Start with the commercial structure
Before capability, before portfolio, before culture fit. How a firm is paid predicts its behaviour more reliably than anything it says about its values, because incentives operate when nobody is watching and values often don't.
- Hourly rewards taking longer. Nobody sets out to be slow, but efficiency reduces revenue, which is a quiet pressure in the wrong direction on every task.
- Traditional fixed price rewards silence. Raising a problem costs the supplier money, so problems surface late, when they're expensive.
- Retainer rewards coasting until renewal approaches. Fine with a real SLA behind it; poor without one.
- Reseller or commission rewards recommending whatever pays best. This is the one most often undisclosed, and the one that matters most.
Ask directly whether they receive any payment from platforms they recommend. The answer, and how comfortably it's given, tells you more than the rest of the meeting.
Twelve questions worth asking
- Do you receive commission, referral fees or reseller margin from any platform you'd recommend?
- What happens to the budget if you discover something you didn't anticipate?
- Who owns the code, the documentation and the credentials during and after?
- If we ended this tomorrow, what would we have and how quickly?
- Who specifically would do the work, and will they still be on it in six months?
- What's the last project you told a client not to do?
- What would make you turn this down?
- How do you handle the parts of the scope we haven't thought of yet?
- What does this cost to run, per year, after you've finished?
- Can we speak to a client where things went badly?
- What do you not do, and who would you send us to instead?
- If we wanted to take this in-house in two years, how hard would you make that?
What the answers actually tell you
"We'd absorb that"
On unanticipated discovery, this is the answer you want, and it should come with a limit. A firm that absorbs everything without qualification is either not serious or will recover it elsewhere. A firm that absorbs nothing has made discovery your risk entirely.
Hesitation on ownership
Ownership of code, credentials and documentation should be immediate and unambiguous. Any hesitation, or any answer involving their proprietary framework, is a lock-in signal. It doesn't make them wrong to work with, but price it.
No example of saying no
If a firm cannot name a project they turned down or advised against, either they haven't, or they don't remember, and neither is encouraging. The specific example matters more than the sentiment.
A vague answer on who does the work
Selling teams and delivery teams differ everywhere; that's normal. What matters is whether you meet the delivery people before signing and whether continuity is committed to.
An immediate proposal
A detailed proposal arriving after one conversation means either the scope was assumed or the proposal is generic. Both are worth probing.
Flags worth taking seriously
- Undisclosed platform relationships — ask directly rather than assuming
- A fixed price arriving without any discovery, on anything non-trivial
- Reluctance to discuss what happens after launch
- No written exclusions in the scope, only inclusions
- Pressure tied to their calendar rather than your deadline
- Case studies with no numbers, or numbers with no methodology
- An inability to explain why a previous project went badly
How to use references properly
Reference calls are usually wasted because the questions are too polite. Three that produce real information:
- What did they get wrong, and how did they handle it? — every project has one
- What would you have done differently in how you engaged them?
- If you were starting again, would you use them for the same piece of work, or a different one?
Ask for a reference where things went badly. Most firms have one and the good ones will give you it. Refusal is itself an answer.
