How to run a vendor selection that the executive team and the board can defend two years later.
Vendor selections are the decisions boards most often wish they could make again. The cycle is depressingly familiar. A vendor is engaged. Implementation begins. Twelve months later, the organisation is in a difficult relationship with a vendor it cannot easily exit, the contract is being renegotiated, and someone is asking how we ended up here.
This article describes what a defensible vendor selection looks like. It draws on selections we have led or reviewed across platforms, managed services, consulting and implementation partners. The principles are similar regardless of category.
The single most common cause of vendor selection failure is that the strategy is implicit rather than explicit. The vendor is being chosen because their pitch deck matched an unstated vision in the head of the executive sponsor, not because the vision has been articulated, agreed and costed. When conditions change — and they always change — the strategy is renegotiated in the contract, and the contract becomes the only place where the strategy lives.
The fix is unglamorous. Before any vendor is engaged, the strategy needs to be written down in plain language, signed off by the executive, and translated into a set of capability requirements that the vendor selection can be measured against. This is usually a two-month piece of work. It feels like a delay. In our experience it is the only thing that saves the eighteen-month delay that follows when the wrong vendor is selected.
The most common mistake in vendor selection is to build a matrix against the vendors on the shortlist rather than against the capabilities the client actually needs. A capability matrix forces the conversation back to "what do we actually need from a vendor" rather than "how do we rank the three people in the room."
Every capability matrix we build has three layers. The first is a published list of capabilities the executive has agreed the vendor must demonstrate. The second is the weighting — explicit, signed off in advance, and not negotiable after the vendor presentations have been heard. The third is the scoring rubric — a written definition of what good, adequate and inadequate looks like for each capability.
If the scoring rubric cannot be written down, the capability cannot be assessed fairly. If the capability cannot be assessed fairly, it should not be in the matrix.
Most reference checks are theatre. The vendor provides three customers. The three customers are at a similar maturity to the client and have had good experiences. The reference call confirms the vendor is good at selling themselves.
Better reference checks start with the client defining what they want to learn. Then the vendor provides references against a brief, and the client chooses who to call based on relevance. Then the reference call is structured — same questions in the same order — and the notes are written up in a consistent format. Then, and this is the part most organisations skip, the client asks the reference for the names of two further customers who had less positive experiences. Those calls are where the truth lives.
Vendor commercials have become more sophisticated than most procurement teams can read. Outcome-based pricing, gain-share arrangements, consumption-based billing, hybrid models — these are all genuinely useful innovations. They are also genuinely difficult to compare across vendors, and they shift risk in ways that are not always obvious in the contract summary.
The defence is to insist that every shortlisted vendor presents its commercial model in the same format, against the same workload profile, over the same contract term. The vendors will resist this — it removes one of their principal levers of differentiation. The defence to that resistance is that the client is entitled to compare like with like. If a vendor's commercial model cannot be expressed in a common format, it is probably too complex for the client to manage.
Most vendor relationships fail not because the vendor was wrong but because the contract was incomplete. Service levels were vague. Acceptance criteria were undefined. Change control was unspecified. Termination rights were impractical. The contract was signed because the relationship felt good. Two years later, the relationship is no longer good, and the contract cannot help.
We are usually brought in to help clients who are about to renegotiate a contract that should never have been signed in its current form. The conversation then becomes painful and expensive. The right time to ask the hard questions about the contract is before it is signed. We have a checklist of contract provisions we believe should be non-negotiable. We are happy to share it.
The board is usually told what was decided, not how it was decided. This is a mistake. A board that has approved a vendor selection on the basis of a recommendation has very limited ability to defend the decision if it turns out badly. A board that has approved a vendor selection on the basis of a defensible process has much more.
The most defensible board paper describes the process before it describes the decision. It explains the strategy the vendor was selected against, the capability matrix that was used, the reference checks that were performed, the commercials that were modelled, the contract terms that were negotiated and the residual risks that remain. The decision then becomes almost self-evident, and the board has the evidence it needs to defend the decision if it is ever challenged.
Vendor selection is one of the highest-leverage activities a technology executive engages in. It is also one of the most vulnerable to the pressure of time, the influence of relationships and the appeal of a confident pitch. An independent perspective early in the selection process almost always produces a better outcome than one brought in late to fix a difficult situation. If you are about to run a selection and would value a sounding board, we would be glad to talk.