Most studio owners can tell you what they spend on marketing. Almost none can tell you what a client costs. Those are different numbers, and the gap between them is where studios quietly lose their margin. The spend is the small part. The expensive part is the ten hours you and your lead designer put into a proposal that went nowhere, priced at what those hours would have earned on a live project. This article turns all of it into one number, carries a single worked example from top to bottom, and gives you a threshold you can test your own studio against tonight.
Why the SaaS formula misleads a project studio
Everyone borrowed customer acquisition cost from subscription software, where it makes sense. You spend money in January, you sign accounts in January, you divide, you get a number. The account then pays you every month for three years, so a CAC that is three or four times the first month’s revenue is not alarming. The whole formula assumes a long, predictable revenue tail behind each signature.
Your unit is a project, not a subscription
A design studio signs a project. The revenue arrives once, over eight to sixteen weeks, and then stops. There is no compounding tail unless you build one deliberately through repeat work and referrals. That single structural difference means a CAC that looks healthy against software benchmarks can be quietly fatal for you. If a SaaS company spends 4,000 to win an account worth 400 a month, they are fine by month twelve. If you spend 4,000 to win a project worth 12,000 gross with 45 percent contribution margin, you cleared 5,400 and spent 4,000 of it getting the work. That leaves 1,400 to cover rent, software, insurance, admin and the weeks between projects, and then you start again from zero.
The denominator problem
The second break is timing. In software, spend and signup happen close enough together that dividing one month’s spend by one month’s signups is roughly honest. In studio work, the enquiry that closed this week came from a talk you gave in March, an article someone read in May, and a referral conversation in June. Dividing August spend by August wins is arithmetic performed on two unrelated populations. We will deal with this properly further down, but flag it now, because it is the single most common reason studios conclude their marketing “stopped working” when nothing changed except the lag.
Spend is not the same as cost
The third break is the one that costs the most. Software CAC includes salaries: the sales team, the SDRs, the marketing headcount. Studio owners almost never include their own selling time, because it does not show up on a card statement. It is the largest line in your acquisition cost by a distance, and leaving it out is the reason the number always looks great and the bank balance never does.
The example we will carry all the way through
Every number below is an assumption in a worked example. None of it is research. Swap in your own figures and the method still holds; the point is the shape of the calculation, not my inputs.
The assumptions
Take a three person studio. Average project value is 18,000. Blended effective billable rate across the team is 120 an hour. Direct delivery cost, meaning the people time it takes to actually do the work, runs at 55 percent of project value, which leaves 45 percent contribution before overhead. Monthly external marketing spend is 2,400, covering paid listings, a contractor who edits and publishes, tools, and a sponsorship. That spend produces 20 enquiries a month. Of those 20, six are worth a real conversation, which is a 30 percent qualification rate. Of those six, you win 25 percent, so 1.5 clients a month.
The three easy numbers
Cost per lead is 2,400 divided by 20, so 120. Cost per qualified lead is 2,400 divided by 6, so 400. Cost per won client is 2,400 divided by 1.5, so 1,600. Monthly revenue attributable to those wins is 1.5 times 18,000, so 27,000. Against 2,400 of spend, that is a return of 11.25 to one.
Eleven to one is the number that gets put in a slide and shown to a business coach. It is also close to meaningless, because it counts none of the labour that turned those 20 enquiries into 1.5 signatures.
What the number does not include yet
It excludes every hour of triage, every discovery call, every scoping session, every proposal, every follow up, and every concept you produced on spec to get over the line. It also excludes the fact that those hours are not free even when you are not busy, because a studio at half capacity that spends its slack on pitching is a studio that is not building the case studies, the systems, or the audit process that would make the next pitch easier. Let us price it.
