Answers · in depth
How to Choose an Operations Consultant: Ten Questions to Ask Before You Pay
Most owners pick an operations consultant the way they’d pick a lawyer at a dinner party — by confidence and referral. That works until it doesn’t. Here are ten questions that separate consultants who read records from consultants who sell decks — with what a good answer sounds like, what a bad one sounds like, and why this list would eliminate us too if we ever answered it badly.
The pitch meeting is the worst place to judge a consultant
By the time a company reaches $5 million a year, the owner has usually sat through at least one consulting engagement that produced a binder and changed nothing. The invoice was the small cost. The real cost was weeks of attention — yours and your team’s — spent feeding interviews and workshops that ended in a summary of what everyone already believed.
The pitch meeting can’t protect you from that, because it is the one setting where every consultant performs well. Everyone shows up prepared, fluent in your industry’s vocabulary, and confident. The differences that matter — what they’ll actually read, who does the work after the signature, how they behave when the findings are thin — never surface on their own. You have to ask.
The ten questions below are the due-diligence checklist we’d want any owner hiring a small-business operations consultant to run — on us included. They aren’t trick questions; a serious consultant should enjoy answering them. And they cut both ways: if we ever answer one of them badly, they eliminate us too.
Questions 1–4: What they read, what they show, who shows up, what you keep
1. “What records of mine will you actually read?” Operations problems live in records — phone logs, CRM pipeline history, quotes, invoices, job schedules — not in opinions about records. A good answer names systems and asks what you run: “Your phone system’s call log, your CRM’s stage history, your quoting tool, your invoices — what are you on?” A bad answer leads with stakeholder interviews and a discovery workshop. Interviews have their place, but as the whole method they produce a well-formatted echo of what you already believed.
2. “Show me a sample finding.” Not a case study, not a logo wall — one page, sanitized, in the actual format you would receive. You’re checking for three things: the record the finding came from, the arithmetic that turned it into a monthly dollar figure, and the specific action it points to. A good answer is a page in front of you before the meeting ends. A bad answer is a speech about proprietary frameworks and confidentiality — confidentiality protects a client’s data, not a consultant’s format.
3. “Who does the work — you, or someone I haven’t met?” In many firms, the person who wins the engagement hands it off the moment the signature dries. A good answer is a name: who reads the records, who writes the findings, who presents them to you. A bad answer is “we’ll staff it with the right team” — which means the person who impressed you was the sales function.
4. “What do I own when we stop?” A good answer: everything — the report, the numbers behind it, the working files — usable with your own team, with a different firm, or alone. A bad answer produces findings that only function as a bridge to phase two. If acting on the report requires its author, it isn’t a report; it’s a sales document with page numbers.
Questions 5–7: Scope, the empty result, and proof
5. “How is the scope fixed — and what happens when you find something outside it?” A good answer sets scope and fee in writing before work begins, and handles adjacent discoveries by noting them in the report rather than billing them midstream. A bad answer is open-ended hourly work that “follows the problem wherever it leads.” Wherever it leads is always more hours.
6. “What happens if you find nothing?” This question divides the field faster than any other. Finding nothing worth pursuing is a real possible outcome of honest diagnostic work, and a good answer treats it that way — in writing, before you pay: you get a clean readout that says so, and you keep the report. A consultant who cannot imagine finding nothing is telling you the conclusion was written before the work started.
7. “How do you verify a result — against what?” The only honest answer involves a dated baseline: before anything changes, the number as it stands on a specific date, pulled from your own records. Improvement is then measured against that record, not against anyone’s memory of how things used to feel. A bad answer measures results in adjectives — “tighter,” “smoother,” “more efficient” — or asserts them months later from goodwill.
Questions 8–10: Refusals, incentives, and who they turn away
8. “What will you refuse to do?” A consultant with no refusals has no shape. Good answers sound like boundaries: “We don’t recommend software we resell.” “We won’t promise a number before we’ve read anything.” “We don’t take implementation work our own audit generated unless the audit stands on its own first.” A bad answer is “we’re a full-service partner” — which means every finding will point, conveniently, at something they sell.
9. “How do your fees shape what you’ll find?” Every fee structure creates an incentive, and a good consultant names their own instead of claiming to have none. Hourly billing rewards duration. A percentage of projected savings rewards aggressive counting. A fixed fee rewards finishing — and removes the reason to pad the findings — though it also forces the consultant to be selective about who they take on. The bad answer isn’t any particular structure; it’s the claim that theirs is the one with no temptation attached.
10. “Who are you not for?” A good answer is specific and comes quickly, because they’ve turned people away before: companies below a certain size, situations where the records are too thin to read, problems that are really staffing or capital problems wearing an operations costume. A bad answer is “we work with businesses of every size and stage.” Everyone-fits is not a scope; it’s a sales posture.
Our answers to all ten are on this site
We wrote this checklist knowing it cuts both ways, so our answers are published where you can read them before anyone at Lumina knows you exist. The Revenue Review is a two-week read of your own records — phone logs, CRM history, quotes, invoices — at a fixed fee of $4,500, set in writing before work begins. Every finding carries a monthly dollar figure, cites the record it came from, and is measured against a dated baseline.
And if we find nothing worth pursuing, the readout says so in writing — and the report stays with you either way. That isn’t a marketing flourish; it’s our answer to question six, and it’s in the agreement.
The rest — the records we read, a sample finding you can hold up against question two, who does the work, what you own when we stop — lives on the pages linked below. Read them the way you’d read any consultant’s answers: skeptically, with this list open.
Asked right after
- What questions should I ask an operations consultant before hiring one?
- Ten cover most of it: what records they will read, whether they can show a sample finding, who actually does the work, what you own when the engagement ends, how scope is fixed, what happens if they find nothing, how results are verified against a dated baseline, what they refuse to do, how their fee structure shapes their incentives, and who they are not for. The answers matter less as facts than as behavior — a consultant who welcomes these questions is telling you something, and so is one who deflects them.
- How much does an operations consultant cost for a $5 million company?
- There is no standard rate — fees run from hourly billing to monthly retainers to percentages of projected savings, and each structure creates its own incentive, which is worth asking about directly. What you can insist on at any price: a fee fixed in writing before work begins, and a stated answer to what happens if nothing is found. Our diagnostic, The Revenue Review, is a fixed fee of $4,500 for two weeks — and if nothing worth pursuing turns up, the readout says so in writing.
The next step
Bring us the process that keeps getting stuck.
A 20-minute conversation about the workflow, the teams and the systems involved — and a straight answer on whether The Revenue Review is the right next step. If it is not, we say that on the call.