Your day rate is a number you can say out loud without checking. Last quarter’s marketing cost probably is not, because it never appeared on an invoice. It came out of Thursday afternoons, out of the ninety minutes before anyone else was awake, out of the gap between a handoff and the next kickoff. That is the real ledger, and most studios never open it. This is an argument for keeping it, and for what changes once you do.
Your constraint is hours, and the P&L cannot see it
“We cannot afford marketing” almost never means money
When a two-person studio says it cannot afford to market itself, it is rarely talking about a budget line. Nobody is agonising over a 40 dollar a month analytics plan. What they mean is that there is no room in the week: the calendar is already full of client work, revisions, scoping calls, invoicing, and the hours that vanish into Slack. Because the problem gets phrased as money it gets answered as money, usually with a subscription or a cheap freelancer, neither of which solves anything.
The founder is the only person who can do most of it
In a studio under ten people, the parts of marketing that actually move work into the pipeline are the parts only the founder can do. The point of view. The opinion that makes someone forward the link. The decision about which client problem is worth writing 2,000 words on. The follow-up email to a lead from March that lands because you remember their budget cycle. You can hire out formatting, scheduling, image prep, and page building. You cannot hire out the part that makes the work distinctive, which means the input is your hours specifically, not generic hours you can buy at market rate.
Cash and time behave differently, and you budget one of them
If you spend 3,000 dollars on something that fails, you are down 3,000 dollars and your week is intact. If you spend 40 hours on something that fails, the money is intact and your delivery capacity for the month is gone. Cash is fungible and replaceable; a week in August is not. Studios plan cash carefully and plan founder time not at all, which is exactly backwards for the constraint they actually face.
How to price your own hour honestly
The three numbers you need before anything else
Work out your realistic annual billable hours, your effective hourly rate, and your utilisation ceiling. Most founders overestimate all three. Take 52 weeks, subtract holidays, illness, dead weeks between projects, and the admin days that produce nothing billable, and you land near 46 working weeks. Then accept that a founder who also sells, scopes, reviews, hires, chases invoices, and handles the awkward client call is not billing 40 hours out of 40. The commonly repeated rule of thumb for a working owner is 50 to 65 percent utilisation, and the honest end of that range is the lower one.
Why “it is free, I did it myself” is the most expensive line in small studio accounting
That sentence is how a studio hides its single largest cost from itself. If you write a case study over three evenings, your books show zero and your capacity shows twelve hours gone. Do that weekly and you have spent the equivalent of a junior salary with no entry anywhere. The consequence is not guilt, it is bad decisions: because the cost is invisible you never ask whether it was worth it, and never notice you spent nine hours re-cutting a portfolio grid that changed nothing while the follow-up emails that close deals sat unsent. The fix is to make the cost visible by logging those hours somewhere you will actually look at them. The total will be uncomfortable inside a month, which is the point.
Opportunity cost is not your salary
The cost of a marketing hour is not what you pay yourself. It is the revenue that hour would otherwise have produced, and only when the hour genuinely had a billable alternative. In a booked week, a marketing hour costs your full rate. In a thin week it costs nothing in cash. Price the hour by the state of the week, not by an average.
A worked example: what eight hours a week actually costs
The inputs, stated so you can argue with them
Assume a founder-led studio. Rate of 125 dollars an hour. 46 working weeks. A 40 hour nominal week, so 1,840 available hours. Utilisation ceiling of 60 percent, giving 1,104 billable hours and roughly 138,000 dollars of founder-delivered billings a year. These are assumptions for the arithmetic, not measurements. Swap in your own numbers; the shape holds at 60 dollars an hour and at 250.
Eight hours a week, priced two ways
Eight hours a week for 46 weeks is 368 hours a year. Priced naively at full rate, that is 46,000 dollars of foregone revenue, a third of your billings. That number is wrong in one direction, because it assumes every one of those hours displaced paid work. It is also wrong in the other, because it assumes the alternative was rest.
Split it. Suppose that across the year you are genuinely capacity-constrained in half the weeks. In those weeks the marketing hours displace billable work directly: 184 hours at 125 dollars is 23,000 dollars of revenue you did not invoice. In the other half, the hours had no billable alternative, so the cash cost is zero, but you spent 184 hours of finite personal capacity, and capacity spent is capacity unavailable for the thing that goes wrong in week 31.
So the honest annual cost of an eight hour marketing week for this studio is about 23,000 dollars plus 184 hours of your life. Not free. Not 46,000 either. If you never run this calculation you will oscillate between treating it as costless and treating it as ruinous, usually within the same month.
The version most people are actually running
Most small studios do not spend eight steady hours a week. They spend zero for seven weeks and then twenty-two hours in one panicked week when the pipeline goes quiet. Same annual hours, much worse return, because the panic week produces your weakest output at the moment you have least leverage.
What that spend has to return to break even
Break even, counted in hours
Keep the whole calculation in one unit and it stays checkable. The 23,000 dollars of foregone revenue is 184 hours you did not bill. At 125 dollars an hour, an 18,000 dollar project is roughly 145 founder hours of delivery. So the marketing year has to produce a bit over one additional project to buy those hours back, and about two before it is clearly worth doing. That is the entire test: one extra project. You can look at the last twelve months and answer it, which is more than you can say for a traffic target, and it is the number to put on the table when someone tells you the answer is to post more often.
The rate lever is usually bigger than the volume lever
Marketing does not only bring more work, it changes which work arrives and what it pays. If publishing consistently moves your average rate from 125 to 140 dollars across the same 1,104 hours, that is 16,560 dollars with no extra delivery hours and no extra risk. That is most of the break even, achieved through positioning rather than volume. It is also why writing about the specific work you want more of beats writing whatever ranks. Clients who arrive having read three of your articles argue about scope; clients who arrive from a directory argue about price. If you are still deciding what to specialise in, note that the studios who win on rate almost always narrowed first, which is also the difference buyers weigh when they compare a freelancer against an agency.
The third return nobody counts: shorter sales cycles
A prospect who has read your process article and looked at two case studies arrives already half-sold. Your proposal call goes from ninety minutes of explaining how you work to thirty minutes of scoping. Across twelve deals that is twelve hours returned to you, a day and a half of capacity that cost nothing extra to create, and the conversation itself starts at scope rather than at whether you are credible. Publishing how you actually run a project, in the same detail you would use in a kickoff, does more for close rate than any amount of polish on the portfolio grid. It also filters: people who dislike your approach opt out before the call, which is a saving even when it feels like a loss. Documenting what you expect from clients has the same effect, which is why a piece on how to brief a design agency tends to earn its hours back faster than a trends roundup.
The feast and famine cycle, mechanically
Step one: you get busy, and marketing is the only thing with no deadline
Two projects land in the same fortnight. Client work has dates, invoices, and someone waiting. Your article has none of those. It is the only item on the list whose deferral produces no immediate consequence, so it is the item that gets deferred, every time, by every rational person under load. Nothing about this is a discipline failure. It is what happens when the only work with a hard deadline is the work someone else set.
Step two: the lag hides the damage for two months
Pipeline responds slowly. Someone who reads your case study today might have a live project in six weeks, might be a year out, and if they do come back it will be through a search or a forwarded link you never see. Plan against an assumed lag of two to three months, call it 8 to 12 weeks, between the work you do and the enquiries it produces. Whatever the true figure is in your studio, it is long enough to hide the mechanism: for two months after you stop, the enquiries keep arriving from the work you already did, which reads as evidence that stopping was fine.
Step three: the cliff, and what it costs
Then the projects finish, the enquiries you stopped generating in June fail to arrive in September, and you have a quiet month. Five weeks at 25 percent utilisation instead of 60 is roughly 70 lost billable hours, or 8,750 dollars at our assumed rate. Worse is the pricing damage: a studio with an empty October discounts to close. Ten to fifteen percent off an 18,000 dollar project is 1,800 to 2,700 dollars given away, and that number becomes the anchor for the referral that client sends you next year. The famine does not just cost the gap, it lowers the ceiling afterwards.