The cost of pitching, which is the real cost
Price your own hours at your own rate
Use your blended billable rate, not your salary divided by hours. If your team bills at 120 an hour, an hour spent writing a proposal is an hour you could have sold for 120. That is the honest opportunity cost when you are near capacity, and it is the only rate that makes the comparison meaningful. There is a fair objection here: when you are at 55 percent utilisation, that hour was not going to be sold anyway, so the marginal cash cost is closer to zero. True, and worth holding in mind. But run your acquisition maths at the billable rate anyway, because the decision you are making is whether to build a machine that consumes half a person forever. You want that decision priced at full freight, not at the discount rate of a slow quarter.
The proposal, itemised
Here is the same month, with the hours logged. Twenty enquiries arrive. Fourteen are not qualified, but they still cost you: reading, a considered reply, sometimes a fifteen minute call to establish that the budget is 3,000 and the deadline was last Tuesday. Call it 30 minutes average across all 20 enquiries, so 10 hours.
Each of the six qualified leads gets a real pursuit. One hour of prep, a one hour discovery call, half an hour writing up notes. Two hours of scoping and estimating, which for anything involving a migration or an existing product is genuinely hard work. Four hours writing and designing the proposal itself. An hour and a half of follow up: the second call, the scope revision, the “can you break out phase two separately” email. That is 10 hours per qualified lead, so 60 hours.
Then, for the two opportunities that reach a shortlist, you go further: a short audit, a homepage direction, a stripped down clickable prototype. Eight hours each, so 16 hours.
Total: 86 hours a month on acquisition. At 120 an hour, that is 10,320.
Fully loaded acquisition cost
Add the 2,400 of external spend. Fully loaded monthly acquisition cost is 12,720 for 1.5 won clients, so 8,480 per won client. On an 18,000 project, that is 47 percent of the contract value consumed by the process of getting it. Against a 45 percent contribution margin of 8,100, you are marginally underwater on every first project.
The spend was 1,600 per client. The cost was 8,480 per client. The difference is 86 hours a month, which is roughly half a person on your team, permanently assigned to selling. Most studios have that person. Almost none of them know it.
Free pitching, spec work, and the invoice you never send
What a free concept actually costs
The 16 hours in the example above is a conservative version of what many studios give away. A “quick homepage direction” for a shortlisted prospect is rarely eight hours. Once you include the research to make it credible, the copy you have to write because the client’s copy does not exist yet, the two rounds of internal critique because you will not show something weak, and the deck to present it in, twenty hours is common in our experience and forty is not rare. At 120 an hour, a forty hour spec concept is a 4,800 gift. Do that four times a year and win half of them and you have given away 19,200 of unbilled work to land two clients, 9,600 each, on top of everything else.
The insidious part is that it is invisible. It never appears in a budget, never gets discussed at a planning session, and never shows up when you calculate whether marketing is paying. It just quietly eats the second half of every second week.
Why clients ask, and why the ask is rational
Do not be offended by it. From the client’s side, choosing a studio is a high stakes decision made with almost no information, and a concept is the only artefact that feels like evidence. The problem is not their motive, it is that a free concept is a bad instrument: produced without research, without stakeholder access, and without the constraints that will actually shape the work. It rewards the studio that guesses prettiest, which is not the same as the studio that would do the best job. If you want prospects to choose well, the more useful thing you can hand them is a clear account of how the decision should be made, which is exactly why we wrote about choosing a web design agency rather than producing more speculative comps.
The paid discovery alternative
The fix that works is to sell the front end. Package the thing you were giving away as a small paid engagement: a two week discovery, an audit with a prioritised roadmap, a costed technical plan, a single key screen designed properly. Price it at 8 to 15 percent of the expected project value so it is a genuine decision but not a barrier. Three things happen. The unqualified prospects disappear immediately, which cuts your pitching hours. The ones who pay arrive at the main project already committed, so your win rate on the follow on proposal goes far above your cold rate. And when they do not proceed, you got paid for the work instead of filing it in a folder called “old pitches”.
