Agency pricing debates often have the wrong starting point. Somebody announces that hourly billing is dead, someone else says value pricing is the only intelligent option, and the poor owner still has to make payroll on Friday.
I've written for NextNW about why AI is making traditional agency pricing assumptions uncomfortable. When software reduces the time required to produce an output, the hours spent are an increasingly imperfect proxy for the client's value. But declaring every hour worthless is not a pricing strategy either.
The useful question is more specific: Which model fairly allocates scope, uncertainty, performance risk and the reward for doing good work? Then, before sending a proposal, check whether the numbers can actually sustain the promise.
Five common agency pricing models
| Model | Best suited to | What can go wrong | Protection |
|---|---|---|---|
| Hourly / time-and-materials | Discovery, undefined scope, expert troubleshooting | Client fears an open bill; agency earns less as it becomes faster | Agreed rates, cap, reporting and checkpoints |
| Fixed project fee | Well-defined deliverable and acceptance criteria | Change requests and revision cycles destroy margin | Written scope, assumptions, change control |
| Monthly retainer | Recurring services with predictable cadence | Every “small extra” becomes an unpriced obligation | Capacity boundaries, priorities and regular review |
| Value-based fee | Clear business outcome with credible value and causal influence | Value is hard to agree, attribution weak or expectations inflated | Discovery, evidence, clear measures and limits |
| Hybrid / milestone or performance element | Base work plus measurable phases or agreed upside | Disputes over definitions and outcomes outside agency control | Floor fee, milestones and precise attribution rules |
Hourly: easy to explain, sometimes hard to defend
Hourly billing is entirely reasonable when neither party yet knows the exact effort. It also creates an uncomfortable incentive. If an excellent team learns to solve a problem in 90 hours rather than 140, and the client pays only for the hours, the agency's bill drops even if quality improves.
Consider a purely hypothetical $140 billable rate. At 140 hours the bill is $19,600; at 90 hours it's $12,600. The saving may be excellent for the buyer, but the agency has to find a sustainable way to earn from its expertise and take responsibility for results. For uncertain discovery work, hourly plus a clear cap and review milestone can still be the most honest commercial offer.
Fixed-fee: price the deliverable, not your optimism
Fixed-fee projects remove uncertainty for the buyer and reward the agency for competent systems. But they transfer scope and time risk to the supplier. If the deliverables, number of reviews, required inputs and dependencies aren't clear, a fixed fee can become a blank check in the opposite direction.
Suppose a strategy and content project is priced at $18,000. This example assumes 130 delivery hours at $75 fully loaded ($9,750), specialist suppliers of $1,250, an allocated $1,800 of overhead, and 12 hours of additional quality review at the same $75 ($900). Total modeled cost: $13,700. The remaining $4,300 is a 23.9% economic contribution margin on that job.
Margin = (Fee − Fully Loaded Cost) ÷ Fee
($18,000 − $13,700) ÷ $18,000 = 23.9% in this example.
For an illustrative 30% margin at unchanged cost, the necessary fee would be $13,700 ÷ 0.70, or approximately $19,571.
The arithmetic does not establish a universal “right” fee or benchmark. It helps you spot a proposal that requires perfect delivery to stay worthwhile. Add contingencies for known uncertainty rather than pretending every project will run exactly to plan. For an account-level worksheet, try the client profitability calculator.
Retainers: predictable for everyone, unless nobody defines capacity
A recurring client relationship can be more effective than a new project negotiation every month. It gives the agency continuity and the client consistent access to a team. The danger is confusing recurring access with unlimited scope.
Define outputs or capacity, responsibilities, turnaround expectations, meetings, what requires a new statement of work and how unused or delayed work is handled. Review the actual economics quarterly. If the work has permanently changed, the retainer should not be preserved through silent overtime.
Value pricing: sell judgment carefully
Pricing around business value makes sense when you understand the buyer's economics, can contribute meaningfully and have a credible scope. It works badly when “value” becomes an excuse to invent an ROI number at proposal stage.
A valuable strategic recommendation may take fewer hours than an elaborate production task. The buyer can reasonably pay for a well-defined expert decision. They should also know what is and isn't promised. In marketing, performance depends on budgets, creative approvals, product quality, competitors, sales teams and external market changes. Avoid guarantees over outcomes you cannot control.
Hybrid pricing: a practical compromise, not a loophole
A hybrid engagement might have a base monthly fee for research and delivery plus clearly defined project milestones. Another version adds a limited bonus for a specific, auditable outcome. Both can work when the buyer and agency agree on attribution, measurement period, excluded events and what happens if the client's internal team delays execution.
Never price a complex delivery obligation entirely on a speculative success bonus unless you can afford the downside and have verified the terms. Risk sharing only works when both sides understand the risk.
AI changes the effort equation. It does not remove accountability
AI can reduce production time or expand what a team can test. That can make the agency more efficient, but it also raises a legitimate question from clients: what are they paying for now? My answer is that the promise must become clearer, not more obscure. A client should understand the work, the expertise applied, the quality checks and the specific deliverables or outcomes under contract.
If automation improves speed while quality, responsibility and agreed scope remain the same, a fixed fee may capture part of that efficiency for the agency. Competition will also influence prices over time. Don't sell the magic of the tool; sell a responsible service.
Seven checks before the proposal leaves your desk
- What exactly is included and excluded?
- Which dependencies are controlled by the client?
- Which costs are direct, shared and likely to change?
- What happens when the buyer requests a materially different scope?
- Who is accountable for approvals and delivery quality?
- What economic margin remains after a realistic contingency?
- Can the commercial promise be explained in one plain-English paragraph?
There is no universal best pricing model. A good model makes both parties clear about what success requires and leaves the agency capable of delivering it. That's a much more useful ambition than winning an argument about hourly rates.
Choose the contract by the uncertainty, not by what's fashionable
Start with two questions: Can we define what we will deliver? and Can we explain what happens when important assumptions change? If the work is genuinely exploratory, a capped discovery phase may be more honest than a fixed price disguised as certainty. If scope and acceptance are clear, fixed fees can reward efficiency. If the relationship is continuous, a retainer can work provided access and capacity are described. If the financial outcome is measurable, a value or hybrid element may be reasonable—but only when the agency can influence and verify it.
Kent's June 2026 analysis of agency pricing and AI starts from the tension between hours, expertise and value. Faster production does not make discovery, quality assurance or responsibility disappear. Nor does the existence of AI justify a made-up price premium.
Rewrite the commercial promise in three sentences
One: The client receives [specific deliverable or recurring capacity]. Two: We depend on [inputs and approvals], and changes to [scope boundaries] are handled through [agreed mechanism]. Three: We are accountable for [work and quality], while [outcomes beyond our control] are measured but not guaranteed. If those sentences cannot be made precise, the problem isn't your rate card yet.
Do an AI-era margin test before picking a model
Consider an illustrative project whose production time falls by 40% through automation. Reprice it at the same fixed fee and you may earn a stronger contribution, assuming quality checks and scope remain unchanged. Change it to hourly billing and the client may benefit directly from fewer charged hours. Both can be fair in context, but neither excuses poor quality or misleading descriptions of what the agency actually does. Before agreeing to a fee, check fully loaded delivery economics and the gap between growth and operating profit.
Price the promise you can actually deliver.
Bring a real scope and cost model to a founder advisory conversation, not a vague debate about whether hourly billing is dead.
EXPLORE FOUNDER ADVISORY ↗