Step four: the panic, and the loop closes
Quiet month means you finally have time, so you market hard for three weeks, land two projects, get busy, and stop. The pipeline empties again 8 to 12 weeks later. This is not seasonality and it is not the market. It is a feedback loop with a delay, and delayed feedback loops always oscillate unless something outside the loop holds the input steady. The thing outside the loop is a fixed weekly block that does not move when the calendar fills.
The asymmetry that makes marketing worth the hours anyway
A billed hour is spent once
You sell an hour, you deliver it, it is gone. Next month you start from zero and sell it again. That is the deal with services, and it does not compound. Nothing you did in March makes April easier unless you deliberately made it so.
A published asset is paid for once and drawn on for years
An article you wrote two years ago can still be answering a question for someone who then books a call. A case study written once gets sent to every prospect in the category for years. The asymmetry is not that publishing is easy, it is that the cost is paid once and the return is drawn indefinitely. Ten hours writing up a project properly, the way we did with the US Construction build, keeps returning for as long as anyone in that sector is searching for someone who has done it before. Ten hours of billable work returns 1,250 dollars and stops.
The compounding is invisible until it is obvious
For the first six to nine months, consistent publishing looks like it is not working, because almost nothing is. Then several things start firing at once: search picks up the older pieces, someone forwards one internally, a prospect mentions an article you forgot writing. While you are inside the flat stretch it is indistinguishable from failure, which is why the decision has to be made once, at the start, and not renegotiated in month four when the evidence is still absent. If you cannot commit for four quarters, do not start; spend the hours on direct outreach instead, where feedback arrives in days.
Not everything compounds, so know which hour you are spending
Before an hour goes anywhere, decide whether it produces something that keeps working or an effect that stops when you stop. Case studies, technical explainers, your services pages, and your about page sit in the first group. Social posts, conference chat, and cold outreach sit in the second. Neither is the better use of an hour, but they behave differently on a calendar, and that is what matters when the calendar is the scarce thing. Compounding work is slow and silent, so it has to be scheduled in advance and defended. Non-compounding work pays back this week, so it always feels more urgent, and it will quietly take every hour you leave unprotected. That is the actual failure: not choosing wrong, but never choosing, and discovering after three years that the fast half absorbed the entire budget.
Protecting the hours structurally
Treat your studio as a client on a retainer
You will not remember to market when busy, and anything that relies on you feeling like it at 4pm on a Thursday in the middle of a launch is already dead. So give yourself a client called Studio. Put four hours a week in the calendar as booked time. Track the hours against it. Set deliverables and dates the way you would for anyone paying you. When a client asks for those hours, respond exactly as you would if another client held them: you are booked, here is the next slot. That sounds like theatre until you notice the whole problem is that internal work has no counterparty, so there is nobody to disappoint by moving it. Inventing one costs you nothing and removes the excuse, and it is the same logic behind a retainer rather than one-off projects: continuous small commitments beat heroic occasional ones.
Fixed blocks, at the least valuable hours you have
Pick the two hours of the week when your billable output is worst anyway. For most people that is late Friday or the first stretch of Monday before anyone is awake in the client’s timezone. Marketing done in your sharpest hours costs the most; done in the dead zone it costs much less and still produces. Just do not put it at the very end of a day, where overruns will eat it.
Batch by mode, not by task
Switching between design work and writing costs real time, more than the task list suggests. Group by the kind of thinking involved: one block for writing, one for capture and admin, one for outreach and replies. Four separate 30 minute slots produce far less than one 2 hour block. Our working assumption, from watching our own tracked hours, is that the first ten to fifteen minutes of any slot goes to reloading context, which in a 30 minute slot is half of it. Treat that as a rule of thumb rather than a constant and it still tells you to stop chopping the week into fragments. It is the same reason your design work suffers on days full of calls.
Pick the cadence you can sustain in your worst month, not your best
Decide your publishing rhythm by asking what you could do during the busiest fortnight you had last year. If the honest answer is one substantial piece a month, commit to one a month and never break it. One a month for two years is 24 assets and a reputation for reliability. Four a month for six weeks, then silence, is 6 assets and evidence that you cannot sustain anything, which is exactly the impression you do not want in front of buyers.
