In our work helping organizations develop their Experience Management (XM) programs, we find that they often want to jump in and start writing surveys. And that makes sense, right, you have to collect some data before you analyze it?!? Unfortunately, this common approach can lead XM efforts in the wrong direction. Instead of focusing first on data inputs, we highly recommend that you start by defining your desired analytic outputs.

Think of it this way: if you’re going on a trip, you don’t jump in a car and start driving without knowing where you’re heading. You select a destination and then map out a path to get you there. That information is critical for making sure your trip goes well. The same is true for an XM program.

To ensure your XM program ends up where you want it to go — driving insightful actions across your organization — start at the end point and use that information to define requirements for all of your data gathering. In other words, don’t just design surveys. Think about the inputs you need to deliver the insights that you want.

How can you reverse the pattern and design an XM program from back to front? Here’s an approach that we often recommend… “Fake reports.”

Before you even think about writing a survey, write a fake report first. Think about who is the key audience for the insights and what information they need to make better decisions. Imagine what the report would look like to fulfill those needs and what specific information you’d like to communicate to those key insight consumers. Sketch it out in PowerPoint, Excel, or Word, and the like, with placeholders for your outcome measures and functional drivers. Spend time thinking about those placeholders, using the guidelines below.

  1. Identify the outcome measures that will be the North Star for your organization. These metrics – be it CSAT, NPS, CES, or some other measure – track what behaviors you care about over a specific time period, such as the next six months or the next couple of years. Do you want your customers to find you the easiest company to work with? Then focus on a customer effort score. Do you want your customers to think about your organization when they have to make a next purchase? Then a Likelihood to Purchase measure might be your outcome measure. Do you want your customers to talk about your organization positively with friends and family? Then NPS might be your North Star outcome measures. When you’ve identified these key metrics, you need to treat them as dependent variables.
  2. Consider how you can drive improvements in these metrics. What can your organization change to increase the score of the metric? These changes that you can make—for example, making a purchase in two clicks on your website instead of seven—are called functional drivers. These represent the independent variables that can be changed to drive your outcome measure up (or down). For your “fake” report, make a list of all the functional drivers that you can think of and assign where you want to be in terms of achieving those drivers in the next six months. Don’t include functional drivers that you cannot change. For example, if you sell software by subscription, and your executives require the locked-in price, don’t focus on pricing changes as a functional driver. Your customers may not like it, but it isn’t a lever you are currently able to change, so don’t use it in your “fake” report.
  3. Socialize the fake report with the key stakeholders. Share the fake report with the people who will be consuming the insights and incorporate their feedback into a final draft. Now that you have a fake report that means something to your stakeholders, use it as a requirements document to define the data that needs to be collected. With this information, you can start to design your surveys. Make them as short as possible, focused only on your outcome measures and functional drivers.

If you start with a clear view of the insights you want to deliver, you’ll do a much better job of creating the appropriate surveys and identifying other data inputs you need that will drive a successful XM program.


Bottom line: Before you even think about writing a survey, write a fake report first.

Carol Haney, Ph.D., is a Distinguished Researcher with Qualtrics, specializing in CX

Elizabeth Dean, is a Senior XM Scientist with Qualtrics, specializing in market research design