Resource · Dashboards and reporting
What belongs on a nonprofit board dashboard (and what doesn’t)
Boards ask for a dashboard because they want to see how the organization is doing without reading a stack of reports. Too often what they get is every number the database can produce, arranged on one page. It looks thorough. It doesn’t help anyone decide anything.
A board dashboard works when it answers the questions the board is actually responsible for, in a form people can read in a few minutes. This guide is for executive directors and small teams who are building one, or fixing one that nobody opens. It covers what to put on the page, what to leave off, and how to design it so the numbers can be trusted.
Start with the board’s questions, not your data
The usual mistake is to start from what’s easy to count. Start instead from what the board has to oversee.
Two primary sources describe that job. BoardSource lists oversight and accountability, fiduciary responsibility for stewarding resources and following legal and ethical standards, and financial stewardship among the common responsibilities of a nonprofit board. The National Council of Nonprofits describes three legal duties: care, loyalty, and obedience, which covers following the law and the bylaws and staying true to the mission. BoardSource’s dashboard book suggests measuring outcomes tied to the mission, strategic initiatives, and risk factors.
For a small nonprofit, that translates into four groups of questions:
- Mission delivery. Are we reaching the people we intend to, and is it helping them?
- Finances. Is money coming in and going out as planned, and can we keep operating?
- Risk and compliance. Are required reports, renewals, and policies current, and is anything getting worse?
- Fundraising. Is the revenue we’re counting on on its way?
Those four groups are a planning device, not a list any source requires. Your organization may need different ones. What matters is the method: write down the questions your board asks in meetings, then find the measure that answers each. If a proposed measure doesn’t connect to a question or a decision, leave it off. What the board is legally required to review is a question for your attorney or accountant, and the National Council of Nonprofits recommends consulting a lawyer or accountant with nonprofit accounting expertise on those points.
Keep the board view small, and separate from the working view
A program manager’s dashboard is a working tool. It might show this week’s caseload, who is overdue for a follow-up, and which records are incomplete. It needs detail, it changes daily, and it may include client-level information.
A board dashboard is a summary. It shows where things stand, which way they’re moving, and what needs attention. Trying to make one page do both jobs produces something too thin for staff and too crowded for the board. Keep two views built from the same underlying data. (If you’re still deciding what to build them in, our guide to choosing a dashboard tool covers that.)
How many measures is too many? No source gives a magic number, and BoardSource’s dashboard book doesn’t either. It stresses preselected, critical measures tailored to each organization. A workable principle: include the fewest measures that let the board answer its own questions, and no more than fit on one page a person can read before the meeting starts. Test each one. If you can’t say in a sentence why it’s there, or it has prompted no question or decision over several meetings, move it to the staff view or an appendix.
Give every measure a comparison and a “so what”
A number alone leaves the reader guessing. Is 410 good? BoardSource’s dashboard book calls for comparative context, so readers can see performance against a benchmark, along with brief explanations.
Pair each measure with one comparison: a target, the plan, last quarter, or the same period last year. Then add a sentence in plain words about what it means and whether the board needs to do anything.
Imagine a food-pantry program that reports “Households served this quarter: 410 (plan: 450).” That tells the board very little. Add “Two Saturday distributions were canceled during a volunteer shortage; recruiting is under way. No board action needed.” Now a board member can read the line, understand it, and move on, or ask a better question. (The numbers in this example are made up.)
Include a few early signals, not only final results
Lagging measures describe what has already happened. Leading measures are early signs that shape what will happen next. The distinction is used widely in performance management. OSHA, for instance, describes leading indicators as proactive and preventive measures, and lagging indicators as measures of events that occurred in the past. That page is about workplace safety, but the idea carries over.
In nonprofit terms: revenue received this year is lagging. Renewals or pledges in progress, and how they compare to last year at this point, are leading. Program completions are lagging. The share of new participants still attending after their first few weeks is an early signal you can act on. A board that sees only lagging measures hears about problems after the chance to change them has passed. Include a small number of early signals where you have a real basis for believing they matter.
