Decile

A decile is one of ten equal groups that a ranked set of data is divided into, with each decile holding ten percent of the observations. To form deciles, you sort the data from lowest to highest on some measure — customer spend, a propensity score, income, order frequency — and then cut the sorted list into ten slices of equal size. The first decile is the bottom ten percent, the tenth decile the top ten percent, and so on. In marketing and analytics, decile analysis is a common way to rank customers or prospects on a metric and then group them into ten tiers, so you can see how the top decile behaves versus the bottom, and target accordingly. The word can refer either to the groups themselves or to the cut points that separate them, but the everyday sense is the ten equal buckets a ranked population falls into.

Deciles matter because they turn a continuous, hard-to-act-on score into ten clear, comparable tiers that support decisions. Ranking customers by predicted value and splitting them into deciles, for instance, immediately shows where value concentrates — often the top decile or two account for a disproportionate share of revenue, while the bottom deciles contribute little. That pattern guides where to focus retention spend, which prospects to prioritize, and where acquisition or credit thresholds should sit. Deciles also make results legible: it is far easier to say "target the top three deciles" than to reason about a raw score. Because each decile is the same size, comparing behavior or response across deciles is a fair comparison, which is exactly what makes decile analysis a workhorse for segmentation, scoring, and prioritization.

Deciles belong to a family of quantiles — cut points that divide a ranked distribution into equal-sized groups — and they are best understood alongside quartiles and percentiles. Quartiles divide the data into four equal groups of twenty-five percent each; deciles divide it into ten equal groups of ten percent each; percentiles divide it into a hundred equal groups of one percent each. So the difference is simply how finely the ranked population is sliced: quartiles give a coarse four-way split, deciles a middle-ground ten-way split, and percentiles the finest hundred-way split. They answer the same kind of question — where does a value or a group sit in the ranked order — at different resolutions, and you choose among them by how much granularity the decision needs.

The practical choice between them is about detail versus manageability. Quartiles are quick and coarse, useful when four tiers are enough and each needs a decent sample. Percentiles are precise but produce a hundred tiny groups, which can be more resolution than a decision warrants and can leave each group thin. Deciles sit in the sweet spot for a lot of marketing work: ten tiers are granular enough to separate a high-value top decile from the middle and the tail, yet coarse enough that each decile still holds a meaningful share of the population. That balance is why decile analysis is so common in customer scoring and targeting. Confusing the three — treating a decile as if it were a quartile or a percentile — misreads how finely the population has been divided, so it is worth being clear about which quantile you mean.

Using deciles well means ranking a population on a meaningful measure, splitting it into ten equal groups, and using the tiers to guide action — concentrating effort where value or response concentrates, and setting thresholds (target the top N deciles, treat the bottom ones differently) that the decile structure makes clear. It means choosing deciles when ten tiers give the right resolution — finer than quartiles, coarser than percentiles — and choosing a different quantile when the decision needs more or less detail. It means remembering that deciles are relative: the top decile is the top ten percent of this population on this measure, not an absolute standard, so comparisons should stay within a consistent population and metric. Used this way, deciles convert a raw score into a clear, fair map of where a population's value and behavior sit.

The failures are treating a decile as an absolute cutoff rather than a relative tenth of a specific ranked population; confusing deciles with quartiles or percentiles and misjudging how finely the data is split; forming deciles on a poorly chosen measure so the tiers do not reflect what matters; and reading too much into small differences between adjacent deciles. The discipline is to use deciles as ten equal, relative tiers of a ranked population — chosen for the right granularity, built on a meaningful measure, compared within a consistent population — so decile analysis genuinely reveals where value and response concentrate and where to direct effort, rather than dressing up a raw score in ten arbitrary-looking labels.

Decile — one of ten equal groups a ranked dataset is split into, each ten percent — sorts a population into tiers for scoring and targeting, finer than a quartile and coarser than a percentile.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where decile is a core concern: