How to brief a consultancy

Engagements that go badly are often traceable to the brief. A brief that specifies a solution gets that solution built, competently, whether or not it addresses the actual problem — and the firm has no standing to say so once it is the agreed scope.

Describe the problem, not the fix

"We need to move to Kubernetes" is a solution. "Deployments take two hours and only one person can do them" is a problem, and it admits several answers, some of which are far cheaper.

You will get better advice by withholding your preferred solution until after the first conversation.

Say what the constraints really are

Budget, deadline, the system that cannot go down, the team that is already at capacity, the audit in March. Constraints shared at the start shape the plan. Constraints discovered in week six invalidate it.

Withholding the budget to avoid anchoring the price mostly produces proposals aimed at the wrong scale.

Start with an assessment

A short, paid assessment with a written deliverable is the cheapest way to find out both what the problem is and whether you want to work with these people. It should be useful even if you then hire somebody else — and if the firm resists that framing, you have learned something.

Agree what handover means

Define it concretely and early: documentation, a walkthrough, your team making the next change unaided. An engagement that quietly becomes permanent is usually a handover nobody specified rather than anyone acting in bad faith.

Ask what they would decline

Ask what work they turn down and when they would tell you not to do something. A firm that has never declined anything is describing a sales process rather than a point of view.