The minimum viable weekly routine when you have no spare time
Three hours, split into three purposes
If four hours is impossible, run three, split like this. Sixty minutes early in the week on pipeline: five follow-ups on open conversations, one message to a past client with something useful and no ask, one revisit of a lead that went quiet more than three months ago. Ninety minutes mid-week on the one publishable asset in progress, which is always in progress, never started from scratch. Thirty minutes at the end of the week on capture, which I will come back to because it is the part that makes the other two cheap.
The follow-up hour is the one to protect first
If you only ever protect one of the three, protect the first. A dormant lead already knows your price, your process, and whether they liked working with you, so the conversation starts at scoping rather than at trust, and none of the expensive groundwork has to be laid twice. The work is 60 minutes of short emails to people who already know you, with no research and no new thinking required. It is easy to spend twelve hours on a new article while a client from eighteen months ago, whose site is now visibly out of date, hears nothing from you. The point of the first 30 days after launch is that the relationship does not end at handover, and the follow-up hour is where that belief becomes revenue.
Never start from a blank page
The ninety minute block should always open something half-built. Our own habit, and the reason we stopped scheduling writing without a starting point, is that a blank page swallows something like the first half hour before a usable sentence appears, which in a 90 minute slot is a third of the budget gone before anything exists. Treat that as a rule of thumb and assume the cost rather than hoping for a good day. Keep one document with the current piece in whatever state it is in, plus a list of the next five topics drawn from real client questions. When the block starts, you continue; you never decide.
Capture as you go, not afterwards
Writing is expensive because you are reconstructing, not recording
Most of the cost of writing a case study is remembering. Why did you drop the tabbed layout? What did the client say in week three that changed the information architecture? What were the load times before? Reconstructing that six weeks after handover takes hours and produces vague prose. Recording it as it happens takes ninety seconds and produces the specific detail that makes the piece worth reading.
What to capture, in thirty minutes a week
Screenshots at every meaningful state change, before and after. One line on any decision you had to argue for. The exact sentence a client used when they described their problem, because it is almost always better copy than anything you would write. Any number you measured: page weight, time to first meaningful paint, form completion, support tickets. Drop them in a folder per project. When it comes time to write, you are assembling rather than remembering, and a piece that would have taken ten hours takes four.
Capture also fixes the honesty problem
Written months later, every case study becomes a smooth story where everything went to plan. Written from notes, it contains the wrong turn in week two, which is the part other people learn from and the part clients find credible. Our writeup of the Pulse Clinic work is more useful because the constraints are in it. The specific reason you moved a client from one stack to another beats any general comparison, and it is only available if you wrote it down at the time.
What to delegate first, and what you can never delegate
Delegate the mechanical layer, in this order
Production and distribution go first: formatting, image preparation, publishing, metadata, internal linking, repurposing one article into shorter posts, scheduling, and pulling the monthly numbers. Then research support: gathering references, checking claims, building the comparison tables you will not use directly. Then first-draft transcription, where someone records you talking for 45 minutes and returns a structured draft you edit. Set that last one up early: talking is fast, typing is slow, and your spoken explanation of a project already contains the opinion you would otherwise spend two hours trying to locate on the page.
What stays with you permanently
The point of view. Which fights you pick. The opinions that make a piece worth forwarding, including the ones that lose you a certain kind of client. The relationships: nobody can send that follow-up email as you. The judgement about what is worth saying at all. If you hand these over, what comes back is competent, generic, and indistinguishable from every other studio’s output, which means it does not do the one job you needed it to do. The same rule governs where AI belongs in a design workflow: it is very good at the mechanical layer and structurally incapable of having a position, and there is a real list of situations where you should not use it at all.
The tell that you delegated the wrong half
Read what came back. If you could put a competitor’s logo on it and nothing would look wrong, you outsourced the wrong layer. Fix it by keeping the argument and handing off everything surrounding the argument.
