If you are here, you want a number, and I am going to disappoint you for one paragraph before I make it up to you. A flat number would be a lie. I have run a design studio, hired firms as the client, and priced work against them as a solo consultant, and the honest answer is that cost follows structure. Once you can see the structure, the quotes stop feeling random.
This essay is the pricing half of my guide to choosing a product design firm. That one covers how to judge quality. This one covers what you will pay and why.
The three shapes a quote arrives in
Almost every firm prices one of three ways. Hourly or day-rate blends, where you pay for time and the risk of overrun sits with you. Fixed project pricing, where the firm carries the overrun risk and pads the price accordingly. And retainers, a monthly commitment for ongoing work, cheapest per hour and easiest to let drift into paying for presence instead of progress.
None of these is wrong. Fixed pricing suits work with a crisp definition of done. Blended rates suit exploratory work where scope is honestly unknowable. Retainers suit a real ongoing stream, and quietly punish you when the stream dries up and nobody cancels. The trap is not the model. It is signing one model for the other model’s kind of work.
If you do take a retainer, build the guardrails in while everyone is still being polite: a deliverable cadence rather than an availability promise, a quarterly re-scope where either side can resize the commitment, and clean cancel terms measured in weeks. A healthy retainer survives those clauses easily. An unhealthy one was counting on you never reading them.
Where the money actually goes
Here is the part no proposal itemizes. Your bill buys two things: the people doing the design, and the structure around them, account managers, project leads, new business, the office, the process itself. At a large agency, the structure can cost as much as the craft. At a boutique, less so. Working directly with a senior consultant, the ratio approaches one.
The question is not whether the rate is high. It is how much of the rate touches your product.
This is why the same project can quote at wildly different prices and why the cheapest bid is rarely the cheapest outcome. A quote is a claim about structure. When you compare two of them, do not compare the totals. Compute the one number proposals never print:
The effective senior rate
Show the math
Slide the numbers to your real proposals. Getting the senior-hours figure means asking each firm who is on the work and for what share, which is itself a revealing question. Junior hours are not worthless, but you should know what fraction of the bill buys judgment versus production.
Run that math on two competing proposals and the conversation changes. A higher total with more senior hours in the files is routinely the cheaper purchase, and now you can show why instead of arguing feel.
Rough shapes, honestly framed
With the caveat that markets and scopes vary: defined projects at boutique firms tend to land in the tens of thousands of dollars. Full product engagements, a product designed end to end with research and a system behind it, run into six figures. Large agencies charge multiples of boutique prices for the same hours, because you are buying their structure too. Senior independent consultants usually sit below boutique totals for the same senior hours, because there is no structure to feed.
Treat those as shapes, not quotes. The spread inside each band is driven by seniority, city, and how well-defined your problem is when you arrive. That last one is the only lever you fully control, and it is worth more than any negotiation.
How to read the proposal itself
When the quotes arrive, most people flip to the number on the last page. The predictive pages are the ones before it. Look for named people with a share of hours next to them, because a proposal that names its team has already committed to staffing it. Look for an assumptions list, because that is where a fixed price quietly stops being fixed. And look for what is excluded: revisions, research participants, front-end build, the design system documentation. The excluded list is the second invoice, drafted early.
One phase deserves special attention: discovery. In a good proposal it is short, sharply scoped, and priced like the diagnostic it is. In a padded one it is a month of workshops that produces a deck restating your brief at firm rates. If you arrive with the problem well framed, say so, and watch what happens to that phase. Firms that resize it are pricing the work. Firms that defend it are pricing the process.
And get the revision terms in plain language before you sign. Not unlimited revisions, which no one honest offers, but a clear statement of what happens when round two reveals that round one missed: whose hours, whose money, how fast. You will use this clause. Every real project does.
How to make any price cheaper
Arrive with the problem, not the solution. A firm quoting redesign our checkout because it leaks can scope tightly. A firm quoting redesign our app is pricing in the archaeology of figuring out what you actually need, at firm rates.
Start small on purpose. A first scope measured in weeks tells you whether the numbers move before the roadmap money is committed. I make this argument from the firm-selection side in the questions I would ask before signing, and it doubles as a pricing strategy: small first scopes keep every later quote honest.
And own your files, your research, and your design system from day one. Lock-in is a price you pay later, at a rate you did not negotiate.
The back-of-envelope worksheet
Before you sign anything, compute what the engagement actually costs, which is never the number on the proposal. Four lines cover it, and this calculator lives with the rest in the toolkit:
True cost of the engagement
- Your team’s hoursDecisions, reviews, user access, data pulls. A real project needs 3 to 6 hours a week from your side; multiply by the weeks and by what those people cost you.
- Revision reserveSomething in round one will miss. If the contract’s revision terms are clean, 15% covers it; if they are vague, budget more and read the terms again.
- Lock-in exit costWhat it costs to change anything after the firm leaves. Zero if files, system, and accounts are yours; a second engagement if they are not.
- Then divide by the metrictrue cost ÷ expected lift on the metric you named. That quotient is the only number that makes two different-shaped proposals comparable.
Show the math
If you cannot fill in the last line because no metric was named, that is the finding. Scope the project until it has one.
What the firm is pricing about you
Having sat on the pricing side, I can tell you the quote is not just about the work. It is about you. Firms price the risk in the client: how many stakeholders have veto power, how fast decisions come back, whether there is real access to users and data, whether the brief will survive its first contact with your leadership team. A project that will need three rounds of internal politics costs more than the same screens for a founder who decides in a day, and experienced firms can smell the difference in the first meeting.
Which means you have pricing power you may not be using. Show up with one empowered decision-maker, a named metric, honest access, and a crisp problem, and you are a cheaper client to serve. Some of that saving reaches your quote, and all of it reaches your timeline.
AI is quietly repricing this market
One more thing worth naming in 2026. AI has compressed the production hours in design work dramatically, the mockups, the variations, the front-end scaffolding, and firms differ wildly in how much of that saving they pass on. Some have rebuilt how they work and price accordingly. Some are still billing the old hours at the old rates and letting the margin quietly widen.
You do not need to audit their tooling. Just ask where AI changes their scope or their price, and listen for specifics. It is the same question I flagged in the evaluation essay, and here it is worth real money: the difference between an AI-honest firm and an AI-silent one, on the same project, can be a meaningful slice of the bill.
The cost that dwarfs the invoice
The most expensive engagement I have ever watched was not the one with the biggest bill. It was a beautiful, professionally run project that solved the wrong problem, because nobody senior stopped to check the brief against the data. The work shipped, the metrics did not move, and the company paid twice: once for the project, and once for the quarter it lost.
A polished project aimed at the wrong problem is the most expensive thing you can buy.
That is the real reason to be careful about who you hire, more than the rate card. Seniority you can trust to say the brief is wrong is the thing actually being priced, and it is the cheapest line on any invoice relative to what it saves.
When a firm is the wrong purchase entirely
A firm is the right buy for big, defined work on a real deadline. If what you actually have is a stuck metric and no clear diagnosis, you need one senior person first, not a team, that is consulting. If you need ongoing direction rather than a project, look at fractional leadership. And if you know exactly what to build and just need hands, a good freelancer beats everyone on price.
If you are trying to place your situation in that map, tell me what you are building and what is stuck. I will tell you which shape fits, including when the answer is a firm and not me.