What Diagnosis Costs

Problem: You want to know what it costs to have someone find out where your company is losing money. You search for it, and every result on page one is an accountant quoting you a price to check your financial statements. Nobody prices the thing you were actually asking about.

Quick Win: An operations audit is a paid look at how your company actually runs, done to find where it wastes time and money, and it is completely different from the financial audit that dominates every search result (AccountingTools). The real published ranges: $5,000 to $25,000 over three to six weeks for a small-business operations audit (Covision Consultants), and $8,000 to $35,000 for the version that also checks whether you are ready to use AI, with most mid-sized companies landing at $15,000 to $25,000 for six weeks (Elite AI Advantage). Our own published number is €4,000 to €15,000. Four things move it, and only four.

What You Get When You Search This, And Why It Is Wrong

Run the search yourself. "Operations audit cost." Page one is accountants.

You get SOC 2 pricing, which is a security check that software companies buy so their customers will sign. Across 171 accounting firms, the thorough version of a SOC 2 report runs $15,000 to $430,000, with specialist firms at $15,000 to $70,000 and the four largest accounting firms at $45,000 to $430,000 (SOC2Auditors).

You get financial statement audits, which check whether your books are accurate. A for-profit company under $5 million in revenue pays $3,000 to $10,000, and accounting firms bill those hours at $150 to $400 (Citrus Audit Group). Companies between $5 million and $50 million pay $15,000 to $35,000 (GuzmanGray).

None of that is your answer. It shows up because the accounting profession has owned the word "audit" for a century. You are searching inside somebody else's vocabulary.

An Operations Audit Is Not A Financial Audit

The definition worth memorizing: an operations audit is "an examination of the manner in which an organization conducts business, with the objective of pointing out improvements that will increase its efficiency and effectiveness" (AccountingTools). One sentence separates the two: operations audits improve performance, financial audits confirm the numbers are reliable.

That last row is why the price behaves differently. A financial audit is a compliance cost with a fixed shape, so it drifts toward an hourly rate. An operations audit is optional. You choose to buy it, so the price is set by how much you put in scope, and you set the scope.

What Has To Be Inside One For It To Be Worth Paying For

Before you compare prices, fix what you are comparing. Five things separate a diagnosis worth five figures from a slide deck.

A ranking, not a list. Twenty findings sorted alphabetically is a to-do list you will never start. One ranked list with a clear number one is a decision. You are not buying findings, you are buying the order you fix things in.

Your numbers, not industry averages. If the analysis compares you to an industry average, you learned about the industry. Every row has to trace back to your own invoices, your own timestamps, your own volumes.

Money attached to every row. Not "improve handoffs." A number: this costs you roughly this much per year, and here is the arithmetic. If nothing is priced, nothing gets prioritized. We covered how to build those numbers in where your business is actually losing money.

Traceable evidence. Every claim points to a named source: this report, that interview, this export. When the department head pushes back, and one always does, you put the receipt on the table in the meeting.

One named first move. The document ends with a single sentence saying what to do Monday. Anything else is a research paper.

That is the shape of a ranked, evidence-backed bottleneck map. A bottleneck is the one step that holds up everything after it. Everything below is about what makes that document cost €4,000 versus €15,000.

The Four Things That Move The Price

There are only four. Anyone who quotes you a number without asking about these four is guessing.

Notice what is not on this list: how big your company is. Revenue does not change the amount of work. A 40-person company with five departments and messy data costs more to diagnose than a 300-person company with one function in scope and clean exports.

Notice what else is not on it: how many days a consultant sits in your office. Independent management consultants bill roughly $1,500 to $4,000 per day for general management work, with a US average around $1,800 per day (ConsultFees). Once you pay by the day, the incentive flips. More days is more revenue for the seller.

Published Ranges, With Sources

Here is every real published number we could find, in one place.

The pattern across all of them is the same: roughly $5,000 for one department, roughly $15,000 to $25,000 for a whole mid-sized company, three to six weeks either way. If you get quoted $60,000 for an operations diagnosis, you are being quoted a multi-year change program with a diagnosis stapled to the front.

One argument for insisting on a fixed fee. Deloitte's Q1 2026 CFO Signals survey polled 200 finance chiefs at North American companies with at least $1 billion in revenue, and 49% cited pressure to invest in new technologies such as cloud or AI as a driver of their cost-management work (Deloitte). Whoever runs your finances is already fielding more technology asks than they can approve. A number they can say yes or no to once beats an engagement that bills by the day and grows a scope. Ours is agreed in writing before anything starts, which is why we publish the range at all.