Two rules if you do this. Make the deliverable genuinely useful standalone, so the client owns something real if they walk away. And do not credit the full fee against the project unless you are comfortable with the discovery becoming a discount mechanism; a partial credit reads as fair and keeps the engagement honest.
When free work is defensible
There are three cases. A tiny, fixed, timeboxed artefact, under two hours, where the point is to demonstrate how you think rather than what you would make. A client you actively want whose budget is confirmed and whose decision maker is in the room. And a public, self initiated piece you would have made anyway, which is not spec work at all, it is portfolio. Everything else is a transfer of value from your studio to a prospect who has not decided to buy.
Win rate is the cheapest lever you own
The arithmetic of moving it
Go back to the example. Six qualified leads, 25 percent win rate, 1.5 clients, loaded cost 12,720, CAC 8,480.
Now the obvious growth move: double the marketing spend. Spend 4,800, get 40 enquiries, 12 qualified, still 25 percent, so 3 clients. But the pitching hours double too, to roughly 172, which is 20,640. Total loaded cost 25,440 for 3 clients. CAC is 8,480. Identical. You doubled your revenue, doubled your workload, and moved your acquisition efficiency not one cent. You also just committed to 172 hours a month of selling, which a three person studio cannot staff without stopping delivery.
Now the other move: leave spend alone, raise the win rate from 25 to 40 percent. Same six qualified leads, same 86 hours, same 12,720. Now you win 2.4 clients. CAC drops to 5,300. Revenue goes to 43,200. Nothing about your marketing changed.
Now the third move, the one nobody wants to make: pitch less. Take four of the six qualified leads, the four you genuinely fit, and put the same care into them. Triage is still 10 hours, pursuit is 40, one deeper piece at 8. That is 58 hours, so 6,960, plus 2,400, so 9,360. Win two of four at 50 percent. CAC is 4,680, a little over half the original, on lower revenue but far better margin and a calmer team.
Qualification is not rudeness
Qualifying is the act of finding out, early and directly, whether this is a project you can win and should want. Ask about budget on the first call, in a number or a range, and say your own range first so it is not an interrogation. Ask who signs. Ask what happens if they do nothing. Ask what else they are considering and who else they are talking to. Ask when the decision gets made and what triggers it. If you cannot get answers to four of those five, you are not in a sales process, you are in someone’s research phase, and you should price your participation accordingly.
The most useful disqualifier we use is simple: has this client bought design before? A company buying its first serious website behaves differently from one on its third. Neither is bad, but the first needs education time that must be priced in, and the second has a shorter cycle and a higher close rate. Our writing on what a custom website costs exists partly so that education happens before the call rather than during it.
Better briefs win more
Win rate is largely decided before you write a word of the proposal. A vague brief produces a vague proposal, which loses to a specific one, and the specificity has to come from the client. The cheapest intervention available to you is to shape the brief yourself: send a short set of questions before the discovery call and treat the answers as the agenda. We keep a public guide on how to brief a design agency and send it to prospects before the first call. A prospect who fills it in has spent an hour on your project before you have spent one on theirs, and people who have done that convert at a different rate from people who have not.
Declining a poor fit is cheaper than winning it
The hardest habit to build is declining work you could technically do. Every yes to a poor fit project costs you the 10 to 18 hours of pursuit, and then, if you win it, costs you again in delivery overrun, scope arguments, and a case study you cannot publish. Two of those a year is a quarter of your capacity spent producing nothing you can sell the next client with. When you decline, be concrete and generous: tell them why the fit is wrong and name the kind of studio or freelancer that would serve them better. In our own log, roughly one in five of those referrals comes back later with a project that does fit.
What a client is actually worth
First project value is the wrong denominator
If you judge acquisition against the first project alone, you will underinvest in marketing and overinvest in discounting. The right denominator includes what the relationship produces afterwards, which for studios comes in three forms: repeat projects, expansion into retained work, and referrals.
