Fixed Fee or Hourly? The Question to Settle Before You Read Either Proposal

The owner of a three-location heating and cooling company needed two years of books untangled before she could refinance, and she asked two firms for proposals on the same described problem: vendor payments miscoded, payroll liabilities sitting in the wrong accounts, and a prior preparer who had left the state. One firm quoted an hourly rate with an estimated range and a promise to check in at the midpoint. The other quoted a single fixed figure, roughly the middle of the first firm's range, with the words "complete cleanup through December 31" on the signature page. She signed the fixed fee on a Tuesday. By March she had paid about sixty percent more than that figure, and she still believed, correctly, that she had chosen the better firm.

Working backwards from the overrun to the hour nobody bought

The change orders, three of them, were not padding. The first covered reconstructing nine months of bank activity from statements because the prior preparer's file was never produced and the feeds only reached back so far. The second covered a payroll service that had been switched mid-year, which meant two sets of quarterly filings to reconcile instead of one. The third covered a line of credit that had been used as a checking account. Each was genuine extra work, each was disclosed in writing before it started, and each was outside the assumptions schedule attached to the fixed fee she signed.

So the failure was not the firm's and it was not greed. It was that she bought a price for a body of work whose size nobody had measured, including the people quoting it. The fixed figure was not a promise about the work; it was a guess about the work, wrapped in a format that reads like a promise. The hourly proposal was the same guess with the uncertainty left visible. She chose the version that hid the thing she most needed to see, and she chose it because hidden uncertainty is more comfortable to sign.

The decision that made the overrun likely came earlier than the signature. It came when she asked two firms to price a problem that neither had looked at.

Decide who actually knows the size of the job

A fixed fee transfers risk from you to the provider, and that transfer is worth paying a premium for when the provider knows more about the work than you do. A firm that has done four hundred of a particular filing knows its own distribution: most take a day, a few take three, and the fixed price carries the average plus a margin. You get certainty and they get the upside on the easy ones. That trade is fair and it is why fixed fees exist at all in professional services.

The trade breaks when neither side knows. If the variability lives inside your records, your history, your other vendors, or your own decision-making speed, the provider cannot price it, and the fixed fee becomes either a number inflated to survive the worst case or a number that will be revised. The cleanup failed this test on its face. The entire question was what condition two years of books were in, and the only people who could answer that were sitting in the firm she had not yet hired.

So the first thing to establish, before you read a single proposal, is whether the uncertainty is theirs or yours. Theirs, and a fixed price is usually the better buy. Yours, and you are asking someone to insure a risk they cannot inspect.

Buy the diagnosis separately from the work

The move available to her, and available to almost any customer facing this, was to pay one firm a few hundred dollars for a scoped look. Four hours in the file, access to the accounting system, a short written statement of what is there and what is missing, and a quoted range for the rest. That spend is small, it is capped, and it converts an unpriceable job into a priceable one. It also tells you something no proposal can, which is how the firm behaves when it is inside your mess.

Good providers rarely resist this. The ones who do resist are telling you either that they do not want their estimate tested against reality or that they intend to recover the diagnostic time inside a larger engagement you have not agreed to yet. Either way the refusal is information. Pay for the diagnosis, read the write-up, and then decide the pricing form with the uncertainty largely gone. At that point a fixed fee is a real promise, because somebody has counted the work.

What each format has to come with before it is safe

An hourly engagement is only as good as its instrumentation. You want a written estimate with a range, a not-to-exceed figure that requires your approval to cross, billing detail at the task level rather than the month level, and a named trigger for a conversation, such as reaching sixty percent of the estimate. Rate matters far less than those four things. A higher rate attached to a firm that works fast and tells you early will cost you less than a low rate attached to a firm that bills in silence until the invoice arrives.

A fixed fee is only as good as its assumptions schedule, which is the part customers skip and the part that decides everything. Read it as a list of the conditions under which the price holds, then ask yourself honestly which of those conditions you can verify today. Read the exclusions the same way. Then ask how a change order gets raised, who authorizes it, and whether work stops while you decide. The Federal Trade Commission oversees how services are advertised and billed, and plain disclosure of what a quoted price does not cover is the ordinary expectation, not a courtesy.

Her second engagement with the same firm, the following year's tax and advisory work, was fixed fee and landed within a few hundred dollars of the quote. Nothing about the firm had changed. What had changed was that the condition of her records was now known to both sides, which is the only state in which a fixed price is a price rather than a forecast.

Building the judgement so it carries to the next engagement

The habit worth forming is to stop treating hourly and fixed as a preference and start treating them as a readout of how much is known. Ask yourself where the variance sits, whether the provider can see it, and what it would cost to make it visible. Ask whether you are buying a defined deliverable or an open problem, because open problems resist fixed pricing no matter how the proposal is formatted. Ask what you would do at the point the estimate is exceeded, and make sure the engagement letter gives you something to do at that point.

Customers who get this consistently right tend to have one more trait, which is that they are willing to spend a small, bounded amount early to avoid an unbounded amount later. Four diagnostic hours are cheap. Two years of surprises are not.

She kept the firm, renewed the following January, and now sends them a short scope note before asking for any number at all. The refinance closed, late by about six weeks, at the rate she had been quoted.