program evaluation, logic model, performance management

KPIs are for driving performance: Four questions to ask about your program's KPI

April 23, 2020 | Pieta Blakely

To evaluate performance, most organizations select one or two KPI(s) — Key Performance Indicator(s) —to focus on. A KPI is the outcome measure that your organization is most focused on. But once you’ve selected your KPI(s), what’s next?


(Haven’t selected your KPIs yet? Check out this post on how to select these measures. Then come back here to read what to do next!)

To measure and use your KPI(s) well, you will want to create an action plan. Typically, you create this plan at some sort of inflection point when you want to improve program performance, e.g. the end of the year, or when you do your strategic planning. To help guide you through the process of creating a strategy to improve program performance, I’ve created a series of questions you can walk through with your program staff and relevant stakeholders. By answering these questions, you’ll be well on your way to creating a strategy to improve program performance.

hand drawing creative strategy with light bulb as concept


Questions to Guide Your Planning Process

1. What is the trend? Create a chart that uses your historic data to help project the trend. For now, keep this simple:
  1. Chart your data.
  2. Look at the level and direction of your trend line.
  3. Continue the trend line in the direction it’s going.
2. Why? Why is this your trend? For example, imagine a classroom attendance chart, where the line remains level until Thanksgiving, after which there is a significant drop off. Ask:
  1. What causes the ups and downs? In the case of the attendance chart, why did it go down suddenly?
  2. What are the real reasons for the change? The danger here is a temptation to simply restate the trend: “Our problem is attendance went down after Thanksgiving.” Ask the deeper questions to find the true problem. (For ideas on how to tackle this problem, see this Causes Diagram from DIY Toolkit.)
  3. What is the cause of the problem? There are two caveats to be aware of when looking for the cause of the problem:
    1. Don’t simply find a correlation. Remember that correlation is not causation.
    2. Don’t let yourself off the hook. Make sure you’re not coming up with explanations that let your organization off the hook. It’s important to look for systemic issues before jumping to conclusions about your target population. 
3. What promising practices do we know of that address this problem? Look for knowledge from those who have experienced and/or studied this issue.
  1. What does research tell us is effective about this program?
  2. What approaches have your or other organizations taken to address this problem? (e.g., maybe there’s an approach that worked with a younger population that could apply to an older population)

4. Who are our current and potential partners in this work? Ask this question because there might be resources these other places have that you don’t. For example, one study showed that the cause of poor school attendance in a particular school district was a lack of clean clothes. The school partnered with Whirlpool to provide laundry services for students.

Download the Process Measure Worksheet


Create Your Action Plan
The answers to the above questions are the building blocks for your action plan. Depending on the depth of your answers, you may have a lot of data to work with — and you may be wondering how to choose your next steps. You’ll need to take your data and figure out the steps you need to take to create a plan that:

  • is evidence-based,
  • uses your resources appropriately, and
  • addresses a real problem.


Using the steps above will ensure that you craft a plan that is rigorous and based in evidence. Furthermore, you will be able to document your decision-making process and check in on your progress periodically. The steps above support a rigorous decision-making process that will result in an action plan that's based on data, supported with evidence, and well documented.

Subscribe to my blog

 

 

Tags: program evaluation, logic model, performance management