These two get used interchangeably and they are not the same purchase. One buys you an answer. The other buys you someone who makes sure the answer gets implemented. Choosing the wrong one is the most common way operators waste money on outside help.

The actual difference

A consultant on a project is scoped to a defined deliverable with an end date. You have identified a problem — the menu, the training system, the local marketing — and you are buying the expertise and the hours to build the fix. When the deliverable is handed over, the engagement is finished.

A fractional COO is scoped to an ongoing responsibility. There is no single deliverable. You are buying weekly attention on the operation as a whole: KPI review, decisions between meetings, and the pressure that keeps change moving when the week gets busy.

The distinction that matters: a project ends when the work is built. Fractional leadership continues because most operational change fails at adoption, not at design.

When a project is the right answer

Pick project work when you can finish this sentence specifically: "I need someone to build me ___." A menu rebuilt on contribution margin. An SOP library. A twelve-week local marketing calendar. An AI workflow for reporting.

Projects suit operations that have a working management layer — someone who will own the thing once it exists. The deliverable lands, your team runs it, and you have bought a capability rather than a dependency.

Projects are also the right answer when budget is finite and you want a defined cost. Scoped work has a number attached to it; ongoing support is a monthly commitment.

When fractional leadership is the right answer

Pick ongoing support when the honest problem is not knowing what to do — it is that nothing survives contact with a busy week.

That is extremely common and not a character flaw. Operators are pulled into the immediate all day; strategic work has no deadline and therefore always loses. A weekly cadence with someone outside the business is, more than anything else, a structural answer to that.

It also suits operations where the decisions are coming faster than the owner can think them through alone — a second location, adding catering, a management transition. In those periods the useful thing is not a document, it is judgement available repeatedly.

When it is neither

Worth saying plainly: there is a size below which ongoing retainer support does not make arithmetic sense. If the monthly fee is a large fraction of your monthly profit, the improvement required to justify it is unrealistic.

The usual sequence that does work: a diagnostic first, then one project against whatever the diagnostic found, and ongoing support only if implementation turns out to be the bottleneck. Buying a retainer before you know the problem is buying attention you cannot yet direct.

Equally: if you have a specific, bounded question — a lease, a pricing sanity check, a second opinion on a hire — neither model fits. That is an hour of someone's time, not an engagement.

Questions that resolve it quickly

  • Can you name the deliverable? If yes, it is probably a project. If you find yourself describing a situation rather than a thing, it is probably ongoing.
  • Have you tried to fix this before? If you have, and it did not stick, the gap is implementation — which is what a cadence addresses and a document does not.
  • Who owns it afterwards? If nobody on your team can own the result, a project will become shelfware. Either build the owner first or buy the ongoing support.
  • What does doing nothing cost per month? If you cannot estimate that, start with a diagnostic, because you do not yet know what you are buying.

How this works here

For transparency, since most consulting sites hide this: the diagnostic is $350, project work is scoped per problem in the $2,500–$50,000 range, and ongoing advisory is $2,000/month aimed at operations above roughly $1M. The $350 credits toward whatever follows, specifically so that the diagnosis does not have to be a separate financial decision from the fix.

If the pricing question is the one you actually came with, what restaurant consultants cost breaks down all four common models.