Percentiles and Quartiles

Your development director walks into a meeting and says "we need to focus on our top donors." Everyone nods. But what does "top" mean? The biggest 100 names? Anyone who gave more than €500? The top 10%? These are all different groups, and the strategy for each one is different. You need a way to describe where someone falls relative to everyone else. That's exactly what percentiles do.

A percentile tells you what percentage of the data falls below a given value. If a €200 donation is at the 90th percentile, that means 90% of all donations were less than €200. The donor isn't necessarily giving a lot in absolute terms, but relative to everyone else, they're near the top. This is a different question than "what's the average?" which we covered in Day 1. The mean and median tell you about the center. Percentiles tell you about position.

You already know one percentile without realizing it. The median, from Day 1, is the 50th percentile. Half the data sits below it, half above. Percentiles just extend that idea to any cutoff you want. The 25th percentile, the 75th, the 99th.

Four percentiles come up so often they have their own name. The quartiles divide your data into four equal groups. The first quartile (Q1) is the 25th percentile. The second quartile (Q2) is the median, the 50th percentile. The third quartile (Q3) is the 75th percentile. Together they split your donors into four buckets of roughly equal size: the bottom quarter, the lower-middle, the upper-middle, and the top quarter.

The distance between Q1 and Q3 is called the interquartile range, or IQR. It captures the middle 50% of your data and gives you a sense of spread that isn't thrown off by extreme values. In Day 2 we saw how standard deviation measures spread. The IQR does something similar, but it ignores the extremes entirely. If your donor list includes one gift of €50,000, the standard deviation will be enormous. The IQR won't flinch.

This matters whenever you need to segment people. Donor management is the obvious case. Quartiles give you a natural way to tier your donors: Q1 is your entry-level base, Q2 is your developing middle, Q3 is your committed core, and above Q3 is your major gift territory. Each group needs a different communication strategy and a different ask amount. If you set your "suggested upgrade" at the 75th percentile of someone's current quartile, you're asking them to move up in a way that's ambitious but grounded in what their peers are actually doing.

Percentiles also show up in grant reporting when funders ask about the distribution of outcomes, not just the average. Saying "the median participant improved by 8 points, and the 75th percentile improved by 15 points" paints a much richer picture than a single mean. It shows that your program works for most people and works especially well for some.

In survey analysis, percentiles help you spot ceiling and floor effects. If your post-program satisfaction survey has a 90th percentile score of 4.8 out of 5, your scale might not have enough room at the top to capture real differences among your most satisfied participants.

Salary benchmarking is another common use. When your board asks "are we paying our executive director competitively?" the answer is usually framed in percentiles. The 50th percentile of comparable salaries is the market midpoint. The 75th percentile is what you'd pay to attract top talent in a competitive hiring market.

The mean tells you the center. Percentiles tell you where you stand relative to everyone else. When you need to rank, segment, or compare, percentiles are the right tool.

See It

Hover over any bar to see its percentile. Click the quartile buttons to highlight each quarter of the donor base.

Reflect

Does your organization segment donors into tiers? If so, how were the cutoffs chosen? Were they round numbers someone picked, or are they based on the actual distribution of giving? What would change if you used quartiles instead?

When you report program outcomes, do you ever share the range or just the average? What story might the 25th and 75th percentile tell that the mean hides?