The Fee Was Fixed. The Work Was Not
· 7 min read · Written by Faceela Research & Editorial Team
Reviewed by Ahmed Hassan Algammal — Founder and Enterprise Systems Consultant
The conversation happens in every partners' meeting eventually. Somebody says the fixed fees are killing us. Somebody else says clients will not accept time and materials any more. A third person suggests a cap, everyone nods, and the firm goes on quoting exactly as it did before.
Nothing was resolved because the question was framed wrongly. There is no best pricing model. There is a model that fits a particular shape of work, a particular client, and — this is the part that is always left out — a particular firm's ability to measure what is happening while the work is being done.
The pricing model is a governance decision wearing a sales hat
Here is the reframing the rest of this piece depends on.
A pricing model decides how the client pays. A governance requirement is what the firm must be able to see in order to survive that decision. Every pricing model carries one, they are all different, and the requirement is not optional — it is the thing that determines whether the model earns money or merely produces invoices.
Fixed fee obliges you to know delivered effort against the fee, while the engagement is running. Time and materials obliges you to keep a time record you would be willing to show a client who queries it. A cap obliges you to see burn against the ceiling before you cross it, not after. A retainer obliges you to know consumption. A success fee obliges you to know the cost of work carried unpaid, including on the engagements that failed.
That is five different measurement obligations, and the firm signs up to one of them every time it sends a proposal — usually without noticing that it has.
Five models, five questions each
Every one of these is right somewhere and wrong somewhere
Fixed fee
- Who it is for
- Work whose shape you know because you have done it enough times — a statutory audit of a familiar size, a formation, a filing cycle.
- Who it is not for
- Anything whose extent depends on what you find in week two. Also any client whose own records decide how long it takes.
- What you must be able to measure
- Delivered effort against the fee, per engagement, while it is running rather than after it has closed.
- Who carries the estimating risk
- You, entirely.
- Where it fails
- When the estimate is made by someone who will not do the work, and nobody ever compares the two afterwards. The second half is the fatal one — a wrong estimate you learn from is a cost; a wrong estimate you repeat is a business model.
Time and materials
- Who it is for
- Work whose extent is genuinely unknown at the start — an investigation, a dispute, a remediation, anything downstream of somebody else's mess.
- Who it is not for
- A client who needs a number for a board before anything starts, and a firm whose time records would not survive being shown to the client who asks.
- What you must be able to measure
- Time, at the level of detail you are willing to put in front of a client who queries the bill.
- Who carries the estimating risk
- The client.
- Where it fails
- On refusal. Clients who have been burned once decline it on principle, and they are usually right about why — the last firm's narrative was unauditable and the hours were reconstructed on a Friday.
Capped fee
- Who it is for
- The narrow case: a client who will not accept open-ended, on work you cannot yet price, where you have decided to buy the relationship and know that is what you are doing.
- Who it is not for
- Any engagement you are relying on for margin this year.
- What you must be able to measure
- Everything time and materials needs, plus burn against the cap, live — because the day you cross it is the day the model changes and nobody announces it.
- Who carries the estimating risk
- You above the cap. The client below it. Which is both halves of the downside and neither half of the upside.
- Where it fails
- Structurally, and it is worth being blunt about it: you cannot exceed the cap and you cannot keep the saving when the job comes in under. Chosen deliberately it is a price worth paying; chosen as a compromise it is the worst of the three.
Retainer
- Who it is for
- Continuing obligations with a predictable rhythm — a compliance calendar, an outsourced function, a standing advisory relationship.
- Who it is not for
- Project work with a beginning and an end, dressed as a retainer because it makes the invoice easier to raise.
- What you must be able to measure
- Consumption against the retainer, monthly, so both sides know where it stands before the renewal conversation rather than during it.
- Who carries the estimating risk
- Shared, and mispriced in both directions at once across a portfolio.
- Where it fails
- Quietly. It is consumed by March and served free until December, or it is never consumed and renewed anyway — and the firm cannot tell which client is which.
Success fee
- Who it is for
- A defined, verifiable outcome you materially control — a recovery, an approval, a transaction that closes.
- Who it is not for
- Advisory work whose outcome depends on the client acting, and any firm that cannot fund the work while it is unpaid.
- What you must be able to measure
- The cost of work carried unpaid — including on the engagements that did not succeed, which is the number that decides whether the model works at all.
- Who carries the estimating risk
- You, entirely, plus the funding.
- Where it fails
- On the definition of success, which is negotiated in ten minutes at the start and argued for months at the end.
And the row that decides all five
Read the third field down each card. Every model on this figure requires the firm to measure delivered effort — the same measurement, in every case, differing only in what is done with it. A firm that cannot produce it is not choosing between five pricing models. It is choosing which of five ways to be surprised, and it should say so before it quotes.
What the argument is actually about
When partners argue about fixed fee versus time and materials, they are almost never arguing about pricing. They are arguing about who carries the estimating risk, using the only vocabulary available to them.
Under a fixed fee, the firm carries it entirely. If the work runs to twice the estimate, the firm absorbs the difference and the client experiences a well-run engagement. Under time and materials, the client carries it, which is why clients resist it — they have usually been on the wrong end of it once, and their objection is not to the principle but to the invoice they could not audit.
