Sign in

Blog · How-to guides

How to read a norm table: the population, the cell count, and the version

A reading guide for the norm table that every gap on a report depends on: the population column that says which customers the norm was drawn from, the cell count that says how many, the greyed cells under the floor, the method column, median or percentile, the version and date, the norms that moved since last quarter and why, and the test of a good norm table, that a reader can defend any gap on the report by pointing at one row of it.

The short answerA norm table has one row per segment cell and per measure: the norm value, the population it was drawn from, the count of customers in it, the method, and the version and date. Read the count first, because a norm from six customers is greyed and should not be producing gaps. Read the population, which should be the company's own full-relationship customers of that kind and never a survey. Read the method, median or a stated percentile. Read the version and the movement since last quarter, because a norm that moved restates every gap under it. The test is that any gap on any report can be defended by pointing at one row of this table.

Every gap on a report is a customer measured against a norm, and the norm table is where the norm is defended. It is the table most readers never open and the one that decides whether the gaps are real. This guide is the reading order: count, population, method, version and movement, and the test.

The columns

Column What it is Read for
Segment cell Size band, sector, type Which customers this norm applies to
Measure Products per customer, spend per unit of size, share What is being normed
Norm value The figure The number the gaps use
Method Median, 75th percentile, mean with reason How it was computed
Population Main-supplier customers in the cell, trailing year Who it was drawn from
Count Customers in the population Whether it means anything
Floor The minimum count The greying rule
Version, date When computed, under which definition Comparability
Movement Change since last version, and reason Whether gaps moved because the norm did

The reading order

  1. Count against floor. Greyed cells produce no gaps; merged cells say what they merged with.
  2. Population. Main-supplier or full-relationship, from the company's own base; never all customers; never a survey.
  3. Method. Median or a stated percentile; mean only with a reason.
  4. Version and movement. Which norms moved since last quarter, and why.
  5. Norm value. Last; it is what the gaps use, and it is only as good as columns one to four.

A worked read

Cell Measure Norm Method Population Count Floor Version Movement
Mid, manufacturing, direct Products used 4 75th pct Main-supplier, trailing 12 mo 84 30 v2026.3, 1 Jul +0
Mid, manufacturing, direct Spend per site $95,000 Median Main-supplier 84 30 v2026.3 +$4,000: recomputation
Small, services, distributor Products used 2 75th pct Main-supplier 31 30 v2026.3 −1: two customers left the population
Large, energy, direct Spend per site greyed Main-supplier 9 30 Merged into Large, industrial

Count: the energy cell is greyed and merged; its customers' gaps are computed against the industrial norm and say so. The small services cell is one customer above the floor and its norm moved by a whole product because two customers left; read its gaps with care.

Population: main-supplier, trailing twelve months, everywhere. Not circular.

Method: percentile for product counts, median for spend. Stated.

Movement: spend per site rose four thousand at recomputation; every gap in that cell shrank by that much times the customer's sites, and the movements page should say so.

The test

Pick any gap on any report. Find the norm table row it used. If the row's count is above the floor, the population is main-supplier, the method is stated and the version matches the report's, the gap is defended. If not, it is not.

Where it goes wrong

Norm value read alone. A figure with no count behind it.

Population 'all customers'. Circular; every customer at norm.

Greyed cells producing gaps. A list built on six customers.

Movement unrecorded. The gaps moved and nobody knows it was the norm.

Every quarter, the norm table first

Covirage publishes the norm table with every report, with count, population, method, version and movement on each row. The norm guide covers how the rows are computed, and the segmentation guide covers the cells.

Questions people ask

Why read the count first?

Because the count is what makes the norm mean anything. A median of six customers moves when one of them has a bad quarter. The floor, thirty is common, is on the table, and cells under it are greyed and merged upward. A gap computed on a greyed cell is a gap nobody should act on.

What is the population column?

The definition of who is in the cell: customers of this size band and sector where the company is the main supplier, in the trailing year. If the population is 'all customers', the norm is circular, because it includes the customers the gap is measured on. The column should say main-supplier or full-relationship, and the table should say how that was decided.

Why does the version matter here more than elsewhere?

Because a norm recomputed at quarter end moves every gap under it. A share of wallet that fell three points may be the customer or the norm. The norm table's movement column, with the reason, recomputation or method change, is what lets the movements page say which.