Index numbers: comparing categories of very different sizes

How do you put a category selling $500 a week next to one selling $50,000 a week on the same chart? Plot them in dollars and the small one is a flat line near zero. Index numbers solve this. In retail, an index expresses each value relative to a chosen base, set at 100, so categories of any size can be compared on how they changed.

The idea is familiar from the Consumer Price Index, which tracks prices against a base period rather than in raw dollars. The same logic works for sales, units, prices or store counts.

How is an index calculated?

Divide each period's value by the base period's value, then multiply by 100. The base period always equals 100. A value of 110 means ten units above the base on a scale of 100. A value of 95 means five below.

Here's a hypothetical with two categories. Category A sells $500 in the base week, then $550, then $600. Category B sells $50,000, then $51,000, then $52,000. In dollars, B's change dwarfs A's. As indexes, A goes 100, 110, 120 and B goes 100, 102, 104. Now it's clear that A grew faster relative to its own size, even though B added far more dollars.

That's the key point. An index shows relative change. It says nothing about absolute size.

How do you choose a base period?

Pick a period that's normal for everything you're indexing. If the base week for one category contained an unusual promotion or a supply problem, every later value for that category will look distorted. A single abnormal base can make steady sales look like a decline or a surge.

Many analysts avoid this by using an average as the base, such as the mean of a full quarter or a full year. That smooths out one-off events in the base and gives a sturdier starting point.

The base should also be recent enough to be relevant. An index based on a period three years ago mixes old changes with new ones. When the base gets stale, it's common to rebase: choose a new base period and recalculate so it equals 100 again.

What can an index hide?

Size, as noted. A tiny category can post dramatic index swings from small dollar changes. Going from $100 to $150 is an index of 150, which looks remarkable and may mean three extra units sold. Always keep the underlying values available, and consider noting them next to the index.

Composition changes can hide inside an index too. If the category gained new items between the base and the current period, some of the index rise comes from a bigger assortment, not from existing items selling more. That can be the right answer to some questions and the wrong one for others.

When should you use an index instead of shares?

Shares tell you how a total is divided at one point in time. Indexes tell you how each part changed over time. If the question is "which categories are growing fastest?" an index fits. If it's "how big is each category relative to the others?" shares fit better.

Regional comparisons often use a related idea: a location's value against a benchmark, with 100 meaning equal to the benchmark. The same caution applies. The benchmark must be relevant, and the underlying volumes should be visible.

NRS Insights reports compare categories month to month, which makes relative measures useful. The July 2026 same-store sales report is a good place to practice reading change by category.