When money can buy the hours back
Contractors are cheaper than your hours, once you can brief them
The moment your effective rate exceeds what a competent contractor charges for the mechanical layer, buying hours back is straightforwardly correct arithmetic. At 125 dollars an hour, paying someone 40 to buy back four hours a week costs 160 and frees 500 of billable capacity. What kills this trade is management overhead: a contractor who needs two hours of briefing to save you three is not worth it in month one, and only becomes worth it if you keep them long enough to amortise the briefing. Hire for eighteen months, not for one campaign, and write the briefing down once so the next person inherits it.
Paid distribution is a different decision, and a later one
Spending money on ads is not a substitute for having something worth clicking. If your positioning is unclear or your site does not convert the traffic it already gets, paid spend buys you a faster route to the same result. The sequencing is: make the site work, then send more people to it. That means the conversion basics first, from landing page structure to credible social proof, before any media budget. The mechanics of channel choice and acquisition cost are their own subject and worth testing properly rather than guessing.
Buying tools is almost never the answer
A new CMS, a new analytics stack, or a new site build will not create hours. They can remove friction, and a publishing setup that takes 40 minutes of fighting per article is a genuine tax worth fixing, which is the honest case for a CMS your team can actually use. But a studio that publishes nothing on a slow stack will publish nothing on a fast one. Fix the cadence first, then remove the friction the cadence exposes.
How the answer changes as headcount grows
One to three people: the founder is the entire function
Everything above applies at full strength. Protect three to five hours, accept a modest cadence, and prioritise compounding assets and the follow-up hour. Do not attempt multi-channel anything. One publishing rhythm and one relationship habit, held for two years, beats a plan you abandon in March.
Four to eight: split creation from production
You now have enough people that someone other than you can run production, and enough revenue that the contractor arithmetic clearly works. The founder’s contribution drops to roughly two hours a week of raw material: talking through a project, marking up a draft, approving an angle. The trap at this size is that the founder’s hours become more valuable, not less, because they are now the only route to new business at the top end, so the temptation to skip the two hours gets stronger just as the cost of skipping them rises.
Nine and up: someone owns the calendar, you still own the opinions
At this size marketing needs an owner with the cadence in their job description and the authority to chase you for your two hours. The founder becomes an editor rather than an author, and the failure mode flips: plenty gets published and all of it is generic, because the owner never gets enough founder time to extract a real position. Budget those sessions, or you get volume without voice.
The handover test
Whatever your size, run this test annually: if you were unreachable for six weeks, would anything publish, and would any follow-ups be sent? If the answer is no, you do not have a marketing system, you have a habit that lives in one person’s head and dies with their next busy quarter.
Measuring it without lying to yourself
Track two things: hours in, and where work came from
Log the marketing hours honestly against the Studio client, and ask every enquiry how they found you, in the actual conversation, not through a form field people ignore. Those two numbers give you cost and origin, which is enough to decide whether to continue. Everything else is decoration at your volume.
At small volume, judge on twelve months, not twelve weeks
If you close 12 to 20 projects a year, quarterly numbers are noise. Two extra referrals in Q2 tell you nothing about whether the writing works. Look at twelve month windows and at leading signs you can feel: are people arriving already knowing how you work, are proposal calls getting shorter, is your average project value moving. Those are visible long before the analytics are. For the parts you can measure properly, decide upfront what is actually worth measuring rather than watching a traffic chart that has no relationship to your revenue.
The one number that should make you uncomfortable
At the end of a year, put your logged marketing hours next to the work that came from them. If 368 hours produced one project you would have won anyway, stop and change something. If it produced three projects, a higher average rate, and shorter sales cycles, then those hours returned more per hour than delivery did, and the correct move is to protect them harder next year rather than let the next busy quarter take them back.
If you are working out how to structure this in your own studio and want to compare notes, write to us at hello@beconfidency.agency. Send your rough hours and where last year’s work actually came from, and we will tell you what we would change and what we would leave alone.
And when the thing eating those hours turns out to be the site itself, the publishing setup that fights you every time you have something to say, tell us what it is doing and we will look at it.
If the hours keep going into a site that fights you instead of selling for you, that is exactly what our web design service is for.