Start with the problem, not the shortlist
Most partner selections begin by collecting suppliers and comparing them. That ordering makes it very hard to tell the difference between a good fit and a good salesperson, because there is no yardstick yet.
Write down what you are trying to achieve, what constraints are real, and what would count as success twelve months in. Two paragraphs is enough. It will change how you hear every conversation that follows, and it will expose the suppliers who never ask about it.
Judge the process, not the portfolio
Portfolio work tells you what a supplier's best team produced under their best conditions, often years ago, and rarely tells you who did it. It is close to uninformative about what your project would be like.
What predicts your experience is how work is run: how requirements are handled, how progress is made visible, how disagreements are resolved, and what happens when something slips.
- How will I see progress between meetings, and in what form?
- What happens in week one, and what will exist at the end of it?
- Who decides when something is finished?
- How is a change in scope priced and agreed?
- What does your handover include if we part ways?
Ask about a project that went badly
This is the single most informative question available, and the answer is difficult to fake. Every supplier with real history has had a project go wrong. One that claims otherwise is either new or not being straight with you.
Listen for specifics and for ownership. A strong answer names what went wrong, what their part in it was, and what they changed as a result. A weak answer blames the client, blames a third party, or stays abstract.
Meet the people who would do the work
Pitches are frequently delivered by senior people who will not be on your project. This is not necessarily deceptive, but it does mean you have assessed the wrong individuals.
Ask to meet the proposed lead engineer and delivery lead specifically, and ask what else they are committed to during your timeline. A named team with visible availability is worth more than an impressive agency brand.
Test communication before you sign
The evaluation period is a live sample of what working together will be like. Slow, vague, or evasive answers now do not improve after a contract is signed; they are the best behaviour you will see.
Send a genuinely difficult technical question and see whether the answer engages with it, admits uncertainty where it exists, or retreats into generality.
Settle ownership and exit before price
Who owns the code, the infrastructure, the accounts, and the documentation should be established in writing early, and the answer should be you, from the first commit.
Ask directly what happens if you leave in six months: what you receive, in what state, and at what cost. A supplier confident in the relationship makes this easy. Resistance here is the clearest warning sign in the entire process.
Price the whole engagement
Day rates are close to meaningless in isolation. A more expensive team that needs fewer people and less rework is routinely cheaper in total, and a low rate frequently comes with a larger team, longer ramp-up, and more management overhead on your side.
Compare the total cost of reaching a defined outcome, and include your own team's time — which is real, and is almost always excluded from the comparison.
Start small if you can
A discovery sprint or a contained first deliverable tells you more than any evaluation process, because it is the real thing at low stakes. You learn how they estimate, how they communicate under pressure, and how their work holds up to review.
Where a small start is not possible, at least ensure the first milestone is early and genuinely verifiable, so a mistake is recoverable.