The person sitting across from you has done this pitch before. Not necessarily to deceive anyone. It is simply that a pitch is a rehearsed artifact: a selected set of past wins, a vocabulary tuned to sound native to your problem, and a confidence level calibrated by repetition rather than by knowledge of your particular situation. A specialist who is genuinely good has one. A specialist who is merely plausible has one too, and it is often the better of the two, because polishing the pitch is where their effort has gone.
This is why interviewing harder does not help much. You cannot out-question a rehearsal. What you can do is set up the engagement so that the difference between good and plausible becomes visible in the ordinary week-to-week work, and becomes visible early, while you still have money left and options open. That setup is almost entirely paperwork. Not much of it, and none of it complicated.
Buy a small piece first, and make it a real deliverable
The single most useful move is to break the first purchase off from the rest. Instead of engaging someone for six months of marketing strategy, buy two weeks of diagnosis that ends in a written document you can hold. Instead of a full system build, buy an architecture review. Instead of an ongoing bookkeeping relationship, buy a cleanup of one quarter's books.
The mechanism here matters. A long engagement pays for presence: hours, meetings, availability. Presence is easy to supply and hard to evaluate, so a plausible consultant can ride it for months before anyone can say what went wrong. A short engagement with a defined output pays for a thing. The thing either exists, addresses your situation, and tells you something you did not know, or it does not. Three people can read it and form the same opinion.
Price the diagnostic properly and pay for it. Free scoping calls produce free-scoping-call output, which is a proposal for more work. When you pay, you own the document, and you can take it to someone else if the relationship ends there. Write that ownership down.
The scope document, and the four lines most of them are missing
Most scopes of work read like a description of activity: "review current processes, interview stakeholders, develop recommendations." Activity language is unfalsifiable. Someone can do all of that and hand you nothing. The fix is to write four specific things into the document before anyone starts.
- The deliverable, as an object. Not "recommendations" but "a written memo of roughly fifteen pages covering X, Y and Z, delivered by the 14th." You should be able to picture it.
- What counts as done. Acceptance criteria, in plain terms. Who reviews it, how long they have, and what happens if they say it is incomplete. Without this, "done" is whenever the invoice arrives.
- What you are supplying, and when. Access to systems, three hours of your operations manager's time, last year's financials by the 3rd. Half of all failed engagements fail on this line. Writing it down turns a later argument into a calendar item.
- Who decides. One named person on your side who can approve, reject, and answer questions within a day or two. Committees stall outside experts, and stalled experts bill for the stall.
A good specialist will read this and push back on parts of it. That pushback is your first real signal. They will say the fifteen-page memo is the wrong shape, or that they need a fourth data source, or that the deadline assumes your team responds faster than teams usually do. A plausible one signs whatever you send, because the document is not where they intend to compete.
The administrative stack, which takes an hour once
Before the first invoice, you want a short file assembled. It is unglamorous and it prevents most of the trouble that shows up later.
An engagement letter covering fees and how they are structured, expenses and whether they need pre-approval, the notice period for ending the work, confidentiality, and who owns the work product when it is finished. That last item surprises people. Absent an agreement, the person who created the material may retain rights to it. If you commissioned a training curriculum or a set of financial models, say in writing that the output is yours.
A W-9 from anyone you are paying as an independent contractor, collected before the first payment rather than chased in January. The Internal Revenue Service oversees the reporting rules for payments to non-employees, and the practical consequence for you is that you need the taxpayer information on file and the year's payments totaled. Getting the form up front turns year-end into a clerical task.
A certificate of insurance where the work carries real exposure: professional liability for advisory work, general liability for anyone on your premises. Ask for it once. Firms that carry it produce it the same day.
Then a change order form, which is nothing more than a half page saying what is being added, what it costs, and what it does to the timeline, signed by both sides. The reason to have the template ready on day one is behavioral. Scope creep is rarely a decision anyone makes. It accumulates through small verbal agreements in meetings. A form makes each addition a small, visible transaction instead.
What the ordinary week looks like when it is working
Set a standing twenty minutes each week and a short written note that precedes it. Three headings: what was completed, what is next, what is blocked or open. The note should take the consultant ten minutes to write. If it takes longer than that, the engagement is too diffuse.
Now watch the notes across three or four weeks, because this is where good and plausible separate, and the pattern is consistent.
A good specialist's questions get narrower over time. Week one they ask broad things. By week three they are asking about a specific line in a specific report, because they have built a working model of your situation and found the place where it does not fit. A plausible one's questions stay at the same altitude the whole way through, or drift toward asking you what you think the answer should be.
A good specialist tells you early when an estimate is going to move, and gives you the reason. Bad news travels fast from people who are actually in the work, because they hit the obstacle in real time. Estimate problems that surface in the final week were usually known in the second.
A good specialist declines things. They will tell you a question is outside what they can answer, or that you need a different kind of professional for one branch of the problem. The willingness to mark the edge of your own competence is difficult to fake, because it costs revenue to say it.
And the open items list should shrink. Items that sit unresolved for three consecutive notes are either blocked on you, which the note makes plain, or blocked on nothing, which the note also makes plain.
Ending it cleanly, whichever way it goes
Write the off-ramp into the engagement letter while everyone is optimistic. A short notice period, a statement that work completed through the notice date is payable, and a handover clause: files, credentials, documentation and drafts returned in usable formats. Then a closing conversation, thirty minutes, covering what they would do next and what they would warn their replacement about.
Run this well once and the second hire is faster, because the scope template, the change order and the weekly note format all carry over. The folder you built for a consultant you were unsure about becomes the standard way your organization buys expertise, and the people who are genuinely good tend to prefer working inside it.