A conservative lifetime value
Carry on with the example. First project, 18,000. Assume 40 percent of clients come back within 24 months for a second engagement averaging 12,000, a phase two, a new product surface, a rebuild of something that has aged. Assume 25 percent produce one referral that closes at the studio average of 18,000, with an acquisition cost close to zero.
Lifetime revenue is 18,000 plus 0.4 times 12,000 plus 0.25 times 18,000, which is 18,000 plus 4,800 plus 4,500, so 27,300.
Against a fully loaded CAC of 8,480, that is 3.2 to one on revenue. Considerably worse than the 11.25 you calculated at the top, and considerably better than the 2.1 you would have got from first project value alone. Both corrections matter.
Now do it on gross profit
Revenue ratios flatter service businesses because delivery is expensive. At 55 percent direct cost, lifetime contribution is 27,300 times 0.45, so 12,285. Against a CAC of 8,480, the ratio is 1.45 to one. That is the honest number, and it is thin. It means the studio in this example keeps about 3,800 per client after acquisition and delivery, before rent, software, admin, insurance, holiday, and everything else. Two bad projects a year wipes it out.
That is closer to typical than most studios want to believe, and the fix is not more leads. It is win rate, pitch discipline, and repeat rate, in that order.
The repeat rate lever
Repeat and referral rates are the only two inputs that improve your economics without any marketing at all, and both are decided by delivery rather than sales. The handover, the first month after launch, the responsiveness in week six when something breaks: that is your growth engine. We wrote about the first 30 days after launch because that window converts a delivered project into a relationship or into a one off. Moving repeat rate from 40 to 60 percent in the example adds 2,400 of lifetime revenue per client and costs you nothing but attention.
Payback, and the cash trap that kills growing studios
When the money actually arrives
Payback period is how long after you start spending on a client before you have recovered what you spent. Take a typical timeline: first touch in week zero, qualified call in week four, proposal in week six, decision in week twelve, contract signed and deposit paid in week thirteen. Delivery runs ten weeks with 40 percent up front, 30 percent at midpoint, 30 percent on completion, so the final payment lands around week twenty five, assuming nobody pays late, which somebody will.
On a pure cash basis, payback is fast, because the actual money you spent was only 1,600 and the deposit is 7,200. On a fully loaded basis, where the 86 hours have to be recovered out of contribution, you clear the 8,480 somewhere around week twenty two, roughly five to six months after first contact. That is your true payback period.
The cash trap
Now scale it. If you decide to grow, you increase pitching hours immediately and receive the revenue five to six months later. Those pitching hours come out of the same people who deliver the current work, so utilisation on billable projects falls in exactly the months where you need cash the most. This is why studios that decide to “push harder on sales” often have their worst quarter immediately afterwards. Every client you win this month is funded by clients you won half a year ago. Plan a growth push with at least four months of runway for the dip, or fund the selling with someone whose time is not billable.
Retainers change the shape
The single structural change that fixes payback is converting some project revenue into ongoing revenue. A client on a monthly arrangement has an acquisition cost paid once and a revenue line that continues, which is the SaaS shape the original formula assumed. This is the practical argument in the retainer versus project comparison: not that retainers are more profitable per hour, often they are not, but that they make your acquisition maths survivable.
The sales cycle nobody measures
Lag destroys naive attribution
If your median time from first touch to signature is 90 days, with a tail running past a year, then this month’s revenue and this month’s spend belong to different cohorts. Cut spend in January and revenue looks fine through March, so you conclude the spend was doing nothing. April arrives and the pipeline is empty. Then you switch it back on and see nothing for a quarter, and conclude it does not work anymore. Both conclusions are wrong and both are extremely common.
Cohort by first touch
The fix is to stop measuring by close date and start measuring by first touch date. Every enquiry gets tagged with the month it first appeared. When it closes, the revenue is credited back to that month, not the month it signed. Your April cohort keeps accumulating revenue through October. This means you accept a hard truth: you cannot judge a month’s marketing for at least two quarters. Report on it anyway, with the cohort still open, and label it as provisional.
