Share of wallet
How much of a customer's spending in your category you actually hold.
01 What it is
Why anyone looks at this number
How much of a customer's spending in your category you actually hold.
Retention tells you they stayed. Share of wallet tells you whether they buy most of what they need from you. Growth inside the existing base is almost always cheaper than the equivalent acquisition.
02 The formula
How it is worked out
Share of wallet = your_revenue_from_the_account ÷ the_account_total_category_spend grain : year × account × category unit : percentage source : CRM and billing, joined to an external or modelled estimate of account size
The denominator never lives in your systems. It comes from a survey, a panel, trade data or a model — so the method has to be stored next to the number. An estimate that changes source between years produces a trend that is entirely artificial.
03 Worked example
The same number, with real inputs
| Revenue from the account | €1.9m |
| Estimated category spend | €8.6m |
| Retention | 98% |
| Calculation | 1.9 ÷ 8.6 |
| Result | 22% |
A loyal account that buys a fifth of what it needs from you. The growth here is not a new logo, it is the other €6.7m already being spent somewhere else.
04 What moves it
Four things that actually change this number
Range coverage
Categories where you are simply not on the list.
Contract structure
Framework agreements that lock in a share, or cap it.
Service performance
OTIF and quality decide whether the second category is ever offered.
Buying centralisation
A group that consolidates purchasing can double or halve your share.
05 Where the number lives
The system, the record and the fields
| System of record | Key record | Fields you need |
|---|---|---|
| CRM and billing, joined to an external or modelled estimate of account size | Account, at group level rather than site level | account_id, parent_id, revenue_ltm, category, estimated_spend, estimate_source, estimate_date |
The denominator never lives in your systems. It comes from a survey, a panel, trade data or a model — so the method has to be stored next to the number. An estimate that changes source between years produces a trend that is entirely artificial.
06 How it goes wrong
Three ways this metric misleads people
Treating retention as share
A 98% retained account can still be a 20% share account, and the plan never notices.
Fix: Report both, side by side, for the top accounts.An undocumented denominator
The estimate changes method and the whole trend line moves for no business reason.
Fix: Version the estimate: source, date and method, stored with the number.Averaging across the base
A portfolio average hides the accounts where a plan would actually pay.
Fix: Work it account by account for the top decile and forget the average.08 Questions
Frequently asked
How do we estimate the denominator credibly?
Three routes, in order of cost: ask the account directly in a review, buy panel or trade data for the category, or model it from firmographics and observed spend in comparable accounts. Record which one you used.
Is share of wallet worth measuring for small accounts?
Rarely one by one. Model it at segment level for the long tail and measure it individually only where a named plan would follow.
One definition, everywhere it is used
SCIKIQ stores this metric once and serves it to every dashboard, board pack and agent that asks.