Ask how much a managed service provider (MSP) should spend on marketing and you will be handed a percentage of revenue. Which percentage depends on who you ask. 2% to 5% to hold your position. 8% to 10% to grow. 10% to 20% to expand. 11% to 12% of topline, with sales salaries and commissions folded in. Some of those numbers count sales and some count marketing alone, and none of the pages I read reconciles them. None of them knows your margin.
The best-sourced number I found comes from The CMO Survey, a survey of chief marketing officers (CMOs) and their peers run by Duke University's Fuqua School of Business with Deloitte and the American Marketing Association. Its 2026 edition found that marketing budgets fell to 9.0% of company revenues. It is a real survey. It is also 308 marketing leaders, 97% of them at vice president level or above, at for-profit companies in the United States. Those are companies with marketing departments run by vice presidents. An owner-run MSP in the $2M to $10M band is not one of them, and borrowing their ratio is borrowing a cost structure you do not have.
My view is that a percentage of revenue is the wrong unit for a business with MSP margins. Set the budget from the gross profit one client returns over its life. Run one channel until you know what a client from it costs and it is paying that back. Then let its results fund the next channel.
For ten years my job was pricing businesses from the buyer's side. Since then I have founded a growth agency that works with MSPs, and today I am fractional Chief Revenue Officer (CRO) at a national MSP in the United States. A buyer's team reads a marketing budget as a claim on margin, so that is how I read it here.
If you want the margin picture first, read the honest MSP margin benchmark, and for per-seat pricing, how much an MSP should charge.
A percentage of revenue ignores the margin it comes out of
Gross margin is the share of revenue left after the direct cost of delivering the service. The trouble is that "direct cost" can be counted generously or honestly, and the answer moves a lot.
At an MSP I work with, the three per-seat service tiers show gross margins of 89%, 77% and 60% at list price when the only cost counted is the tools. Count the technician labor that delivers the service, and you get what I call fully loaded gross margin: 49%, 52% and 39%. Same clients, same prices, far less room in every tier. The order changed too. The tier that looked best on tools alone is not the best once the people are in.
Fully loaded gross margin counts the technician labor that delivers the service. Same clients, same prices.
Now put a percentage budget on top of that. A marketing line of 10% of revenue is the same number whether the margin underneath is 89% or 39%, but it is a very different share of what the business keeps. The percentage cannot see the difference. Your bank account can.
The profit line below gross margin is thinner still. The benchmark I linked above puts average MSP earnings before interest, taxes, depreciation and amortization (EBITDA, roughly operating profit before financing and accounting charges) at around 15%, and finds that roughly a third of MSPs run at a loss. Simple arithmetic on those figures: at 15% EBITDA, a marketing budget of 10% of revenue would consume two-thirds of the profit. At 5% it would consume a third. Your own numbers will differ. This is what the percentage rule does to a business shaped like the average one, and it is why owners who have been told "spend 10%" feel, correctly, that someone is guessing with their money.
The right unit is the gross profit one client returns
Lifetime value (LTV), as I use it, is gross profit, not revenue: what a client pays you over the whole time it stays, multiplied by your gross margin. It is the same definition I use in the explainer on client lifetime value and unit economics. Customer acquisition cost (CAC) is everything you spend to win one new client, meaning the channel spend plus the agency or staff time behind it. Divide the first by the second and you know how many times over a new client pays back what it cost to win.
Take one client as a worked example. It pays you $150k over its life, and it cost $15k to win. At a gross margin of 40% to 50%, which is roughly where the fully loaded figures above sit, that client returns $60k to $75k of gross profit. Against $15k of acquisition cost, LTV to CAC is 4x to 5x. Run the same client through that MSP's actual fully loaded tier margins of 39% to 52% and you get $58,500 to $78,000 of gross profit, or 3.9x to 5.2x.
I have gotten this wrong out loud. I have described exactly this client, $150k against $15k, as a 10x ratio, and argued that MSPs need a higher ratio than enterprise software as a service (SaaS) companies, which are happy at 3x, because MSP margins are smaller. The instinct was right. The arithmetic was not. $150k is revenue, so the 10x was a revenue ratio sitting next to a margin argument. On gross profit the same client is 4x to 5x.
