Every penetration, fit, share or whitespace figure needs a denominator: what a customer like this one usually buys. This guide sets out how to build that norm from your own customers by banding them on type and size, the majority rule, the minimum band size, and how to record overrides so the gap list stays credible.
Every article on this site that measures penetration, fit, whitespace or range has the same quiet dependency: the set of products a customer like this one usually buys. Get that denominator wrong and every gap list is either trivially long or empty. This guide sets out how to build it from the firm's own customers, which is the only source that survives being questioned by a salesperson.
A customer's gap is the difference between what customers like them buy and what they buy. "Like them" means the same type and roughly the same size. What such customers buy is a fact in the firm's own order history. So:
Type is whatever the sales team already uses: sector, channel, kitchen type, contractor type, client segment. Size is a band on revenue, headcount, assets or beds, whichever the firm has. Two dimensions are usually enough; three makes bands too small.
| Band | Customers |
|---|---|
| Mid-market manufacturer, 50 to 250 employees | 84 |
| Mid-market manufacturer, 250 to 1,000 | 41 |
| Small contractor, under 50 | 312 |
| Regional bank, under 5bn assets | 9 |
The last band is too small. Merge it with the next size up, and note the merge on the report.
For each band and product, the share of customers buying it in the trailing period. Products above the threshold are in the norm.
| Product | Share of band buying | In norm at 50%? |
|---|---|---|
| Bearings | 96% | Yes |
| Seals | 71% | Yes |
| Couplings | 58% | Yes |
| Service | 44% | No |
| Gearboxes | 12% | No |
A customer in this band buying bearings and seals has one gap: couplings. Service is not a gap under this norm, whatever the service team hopes.
Product marketing will want service in the norm. The rule is that it can be, as a labelled override: "Service added to the mid-market manufacturer norm by product marketing, 1 September 2026, because of the new service contract launch." The gap list then shows service gaps with the label, and the field can see which gaps are majority-driven and which are strategy-driven.
Norms drift as the customer base changes. Recompute on a schedule, quarterly is usual, and keep the history, so a gap that appeared because the norm changed can be explained as such.
Bands too small. The most common failure. Merge until the majority means something.
The whole catalogue as the norm. Every customer has gaps, none of them credible. The majority rule exists to prevent this.
Norm from one big customer. A band dominated by one customer's unusual purchasing is that customer's norm. If one customer is more than a quarter of a band's revenue, band by count, not by revenue, and consider excluding it as an outlier.
Reference customers ignored. Customers above the norm are the proof the norm can be exceeded. Show them to the field as examples, not as anomalies.
The norm is the denominator in the SaaS whitespace, commercial banking cross-sell, industrial distributor penetration and manufacturer catalogue fit guides, and in every other fit measure on this site. Covirage computes it from the uploaded rows and shows the bands and thresholds on the report. The glossary entry on penetration has the short version.
Industry tables describe an average firm selling an average range. Your norm describes what your own customers actually buy from you, which is the only thing a gap list can credibly promise. External data is useful for wallet share, where the total is outside your systems; for fit it is unnecessary.
Big enough that the majority rule means something. Twenty customers is a comfortable minimum; below ten, merge with the nearest band and say so. A band of four has no norm.
Half is the honest default: a product most of the band buys. A lower threshold produces longer gap lists with weaker claims; a higher one produces shorter, stronger lists. Pick one, apply it everywhere, and let the field's feedback move it.