Founders default to asking for a project, even when what they actually need is a retainer — because a project sounds more concrete and easier to justify to a board. That instinct is understandable and often wrong. Here's the distinction I actually use when someone asks which one fits.
A project has a defined answer waiting at the end
A project engagement works when you can name the deliverable before we start: a QA architecture assessment with a written report, a due diligence read before a term sheet, a specific automation framework built into your repo. The scope, timeline, and "done" state are all fixed up front. If your question is "what's wrong with my release pipeline and what do I do about it," that's a project — it has a clear finish line.
A retainer works when the question keeps changing
A retainer is the right shape when what you need isn't a report, it's a second opinion that shows up before decisions get made instead of after they've already gone wrong. Architecture calls, a hire you're unsure about, a vendor pitching you an approach that sounds plausible but you can't independently evaluate — these don't arrive on a schedule, and a project engagement with a fixed scope can't flex to catch them.
The tell that you need a retainer, not a project: you keep having the same category of decision recur every few weeks, and each time you're either guessing or waiting for someone with the right context to become available.
Where founders get this wrong
The most common mistake is buying a project to solve what's actually a recurring-decision problem. You get a clean report, implement the recommendations, and three months later you're back in the same uncertainty — not because the report was wrong, but because the underlying need was ongoing judgment, not a one-time diagnosis. The second most common mistake is the reverse: paying for an open-ended retainer when the real need was one specific, bounded question that a two-week project would have answered for less money and more clarity.
A rough test
Ask: if I get the right answer today, does the need go away, or does a new version of the same question show up again in six weeks? If the need goes away, it's a project. If a new version of the same question is basically guaranteed to resurface, a retainer costs less over a year than repeatedly re-scoping projects to answer the same category of question each time.
For full product builds rather than advisory, that's a different engagement altogether — see Aarohii's Model A and Model B for how that's structured.
Facing this decision right now?
Tell me what is happening, what it is costing the company, and what you need to decide.
Request an Advisory Fit Call