This blog originally appeared on the Stanford Social Innovation Review website.

A nonprofit executive recently told me about the dismal state of her organization’s performance management data system. Despite carefully selecting a vendor and spending significant resources, staff compliance with data entry is low, senior leaders rarely view the system’s auto-generated reports, and adding new fields and queries requires moving mountains.

She’s not alone in her struggles. The more than 150 nonprofit leaders who have taken Bridgespan’s Performance Measurement Diagnostic, a survey that encompasses 10 elements of measurement, gave the lowest marks—by far—to their data systems. The challenge is particularly acute for organizations with budgets greater than $25 million; they are roughly twice as likely to rate their data system as a weakness compared to organizations with budgets less than $5 million.

What’s going wrong here? In our view, too many nonprofits are falling into the “data system trap”—that is, they are building or buying their performance management data system before they are really ready and have laid the essential groundwork.

Here are five strategies to keep you out of that trap:

1. Don’t start the process until your theory of change and indicators are in order. A data system is only as good as the measurement approach to which it is anchored, and a good measurement approach starts with a clear theory of change. If you haven’t articulated your target population, the benefits you are seeking for that population, and the activities, strategies, and resources you will use to get to those benefits, you won’t have the right input, output, and outcome indicators to measure, and you won’t have the right data system to track them.
2. Go through at least one data collection cycle to ensure the tools and reports are solid. For nonprofits launching new programs or initiatives, oftentimes the data collection tools and reports used in the first year don’t end up being the exactly the right ones. So despite the frustrations of having to use clunky Excel spreadsheets or some other makeshift data system early on, this is often the right call, as it helps you avoid locking into the wrong data system for your ultimate needs.
3. Make sure that everyone who will contribute to the system has a voice in designing it. Oftentimes, the success or failure of a data system will hinge on whether front-line staff and other contributors view inputting data as a chore or as absolutely integral to providing the right supports to clients. And yet, most nonprofits don’t engage staff until well after the planning process is complete. This is a huge mistake, because people are inclined to support what they help create. If you don’t give staff members the chance to tell you what reports they need to do their jobs better, you’ll create a system that serves only leadership well. This kind of a system won’t be around for long.
4. Use a cross-functional planning team to get a cross-functional system. Performance management data systems first and foremost need strong client tracking, but program data alone isn’t enough for organizations to make good decisions. A cross-functional planning team can help clarify important decisions the organization regularly makes, and then work backward to figure out the additional data sources the organization really needs to inform those decisions (for example, costs and revenues, social media, government data, and volunteer data). While one multi-faceted system integrating all of this data may not be the right answer for every organization, it’s important to go into the process with a clear understanding of information needs and priorities.
5. Treat it as a change management, not a technology, project. Most nonprofits ignore the massive cultural and behavioral change necessary for a new data system to be effective. Change management thinking says you need, among other elements, a compelling case for change (with a strong emotional appeal), supporters mobilized at all levels, and fast feedback loops. Make sure the team leading the effort is well versed in strategies for change before getting started.

Using these five strategies will greatly increase the odds that you are developing your performance management data system on a strong foundation. And if you get to the point of choosing a vendor, don’t just rely on what they tell you. Talk to nonprofits in your field that have experience with each of the leading contenders. Just make sure you probe during these reference checks, because often the root cause of unhappiness with a data system is less about the vendor they selected, and more about their inability to have laid the right groundwork upfront.

What items am I missing on my list? What do you wish you had done before setting up your current system?