That is still above 3x, and I think it should be. A business whose margin goes to paying the technicians who deliver the service needs more headroom per client than a software company does. It just does not have 10x of it, and anyone budgeting as if it did will overspend.
Run the arithmetic backward and you get a budget. This is an illustration, with round numbers, not a benchmark. A client worth $150k over its life at 50% gross margin returns $75k of gross profit. If you want 5x back, you can afford $15k to win the next client like it. If you would accept the 3x that enterprise software companies are happy with, you can afford $25k. That figure is your allowable acquisition cost: the most you can pay to win one more client and still get the return you chose. Multiply it by the number of new clients you are trying to add and you have a marketing budget built from your own economics, with no borrowed percentage anywhere in it.
The fair objection is that MSP margins are too thin to leave much for marketing at all. With a third of MSPs losing money, that deserves a straight answer. Thin margins lower the ceiling, and the calculation above already carries that: plug in 39% instead of 50% and the allowable cost falls with it. If the ceiling comes out uncomfortably small, the fix sits upstream in price and delivery cost, which is what the pricing piece linked above is about. Buying more demand at a loss will not fix it. The objection also cuts the other way. A business that can afford less per client can afford much less waste, and a small budget spread across several channels tends to produce exactly that.
Run one channel until you know what a client from it costs
My rule is short. Always start with one channel, deliver the results, then take the money from those results and put it back in.
The pattern I run into looks different, and this is an illustration rather than a statistic. An owner splits a modest budget four ways, a quarter each to paid search, content, events and an outbound agency. Each channel gets too little to produce enough clients to measure. Months later the owner has four partial stories and no idea what a client from any of them cost. Nothing failed clearly, so nothing gets cut, and nothing succeeded clearly, so nothing gets more money.
One channel fixes that because it produces a number. The gate for adding a second channel is that the first one's CAC is known and paying. Known means you have counted it: what you spent on that channel, divided by the clients it actually closed, checked in your customer relationship management (CRM) system rather than estimated. Paying means those clients return more gross profit than they cost, at the ratio you chose above. Until both are true, a second channel only makes the first one harder to read.
The gate also tells you when to stop. In my agency's own test, selling to MSP owners, we spent 1,628 euros (EUR) on Google Demand Gen and Search ads from June to August 2026. That bought 1,086 clicks and 3 tracked conversions, about EUR 543 each, and no client we could trace. Search clicks cost EUR 17 to 37, and MSP owners barely search for what we sell and rarely click an unknown vendor's ad. We stopped it, along with our Microsoft and ChatGPT ads, in September 2026.
One limit on that result: it is one buyer, MSP owners, and I would not generalize it to an MSP selling to small businesses, where a trigger sends buyers to search and paid intent can work. And one thing in its favor. The test did its job. It was small, it was measured, the CAC came back unknown and plainly not paying, and we stopped before it cost more.
Then let the first channel pay for the next
The reinvestment step is where the method earns its keep. An MSP I work with started with one channel on a $3,500 retainer. The same MSP now runs a $20,000 retainer, because each channel's results paid for the next.
How it got there matters more than the endpoints. The budget grew because results had already arrived. That is a very different conversation with yourself than "we should be spending 10%." It is also a far easier one to have with a partner or, eventually, a buyer's team, because it comes with receipts.
It also changes what you ask of the next channel. The second channel does not need to be justified from scratch. It needs to beat, or at least match, the cost per client of the first one, and you already know what that is.
What to do this month
Take your last three new clients and do the arithmetic on paper.
- Estimate what each will pay you over its life: the monthly amount times the months you realistically expect to keep it.
- Multiply by your fully loaded gross margin, with the delivery labor counted, to get each client's gross-profit lifetime value.
- Pick the return you want on acquisition, on gross profit, and divide. That is your allowable cost to win the next client like these.
- Choose one channel and give it enough of that budget to close clients you can count.
- Write down the gate before you start: the cost per client you will accept, and the date you will judge it. Add nothing else until it is met.
If you want to work through these numbers alongside other MSP owners, Know Your Number is a private room for MSP owners with a regular call on sales and marketing.