Attribution honesty
Ask every client, in conversation, how they came to talk to you, and write down what they actually say rather than what your form field recorded. You will hear things like “I read something of yours a year ago, then a friend mentioned you”. Record both. Multi touch attribution for a studio doing twenty enquiries a month does not need software, it needs a column called “first touch” and a column called “what made them contact us”. The gap between those two is where most of your marketing budget is really working, and it is invisible to any analytics tool. That said, do measure the tool measurable parts properly; our notes on what to actually measure on a website apply to your own site as much as your clients’.
Cheap leads and winnable leads are different products
Cost per lead is a vanity metric
A channel that delivers leads at 40 each and qualifies at 5 percent costs you 800 per qualified lead. A channel that delivers leads at 400 each and qualifies at 60 percent costs you 667, and crucially it consumes a fraction of the triage hours. Since triage hours are a real cost in your loaded CAC, expensive leads are frequently cheaper. Optimise for cost per qualified lead first, and cost per won client second. Cost per lead should never appear in a decision.
The four signals of a winnable lead
A winnable lead usually arrives with most of these: a stated budget or a willingness to state one, a date driven by something real such as a funding round or a product launch, a named decision maker on the first call, and a written description of the problem rather than a description of the deliverable. “We need a website” is a deliverable. “Our sales team keeps losing deals because prospects cannot tell what we do” is a problem, and problems convert.
Score every enquiry out of four on those signals in your log. After six months you will know exactly which sources produce threes and fours and which produce ones. That is the report that should drive your channel decisions, not traffic. Which channels to run, and in what order when you have nothing to spend, is a separate subject with its own article; the only point here is that you cannot judge a channel until you have scored the leads it sends you.
Portfolio hours are acquisition cost and nobody books them
Case studies never appear in acquisition maths, because nobody invoices them. Put them in. A case study that takes six hours of your time is 720 at a 120 blended rate, and it belongs in the month you wrote it, not the month it produced an enquiry. Book a year of them that way and the activity you have been treating as a chore between projects usually turns out to be the cheapest line in your loaded CAC table, because it is the one channel where the cost stops and the enquiries do not.
The lag is what makes this hard to read. The piece you publish in March converts someone in November, so for two quarters it looks like pure cost, and then in the month it lands it looks free. Neither reading is right. Charge the hours to the month they were spent, credit the win to the first touch date, and judge the asset only when the cohort closes. How to write a case study so it converts at all is a different question, and we published a separate playbook for it rather than repeat it here. The accounting point is narrower: unbilled hours are still cost, and portfolio hours are the ones studios forget to count.
Instrumenting all of this without a CRM
One spreadsheet, one row per enquiry
You do not need a CRM at twenty enquiries a month. You need one sheet you actually fill in. Columns: date of first touch, name, company, source as stated by them, source as recorded by your site, the four signal score, budget stated, qualified yes or no, reason if no, hours spent in pursuit, outcome, close date, contract value, and delivery hours once it ships.
The column that matters most and that nobody keeps is hours spent in pursuit. Without it you cannot compute loaded CAC, and loaded CAC is the entire point. Log it in fifteen minute increments, in the moment, for ninety days. Reconstructed at the end of the month it is fiction: nobody remembers the forty minutes they spent rereading a brief on a Sunday, and those are exactly the minutes that make the number honest.
The monthly ritual
Thirty minutes on the last working day of the month. Sum external spend. Sum pursuit hours, multiply by your blended rate. Count enquiries, qualified leads, wins. Compute cost per qualified lead and loaded CAC. Update the first touch cohorts with anything that closed. Write two sentences about what changed. Twelve of those and you have a real picture of your business, at a total cost of six hours a year.
