There is a lot of confident advice about agency pricing, most of it arguing that hourly billing is obsolete and value-based pricing is the answer. It is more useful to understand what each model actually does to your risk, because the right answer changes by client, by service line, and by how well you can estimate.
Hourly
You are paid for input. The client carries the risk of the work taking longer.
Hourly gets dismissed as unsophisticated, which underrates it. It is the only model that is automatically fair when nobody can predict the shape of the work — open-ended discovery, ongoing support, anything where the client will change direction and both parties know it.
Where it fails:
- Your upside is capped and your downside is not. Getting faster makes you less money. Getting better at something reduces your revenue from it, which is a genuinely perverse incentive.
- Every efficiency is a price cut you gave away.
- It invites scrutiny of hours rather than outcomes, which is a worse conversation to be having with a client.
Hourly works best when the scope is genuinely unknowable and the relationship is built on trust. It works worst when you are good at something repeatable.
Fixed fee
You are paid for an output. You carry the risk of it taking longer.
The client knows what they will pay, which is often worth more to them than the absolute number. You keep the benefit of getting faster. Most agency work is sold this way for good reason.
Where it fails:
- Estimation error goes straight to your margin. A job quoted at 40 hours and delivered in 65 has lost 38% of its value, and nothing about the invoice reflects that.
- Scope creep is invisible until you compare hours to fee, which many agencies only do at the end.
- It punishes optimism, and estimation is a systematically optimistic activity.
Fixed fee requires the thing most agencies do not have: honest historical actuals, by project type, to estimate from. Without them it is a bet placed on feel. With them it is the best model for most work.
The prerequisite is therefore tracking time even on fixed-fee work — which people question, because the hours are not billed. They are not billed; they are the only thing that tells you whether the price was right. See how to tell whether a project made money.
Value-based
You are paid a share of the outcome's worth. Both parties carry a different risk.
When it fits, it is the only model where your ceiling is not set by capacity. A piece of work that takes 60 hours and makes the client $2m can reasonably be priced well above cost-plus.
Where it fails:
- Most agency work has no measurable, attributable value. A rebrand's contribution to next year's revenue cannot be isolated, and pretending otherwise makes the pricing conversation an argument about attribution.
- It requires access to the client's commercial numbers and a buyer senior enough to discuss them. Most agencies are not selling at that level most of the time.
- It is sold as the sophisticated choice, which leads agencies to attempt it on work where it does not apply and end up with fixed fee plus a story.
Value-based pricing is excellent for a narrow band of work: measurable commercial outcomes, senior buyer, demonstrable track record. It is a poor default.
The practical answer
Most agencies should run a mix, chosen by the shape of the work:
| The work | The model |
|---|---|
| Scope genuinely unknown, direction will change | Hourly, or a capped estimate |
| Repeatable, well-understood, good historical data | Fixed fee |
| Ongoing, predictable volume | Retainer with a defined scope |
| Measurable commercial outcome, senior buyer | Value-based, occasionally |
And underneath all four, track the hours regardless. The model determines what appears on the invoice. It has no bearing on whether you need to know what the work cost — and the agencies that stop tracking when they move to fixed fees are the ones that discover two years later that a whole service line has been losing money at scale.
The number that tells you the model is working
Whatever you charge, the check is the same: effective hourly rate — revenue divided by hours actually worked — measured against your true cost per hour.
Hourly work should land near your rate card. Fixed-fee work should land above it, because you are being paid for the risk. Value-based work should land well above it, or you have taken the risk without the reward.
If fixed-fee work consistently lands below your hourly rate, you are not doing sophisticated pricing. You are giving a discount for taking on the delivery risk, which is the worst of both.
