What is category management?

Category management is a way of organising procurement around groups of related third-party spend rather than around individual purchases or individual suppliers. Each category is given an owner, a fact base, a cross-functional decision forum and a multi-year plan covering price, specification and process. It is a governance discipline, not a sourcing event.

Where the term comes from, and why that matters

Category management did not start in procurement. It started in grocery retail. Brian F. Harris, then director of the Food Industry Management Program at the University of Southern California, originated the concept in the late 1980s. The eight-step model that became the industry standard was codified in the early to mid 1990s through the Efficient Consumer Response programme, with Harris and his firm The Partnering Group central to it. Retailers used it to manage a shelf: define the category, define its role in the store, set targets, then decide assortment, pricing and promotion.

Procurement adopted the vocabulary in the 1990s and largely stopped citing the source. That is why the “four Ps” framing carried by some published procurement guidance still includes promotion, which means nothing when you are buying maintenance services.

The origin is not trivia, because what transfers is the method rather than the object. Retail category management looks at what a business sells and runs a category for revenue and margin on the shelf. Procurement category management looks the other way, at what a business buys, and runs a category for cost, value and risk. The mechanics survive the switch intact: group related spend, give it an owner, build a fact base, agree a target position, work to a plan.

That is also why the discipline transfers across industries and across category types with very little adaptation. It applies to direct materials in a manufacturing business and to professional services in a financial one, to goods and to services, to capital and to operating expenditure. What changes between them is which levers are available and who has to agree to them. The way the category is run does not.

What actually counts as a category

A category is a group of spend that faces a common supply market and can therefore be acted on through a common set of decisions.

That definition does real work. It rules out grouping by internal budget line, which is how most first attempts go wrong. “IT” is a budget, not a category: laptops, telecoms and software licensing face three unrelated supply markets, three different supplier populations and three different negotiating positions.

A workable category structure runs three to four levels deep. Below the fourth level the distinctions stop being actionable and start being clerical.

Category management and strategic sourcing

These two are used interchangeably across most published material. They are not the same thing, and they are not alternatives either. They sit at different altitudes, and the relationship runs one way: category management sets the direction, and sourcing events deliver it.

A category strategy states where a category should be in three to five years and what has to change to get there. Where a category has been handled loosely, the objectives are recognisable: streamline the supplier base, reduce the assortment, take working capital out, shorten lead times, move the specification. Those are strategic objectives, not activities.

Delivering them takes sourcing events, usually several, often sequenced across years. Which events, in what order, is itself part of the strategy.

	Category management	Strategic sourcing
Answers	Where should this category be, and what has to change	How do we deliver one of those changes
Unit of work	A category, continuously	An event, scoped to one or more of those objectives
Horizon	3 to 5 years, ongoing	Weeks to months, then it ends
Output	A target position and a route to it	A delivered change: a contract, a consolidated base, a reduced assortment, shorter terms
Relationship	Sets the objectives	Realises them, one or several events at a time

A simple way to tell which one you are looking at: ask someone to describe a category without naming a supplier event.

If they can tell you where the category should be in three years, what has to change to get there and roughly in what order, there is a strategy, and the events sit underneath it. If the only thing they can describe is the next tender, the organisation has a sourcing calendar. A sourcing calendar is worth having. It is not the same thing.

The difference shows up in a quiet year. A category under management still moves when no event is running: specifications get reviewed, demand gets challenged, the supplier base gets tidied, terms get prepared ahead of renewal. A category on a calendar sits still until the next event falls due.

A sourcing event is not one thing

The common mistake, and one a good deal of published material encourages, is to read “sourcing event” as a synonym for “tender”. It is a family of approaches, and choosing between them is where the return is decided. Some examples on types of sourcing events (list is not exhaustive):

• Competitive tendering. Contest the market. Right where the supply market is genuinely contestable and the requirement can be written down.
• Structured renegotiation. Tender-grade rigour at conversational speed, run across a batch of suppliers. Right where switching is impractical but the terms have drifted away from the market.
• Category optimisation. The deepest of them. Puts specification, demand and process in scope alongside the commercial terms. Right where the value sits outside price.
• Demand and consumption work. Reduce what is bought rather than what is paid per unit. Frequently the largest single lever, and the one most often skipped.
• Supplier-driven improvement and partnership development. Right where the supplier holds knowledge you do not and the relationship can carry a joint agenda.

Every one of these works on total cost of ownership rather than unit price. What separates them is depth and which levers they put in scope, not whether they are commercial. A category optimisation event reviews demand, specification, supplier base and price. A tender against a fixed requirement reviews supplier and price. Both are sourcing events, and both can change what you buy, not only what you pay.

Which is also why “procurement-led” does not survive contact with practice. Any approach that challenges specification, demand or service level needs the business in the room with the authority to decide. That holds in a renegotiation, in a tender where the requirement is being rewritten, and most of all in a category optimisation. Procurement runs the process. It does not own the decisions.

The three families of levers

Category management works on total cost of ownership rather than unit price. The levers fall into three families.

• Commercial: competition, volume aggregation, contract structure, payment and delivery terms, index and currency clauses.
• Technical: specification, standardisation, variant reduction, material substitution, design, including also e.g. make vs. buy decisions.
• Process: demand control, ordering behaviour (e.g. MOQs), consumption, compliance, inventory and service levels.

The technical family is what separates the deeper approaches from a price negotiation, and it runs in both directions. The right answer is sometimes a plainer specification. It is sometimes a more demanding one, where a higher specification lowers lifetime cost through longer service intervals or fewer failures. Anyone who only ever specifies downwards is running a cost-cutting exercise with a category management label attached.

The precondition almost nobody writes down

Two of the three lever families take value out of somebody else’s budget. A plainer specification is a loss for the function that wrote it. Fewer variants is a loss for whoever chose them. A longer lead time is a loss for operations.

Category management therefore requires a decision forum willing to accept a local cost for a group gain. In practice that means three things: a sponsor positioned above the affected functions, a forum that decides rather than one that is updated, and the specification owner and the budget holder present when the decision is taken.

Secure this at the start, before the fact base is built. Where it is not secured the pattern is predictable. The commercial levers survive, and the technical and process levers are withdrawn late, after the analysis has already been paid for.

Why organisations rarely do it well

The barriers are structural rather than personal. Category work competes with the renewals already in the diary, and the diary usually wins. Where the capability exists it is thin and concentrated in a few people. And the mandate, written or unwritten, is frequently commercial only, which removes two of the three lever families before anyone begins.

The evidence supports the pessimism. Research by Future Purchasing with Henley Business School found in 2024 that 68% of available value remains uncaptured across cost, risk and stakeholder value.

Does any of it reach the P&L?

This is the question a CFO asks, and we have not found it answered in the widely read practitioner material.

A negotiated saving is not a realised saving. A CFO Leadership survey reported by BCG in January 2026 found that only 38% of CFOs have high confidence that procurement savings reach the P&L, falling to 29% in large enterprises. The underlying methodology is not published. McKinsey has put the gap between reported procurement savings and P&L impact at up to about 20%.

The remedy is not better negotiation. It is a savings definition finance has agreed in advance, a baseline that survives changes in volume and mix, and an explicit rule for when a saving counts. We work to five stages: estimated (a percentage applied to a spend base), validated (finance has agreed the baseline), identified (a named supplier and a priced delta), signed off (approved against that baseline, with a run rate and an effective date), and realised (the money has left the cost base). Only the last two carry weight with a CFO.