The capped fee is the compromise both sides reach when they cannot agree, and it deserves to be described honestly rather than sold. Under a cap, the firm carries the whole downside and gives away the whole upside: it cannot exceed the ceiling, and when the job comes in under, it bills the lower number. That is a real cost, and it is sometimes worth paying to win a relationship. What it is not is a middle position. A firm that reaches for a cap because it cannot decide has chosen the most expensive of the three.
Every model fails, and each one fails differently
| The model | How it fails | The first sign, if anyone is looking |
|---|---|---|
| Fixed fee | The estimate was made by someone who will not do the work, and nobody compares the two afterwards | Effort passes the fee in month two and the engagement runs to month five |
| Time and materials | The time record cannot survive being shown to the client who asks for it | A narrative reconstructed on a Friday, then a query, then a write-off |
| Capped fee | The cap is crossed and the model silently becomes pro bono | Burn reaches the ceiling with deliverables outstanding and nobody is told |
| Retainer | Consumed by March or never consumed at all — and the firm cannot tell which client is which | Renewal season arrives with no consumption figure on either client |
| Success fee | The definition of success was negotiated in ten minutes and is argued for months | The outcome arrives and two parties describe it differently |
| Any of them, mixed across one client | Nobody can say what the relationship earned | The profitability report has one line per invoice and none per client |
The right-hand column is the only one that matters. Every failure in this table has a first sign that appears weeks or months before the loss is realised, and every one of those signs is invisible in a firm where effort is recorded after the fact and reviewed at month end. The pricing model did not fail. The firm found out too late to act.
The mixed portfolio, which is what you actually have
No firm runs one model. A real practice has fixed fees on the compliance work, time and materials on the disputes, a couple of capped arrangements it regrets, retainers on the long relationships, and one success fee somebody agreed to in a meeting.
That is fine. It is the correct answer, in fact, because the work genuinely differs. What is not fine is the consequence: the firm now needs all five measurement obligations satisfied at once, out of the same record, and most cannot satisfy one.
It gets worse where a single client sits under two models — a retainer for the ongoing work and a fixed fee for the annual piece, which is entirely normal. Ask what that client earned last year and the honest answer in most firms is that nobody knows, because the two arrangements are held separately and nothing joins them at the client. That is the same failure described in client profitability, arriving from the pricing side rather than the cost side.
Changing the model does not fix a measurement problem
This is where firms lose the most money, and it is worth stating plainly.
A firm with fixed fees that keeps losing money on them concludes that fixed fees are the problem, and moves to time and materials. The margin does not improve, because the reason the fixed fees lost money was that nobody could see effort passing the fee — and time and materials requires a better time record, not a worse one. The firm has moved to a model whose governance requirement it satisfies even less well than the one it left.
The same trap runs in the other direction. A firm whose time and materials invoices keep getting queried moves to fixed fee to stop the arguments. The arguments stop. The losses do not, they merely become invisible, which the firm experiences as an improvement for about a year.
The order is: measure first, then choose. A firm that can see delivered effort against price on a live engagement can run any of the five, and can price the next one from evidence. A firm that cannot is choosing which way to be surprised.
What to make a vendor show you
On a live system, on real engagements, not on slides.
- Show one engagement of each pricing model in the same list, with the right measurement on each — effort against fee, time against rate, burn against cap, consumption against retainer.
- Show effort passing the fee on a fixed-fee engagement, as an alert while the work is running, not as a report after it closed.
- Show burn against a cap today, with the deliverables still outstanding beside it.
- Take one client with two arrangements and produce a single profitability figure across both.
- Produce the time narrative for one time-and-materials month in the form you would send to a client who queried it.
- Show what the last three engagements of this type were estimated at and what they actually took, as the input to pricing the next one.
- Show what a success-fee engagement has cost so far, including the ones that did not succeed.
Item 6 decides it. Everything above it is visibility on work in progress, which is valuable and which several systems do adequately. Item 6 is the loop closing — the firm's own history becoming the basis of its next quotation — and it is the only mechanism by which pricing gets better rather than merely more confident. A system that cannot produce it leaves the firm estimating from memory for ever.
Item 2 is the quiet one. Almost every vendor will show you a report of engagements that went over. Very few will show you the moment one crossed, on the day it crossed, to a person who could still do something about it.
The short version
There is no best pricing model. There are five, each right somewhere, and the choice is governed by the shape of the work, the client in front of you, and what your firm can currently see.
Every model carries a measurement obligation, and they are all different. Fixed fee needs effort against fee, live. Time and materials needs a record you would show a client. A cap needs burn against the ceiling before you cross it. A retainer needs consumption. A success fee needs the cost of the failures.
Changing model to escape a measurement problem moves the loss rather than removing it — and usually into a model with a stricter requirement than the one you left. Measure first, then choose.
What that looks like when the engagement, the effort and the price live in one record is on the professional and business services page, and the system is ServX. The neighbouring pieces are scope creep and engagement letters, realisation rate and write-offs, and retainers and deferred revenue. For the capacity side of the same decision — because a fixed fee priced on a grade mix you cannot staff is a fixed fee you will lose money on — utilisation and capacity planning.