What Those Weeks Actually Buy You

Strip the calendar back and a diagnosis is three activities, and only three.

Interviews. Structured conversations with the people who do the work, not the people who manage them. Six to twenty of them depending on scope. The gap between what actually happens and what the process document says happens is usually where the money is.

Rebuilding the numbers. Pulling volumes, timestamps, and costs out of your own systems to work out what each step really takes. The unglamorous half, and the one that makes a finding defensible instead of a story someone told you.

Ranking. Putting a euro figure next to each bottleneck and sorting. The ranking is the deliverable. Everything before it is raw material.

If a proposal describes weeks of "stakeholder alignment" without naming who gets interviewed and which systems get pulled, you are buying calendar.

When You Should Not Buy One

Four situations where the honest answer is don't.

You are below roughly $1 million in revenue, or you have not proven people will buy repeatedly. Your constraint is demand, not efficiency. Diagnosing a process you are about to throw away is expensive procrastination.

You already know the answer. Plenty of CEOs can name their bottleneck in one sentence and want a document to win an internal argument. That is a real need, but a smaller and cheaper one. Say so and pay for that instead.

You need a person, not a document. If nobody is making the recurring decisions that cross departments, a ranked list will not fill that hole. We wrote about the trade-off in fractional COO versus installed operations.

You have already started building. You are past diagnosis and into a measurement problem. Capture the before numbers first, which is the argument in baseline it first.

Where This Breaks: The Audit That Recommends Another Audit

The failure mode is not a bad diagnosis. It is a good diagnosis that nothing happens to.

One consultancy's own estimate puts the share of consulting recommendations never fully put into practice at 30 to 50%, though it cites only "industry data" without naming a study, so treat it as directional (Jalubro). The mechanism it describes is the believable part: by month six the executive who paid for it has a new priority, part of the team has turned over, and anything needing two departments to cooperate stalls.

The harder numbers point the same way. S&P Global Market Intelligence surveyed over 1,000 companies and found 42% abandoned most of their AI projects in 2025, up from 17% the year before, with an average of 46% of trial projects never making it into real use (Telecom Reseller, reporting S&P Global).

So watch for three specific tells when you read a proposal.

The diagnosis that ends in a second diagnosis. If the recommended next step is a "deep dive" to work out the size of the opportunity, you did not buy a diagnosis. You bought a proposal for one, at full price.

Findings with no number. "Reduce handoff friction" cannot be prioritized, funded, or checked afterwards. If the top three findings have no euro figure, the ranking is decoration.

A recommendation nobody in the building can execute. A fix that needs three departments to change how they work, with no owner named and no first step, dies quietly in month two. The audit should tell you what to fix and who moves first.

The defense against all three is boring: insist the deliverable is ranked by money, traceable to your own data, and ends with one named move. Then pay a fixed fee for exactly that.

Frequently Asked Questions

Is a cheaper audit worse?

Not automatically, but check what got cut. At the low end of every published range the scope narrows to one department and the analysis only describes problems instead of pricing them (Covision, Elite AI Advantage). That is exactly right if you already know which department is the problem, and wrong if you do not, because the value of a company-wide diagnosis is finding out you were looking at the wrong department.

Should I pay hourly or a fixed fee?

Fixed, for a diagnosis. Hourly makes sense when nobody can predict the shape of the work, which is true for a financial audit of messy books and false for a scoped diagnosis. Day rates of $1,500 to $4,000 (ConsultFees) mean every extra week of discovery is revenue for the person deciding how long discovery takes.

Does the audit have to come from the firm that builds the fix?

There is a case for separating them: one firm doing both means every diagnosis conveniently recommends what that firm sells. The counter-argument is that a diagnosis with no build attached is the one most likely to end up in a drawer. The practical middle is to buy the diagnosis as a standalone fixed fee you own outright, whether or not you build anything with the same people afterwards.

---

The short version is that the market has quietly settled around $5,000 for one department and $15,000 to $25,000 for a whole mid-sized company in three to six weeks, and almost nobody says it out loud. We publish our range, quote it fixed before we start, and hand back a ranked map costed in euros from your own numbers that you keep either way. See what we build for companies →