Most bad agency relationships were predictable during the first call. The warning signs were there; the buyer simply did not know which questions surface them.
Here are nine, in the order we would ask them, along with what a good answer sounds like and what should worry you.
1. Who will actually do the work?
Good answer: named people, with their role and experience, available to meet before you sign.
Worry if: you meet a charismatic salesperson and are told the "delivery team" will be assigned later. The gap between the people who win work and the people who do it is the single most common source of disappointment in this industry.
2. Can I see a project that did not go well?
Good answer: a specific story, what went wrong, what they changed afterwards.
Worry if: they cannot think of one. Everybody has had a difficult project by year three. An agency that presents an unbroken record is either very new or not being straight with you.
3. How do you price, and what happens when scope changes?
Good answer: a clear model — fixed price or hourly — and a defined written process for change requests.
Worry if: the answer is vague, or if scope changes are described as "we will sort it out as we go." That sentence is where budget overruns are born.
4. Who owns the code, the files and the accounts?
Good answer: you do, on final payment, including source files and administrative access.
Worry if: they host on a proprietary platform you cannot leave, hold the domain in their own account, or become evasive. Ownership is the difference between a supplier and a hostage situation.
5. What do you need from us, and when?
Good answer: a specific list — content, approvals, access — with dates, and a clear statement of what happens if you are late.
Worry if: nothing is asked of you. Projects fail on client-side delay at least as often as on agency-side delay, and an agency that has not planned for it has not run many projects.
6. How will we know if this worked?
Good answer: a metric agreed before starting, with a baseline measured before any changes.
Worry if: success is described in terms of deliverables ("a beautiful new website") rather than outcomes. Deliverables are what you buy; outcomes are why you bought them.
7. What happens after launch?
Good answer: a defined support window included, and a clear optional plan afterwards.
Worry if: support is unmentioned until you ask, or is only available as an expensive open-ended retainer.
8. Can I speak to a client from two years ago?
Good answer: yes, and here are two.
Worry if: every reference is from a project that finished last month. Recent clients are still in the honeymoon phase. The interesting question is whether the work held up.
9. What would you tell us not to do?
Good answer: something specific about your situation that costs them revenue to say.
Worry if: everything you propose is enthusiastically endorsed. An agency that agrees with all of your ideas is selling, not advising.
Two things that matter less than you think
Agency size. A four-person team and a forty-person team can both deliver well. What matters is whether the senior people touch your project, and small teams often score better on that.
Portfolio beauty. Portfolios are curated and frequently include work the agency did not lead. A single detailed case study with a measured result tells you more than thirty attractive screenshots.
The simplest test
Ask for a written scope before signing anything. An agency that can describe precisely what you will receive, by when, for how much, has thought the project through. One that cannot has not — and you will be paying for that thinking later, in change requests.
We publish our own scoping process and our pricing on the
and four detailed case studies — including what we got wrong — in our