What not to bother with
Do not build dashboards. Do not install a tracking stack that takes a fortnight. Do not attempt per channel attribution modelling at this volume; the sample size cannot support it and you will draw confident conclusions from noise. Three data points a month for a year is enough to see a trend, and a trend is the most that twenty enquiries a month can honestly support.
The test: is my marketing working?
The threshold
Here is the test we use, and it is deliberately strict.
Take the trailing twelve months. Compute fully loaded acquisition cost: external spend plus all pursuit hours at your blended billable rate. Compute gross profit on the first projects won in that period, meaning contract value minus direct delivery cost, ignoring repeat and referral entirely. Divide.
If first project gross profit is less than 2 times fully loaded acquisition cost, your acquisition is not working, regardless of how good the revenue looks. Between 2 and 4 times, you have a functioning machine that is fragile: it survives, but a single bad quarter or one overrunning project erases the year. Above 4 times, you are underspending and should push more money and more attention into whatever is producing your best leads.
The reason the threshold sits on first project gross profit rather than lifetime value is that lifetime value is a forecast and forecasts are where wishful thinking hides. Repeat and referral revenue should be upside that makes a good business great, not the assumption that makes a bad one look acceptable.
Three failure modes and what each looks like
The lead problem: cost per qualified lead is high, win rate is fine. You are not reaching enough of the right people. This is the only failure mode where spending more money is the correct response.
The conversion problem: qualified leads are plentiful and cheap, win rate is under 20 percent. Do not spend another cent on marketing. Fix the pitch, the qualification, the pricing conversation, and the way you present work. Studios lose more deals in the presentation than in the proposal document.
The margin problem: win rate is good, acquisition cost is reasonable, and the ratio is still under 2. Your projects are underpriced or overrunning. That is a delivery and pricing problem wearing a marketing costume, and no amount of lead generation fixes it. It is the same underlying failure we describe from the buyer’s side in the real cost of a cheap website, except this time you are the one absorbing it.
When the numbers say stop
Cut, fix, or wait
If the test fails, work out which of the three it is before you touch anything. Cut a channel only when it has produced a full cohort’s worth of data and its qualified lead cost is more than double your next worst channel. Fix the pitch when the win rate is the problem, and give it at least ten opportunities before judging the change. Wait when the only issue is that you started four months ago, because almost nothing in studio marketing produces a readable signal inside one sales cycle.
Where the money goes instead
When acquisition is not paying, the highest return use of the same money is usually not more marketing. It is reducing the cost of the pitch: templates and reusable scoping components so a full pursuit takes six hours instead of ten, a standing paid discovery offer so spec work stops, a pricing structure you can explain in two minutes, and two or three case studies written properly for buyers. Every one of those raises win rate or lowers pursuit hours, and both go straight into the denominator of the number you are trying to fix. Run that through the example we carried through this article. Cut pursuit time per qualified lead to six hours, and pitch four leads instead of six: 10 hours of triage, four pursuits at six hours, one deeper piece at eight, so 42 hours. At 120 an hour that is 5,040, plus the 2,400 of spend, so 7,440. Hold the win rate at the 50 percent you get from pitching only the leads you actually fit and that is two clients, so a loaded CAC of 3,720 against the 8,480 you started with, without changing a single thing about your marketing.
If you want a sense of how we structure the front end of an engagement so the scoping work does double duty as sales work, the process we describe for a website redesign is the same one we use to write a proposal: understand the current state, name the constraints, sequence the work, price the sequence. The document that wins the project and the document that runs the project should be the same document.
If you are working through your own version of these numbers and want a second pair of eyes on where the leak is, write to us at hello@beconfidency.agency. We will tell you what our own figures look like, including the parts that are not flattering, because comparing methods with another studio is more useful than any benchmark you can find published anywhere. There is no pitch attached to that conversation and no concept deck at the end of it.
Log your pursuit hours for the next ninety days, because you cannot fix a number you have never measured.
And when the pursuit maths says the leak is the site doing the convincing, that is exactly what our web design service is for.