashboard built using Microsoft Excel
5. If you're building your dashboard in a spreadsheet, use the same narrow width for all columns and merge cells to allow for flexible vertical placement and configuration for all dashboard components.
Spreadsheets can support flexible dashboard visualizations simply by making all the columns very narrow and then merging cells: Dashboard widget created within a spreadsheet
6. Review the dashboard with your primary audience, determine whether they can easily get the information they expect, and incorporate their feedback.
Ask (either directly or by observing their reaction) if they can easily get the information they are expecting. Because the dashboard will be standardized and recurrently used by the same audience, it is okay if they have to learn how to read the dashboard to some extent. But the less this is required, the better. Incorporate their feedback judiciously. There will likely be requests to add data to the dashboard, but there will never be requests to remove data from the dashboard. While well-intended, this can lead to dashboard bloat – a multi-screen dashboard that the audience actually finds unwieldy and unhelpful. Always ask if any requested addition is: 1) truly needed as an ongoing and always-available metric (rather than a one-off analysis request), and, if so, 2) whether this dashboard is the most appropriate delivery mechanism for the information.