What B2B Software Buying Committees Actually Look Like, Category by Category

"B2B software is bought by committee" is true, repeated constantly, and almost useless as written. It tells a vendor to expect more than one person and nothing about who those people are, what each of them is actually weighing, or which of them can stop the deal without being able to start it.

The composition is not generic. It varies by category in ways that are stable enough to plan against, and we publish the room for each of the categories we cover. Read across them and some patterns hold everywhere, while others reverse completely.

The seat that decides is rarely the seat that signs

This is the most consistent finding across categories, and the most commonly ignored.

In recruiting software, a Head of Talent signs. The recruiters and coordinators who live in the system daily decide, because their resistance kills a rollout regardless of what leadership chose. In endpoint security, a CISO selects, but IT operations can veto on agent footprint alone — endpoint performance complaints land on the help desk, and that is enough to stop a deployment. In revenue intelligence, a CRO buys and reps determine whether it survives, because a product they route around produces data nobody trusts.

The practical consequence: a vendor who sells exclusively to the economic buyer is selling to the person who can say yes and not to the person who can quietly make yes meaningless.

Finance arrives at different moments

Finance is in almost every room. When it arrives is the variable.

In AP automation finance is the room — controller, AP manager, CFO — and the conversation is theirs from the first call. In endpoint security finance and procurement are increasingly present early, as security spend moves out of the security budget and into a scrutinised line. In observability, finance arrives late and unhappily, usually after a bill has jumped, which is why more evaluations in that category start with an invoice than with an outage.

A vendor who leads with ROI in a category where finance arrives last is answering a question nobody has asked yet. One who avoids pricing in a category where finance is already in the room looks evasive.

Some rooms are smaller than the category's reputation

Compliance automation is the clearest example. It sounds like an enterprise purchase and is usually not: a company chasing its first SOC 2 typically has a founder or COO, whoever has been nominated to own compliance, an engineering lead, and sales leadership — who are in the room because a blocked deal is what started the project. There is frequently no security function at all, which is precisely why the category exists.

Fleet telematics is similarly compact at the operating end: an operations or fleet manager, a safety manager, an owner or CFO, and drivers whose influence is entirely indirect and entirely decisive.

Selling an enterprise motion into a four-person room slows the deal and signals that you have misread the buyer.

Some rooms grow while you are in them

ERP is the extreme case. It starts with a CFO and a controller and accumulates a head of financial systems, IT and security, operations, and divisional leads in a multi-entity group. Each addition brings a requirement, the business case stops matching the quote, and the project goes back for re-approval. This is why ERP evaluations die in the middle rather than at the end, and why a quiet period is more often scope creep than a competitive loss.

HRIS grows in a more predictable direction: HR owns it, but payroll and — the moment a second country appears — legal and data protection become gates rather than stakeholders. The category where this matters most is the one where a single new jurisdiction can disqualify most of a shortlist.

The seat nobody invites

Several categories have a participant who is never in the meeting and routinely determines the outcome.

Drivers in telematics, who will accept vehicle tracking and resist an inward-facing camera, and in a unionised fleet turn the rollout into a negotiation. Hiring managers in recruiting software, whose unwillingness to review candidates in yet another system quietly ends adoption. Reps in revenue intelligence. Application owners in identity management, who decide whether "supported" means supported without custom work.

These are implementation risks that present as buying signals. A deal that closes and never deploys looks like a win in the pipeline and is not one.

How to use this

Three things, none of which require new tooling.

Map the room before the first call, by category rather than by account. The ERP room and the compliance automation room are not the same shape, and a discovery call built for one wastes the other's time.

Work out who can veto. In most categories the blocker is operational — agent footprint, daily usability, integration effort — and sits outside the function you are selling to. Addressing it unprompted is more persuasive than any feature.

Ask what else they are looking at, and listen for who said it. A shortlist named by a practitioner is a different signal from one named by a CFO. The first is a real evaluation; the second is often a budget exercise.

We publish the committee for each category we cover, alongside what starts an evaluation there and which vendors buyers actually put side by side. It is the same information our own programmes run on, and it is more useful to a vendor before the first call than after it.