Report outcomes, and be honest about activity and small numbers
The Kellogg Foundation’s logic model guide separates outputs from outcomes. Outputs are the direct products of activities, such as workshops held and people enrolled. Outcomes are the changes in participants’ behavior, knowledge, skills, status, or functioning. The guide’s point is that doing an activity is not the same as achieving results from it.
So label activity counts as activity. “People served” and “sessions held” are useful for oversight of capacity, but they don’t show that anyone is better off. Where you have an outcome measure, put it next to the activity count. Where you don’t yet, say so on the page rather than letting effort stand in for impact.
Small programs need extra care. When a program serves 20 people, one person shifts a percentage by five points. Show the count and the total behind every percentage. Avoid arrows and color changes for movements that a single case could cause. Combining periods can help: one Washington State Department of Health standard notes that combining years or collapsing categories improves statistical stability in reporting. A quarter with six exits is a story to tell, not a trend to chart.
Show your work, and protect the people behind the numbers
Trust in a dashboard erodes fast when two people calculate the same number differently. For each measure, keep a short definition note: what counts, what time period it covers, which system it comes from, who updates it, and when it was last refreshed. Put the “as of” date on the page. A shared one-page data note is enough for a small team, and it lets the dashboard survive staff turnover and a funder’s follow-up questions.
If the dashboard shows information about people served, the board almost never needs individual records. Show aggregates. Then watch for small groups, because a count of one or two in a narrow category, such as one participant in a particular age band at one site, can point at a real person. The Washington State Department of Health standard suppresses non-zero counts below ten in its reporting, and warns that suppressed numbers can sometimes be recovered by subtracting from totals, so related cells may need to be hidden too. That is one agency’s rule for its own data. HHS’s HIPAA de-identification guidance says the Office for Civil Rights doesn’t set a universal minimum count and that HIPAA applies to covered entities and their business associates. The right threshold for you depends on your data, your funders’ and partners’ requirements, and any agreements you’ve signed. Settle it with your privacy contact or counsel before publishing, not after.
Make it readable for everyone
Accessibility is good design, and some funders, contracts, or laws may require it. Ask your counsel or funder what applies to you. The W3C’s Web Content Accessibility Guidelines (WCAG) give specific tests you can apply to a web dashboard:
- Color can’t be the only signal. Success Criterion 1.4.1 (Level A) says color shouldn’t be the only visual way to convey information. If you use red, yellow, and green, add a word or symbol: “On track,” “Watch,” “Off track.” The Urban Institute’s style guide suggests patterns such as stripes or hashmarks as additional encodings.
- Text needs contrast. Criterion 1.4.3 (Level AA) asks for at least 4.5:1 for normal text and 3:1 for large text.
- So do the chart marks. Criterion 1.4.11 (Level AA) asks for at least 3:1 for graphic parts needed to understand the content, such as chart lines and bars.
- Text should scale. Criterion 1.4.4 (Level AA) says text can be resized up to 200 percent without losing content or function.
- Charts need a text version. Criterion 1.1.1 (Level A) calls for text alternatives. For charts, that means a short label plus a longer description summarizing the data and trends, with a data table where practical.
For chart choices, keep it plain. The Urban Institute’s guide recommends line charts for trends over time and bar charts for comparing categories, says bar chart axes should always start at zero, and advises keeping pie charts to fewer than five slices. It also suggests labeling data directly instead of using legends, using gray for context and color for the one thing you want noticed, and writing active titles that state the point. For a single measure, a large number with its comparison and one sentence beats a chart.
Set a rhythm and name an owner
Refresh each measure when the underlying data actually changes, and at least as often as the board meets. Refreshing something faster than it is entered only creates work. Put one person’s name on the dashboard: they update it, check the numbers against the source, and write the “so what” lines. Name a backup. Once a year, ask the board which measures led to a question or a decision and which nobody mentioned, and retire the ones nobody used.
A practical first step
Pull your last two board packets. For every chart or table, write down the question it was meant to answer and the decision it informed. Keep the ones with a clear answer. Then sketch a one-page version on paper, with the measure, its comparison, and one sentence for each. Show it to two board members and ask what they would ask about it. That conversation will tell you what belongs on the page.