The usual answer is a percentage of revenue, traced back to the SBA’s seven-to-eight percent for businesses under $5 million, rounded up to ten by everyone repeating it. Sounds authoritative right up until you turn it into line items. Ten percent of $250k is $25,000, and a business that size spending $25,000 is almost always buying things it has no business buying yet. So here are annual budgets in dollars, at roughly $250k, $500k, $1M, and past $2M. Including what’s deliberately not in them.
How much should a small business spend on marketing? Add up the jobs.
Answer first: pick the jobs your business needs done this year, price them, add them up. That’s your budget. Run the percentage afterward as a sanity check.
The rule works when revenue is smooth and there’s a marketing department to hand the money to. A local service business has neither. Your revenue arrives in clumps, and your “marketing department” is you, at 9pm, wondering why the contact form emails go to an inbox nobody checks.
And look, if that’s you, there’s a time question under the budget question. Hours spent wrestling with your own website are hours you weren’t quoting, weren’t billing, weren’t home. That cost never lands on an invoice, which is why it gets underestimated. Hiring it out buys back the Tuesdays.
There’s also a feedback loop baked into the rule: slow year means smaller budget means less visibility means slower year. And the costs that keep your site alive don’t shrink because you had a soft quarter. Closer to rent than discretionary spend.
What order should you buy things in?
Site, then Google Business Profile and reviews, then content, then ads. In that order. Skipping ahead makes budgets disappear with nothing to show.
The logic is boring. Ads multiply whatever is already there, so if your site turns ten visitors into zero calls, paid traffic turns a thousand into zero calls. Only now you paid for the thousand. Google Business Profile earns second slot because it’s free and it’s often the first thing a local customer sees.
Content sits third because it compounds slowly. A page published in March isn’t doing its best work in April. Ads sit last because they’re the one line you switch on Monday and off Friday. Great once everything underneath works. Terrible when it doesn’t.
What does the budget look like at ~$250k in revenue?
Roughly $4,700 to $5,200 a year, plus hosting. A real site, properly maintained, with a couple of new pages. No ad spend, no agency retainer.
| Line item | Annual cost |
|---|---|
| Website build, amortized (our Launch package is $2,400 flat, spread over three years) | ~$800 |
| Maintenance: Care plan at $300/month | $3,600 |
| One or two new service pages at $275-$375 each | $275-$750 |
| Domain and hosting | Your host’s renewal rate (look it up, the intro price expires) |
Google Business Profile, review requests, and answering the phone don’t get a row, because they cost $0 and some Tuesdays. Do all three. Just know what they are: pointers. Each ends with a stranger landing on your site, deciding fast. Free attention aimed at a site that can’t close is still zero.
The build is amortized because a good small business site lasts three to five years. Stuck on a platform you can’t export from? Add a one-time migration, ours run $700 to $1,800.
And if you already took a run at this yourself, maybe with an AI builder that spits out a page in ninety seconds, you probably hit the wall everybody hits. Fine in the preview, wrong on a phone. Copy describing a business, just not yours. A form submitting somewhere nobody checks. That’s not a you problem, and it isn’t one more prompt away. Those tools produce a starting point, and a starting point is not a business asset. The gap is structure that matches how people buy, copy that says what you charge, forms that land where you’ll see them, and somebody who owns it when it breaks. Closing that gap is what we actually sell.
Deliberately not in this budget: paid ads, an SEO retainer, social media management, a CRM you won’t configure, and a second website for a service line you launched in your head. Your cheapest growth is reviews and referrals, and the site’s job is to not lose the people they send you.
What changes at ~$500k?
Roughly $9,000 to $10,000 on the web side, before any growth retainer. The new money goes to content and to knowing where leads come from.
| Line item | Annual cost |
|---|---|
| Maintenance: Grow plan at $650/month | $7,800 |
| Content: four to six full-service pages at $275-$375 each | $1,100-$2,250 |
Two rows. Build amortization and hosting still apply, just no longer the interesting part. Feature work comes and goes (custom features start at $6,500), so leave it out of the annual figure.
Add one thing before you spend a dollar on ads: knowing where a lead came from. Analytics someone looks at, a hidden source field on every form, the habit of asking callers how they found you. Unglamorous. Also the difference between an ad budget and a donation. That source field is a care plan ticket, not a project.
This is where a growth retainer starts to make sense, because you finally have enough traffic and closed jobs for someone to optimize something. Before that, the same dollars do more as pages and site improvements, which keep earning after the invoicing stops. Our sister studio ZealousGrowth handles the growth half when you’re ready, and publishes prices the way we publish ours.
What changes at ~$1M?
You buy capacity instead of projects. Expect roughly $18,000 to $19,000 a year on web and content, then a separate acquisition budget on top.
| Line item | Annual cost |
|---|---|
| Grow plan at $650/month (monthly content updates, SEO monitoring, a named dev block) | $7,800 |
| One funded project: custom features from $6,500, online store from $8,500, membership from $9,500 | $6,500+ |
| Content: roughly a page a month at $275-$375 | $3,300-$4,500 |
A retainer beats one-off projects here for a boring reason: scheduling. You have ideas faster than a project queue absorbs them, and a page that ships in a week beats one you’re still scoping. Ours are month-to-month.
On ad spend: pick a number you could fund for two straight quarters without flinching, then start at a fraction of it. Ads you switch off in week three teach you nothing. Your measurement stack (analytics, call tracking, a CRM) sits outside it, at whatever those vendors charge.
What changes past $2M?
The constraint stops being the budget and starts being the queue. You have more ideas than a project schedule absorbs, and each one waits behind the last.
Two ways out, and they solve different problems.
A Dedicated developer is $5,000 a month and buys $10,000 of rate-card work every month. Projects stop being their own line item, because the retainer absorbs them. At $2M in revenue that is about three percent, and it comes without the salary, the benefits, or the months it takes to find and train someone.
Zealous Growth is the other path, and the better one if you are still climbing toward $2M rather than sitting past it. It carries less build capacity and far more automation and growth tooling.
| Line item | Annual cost |
|---|---|
| Dedicated retainer at $5,000/month, $10,000 of rate-card work each month | $60,000 |
| Content: two pages a month at $275-$375 | $6,600-$9,000 |
Roughly $67,000 to $69,000 a year, with acquisition spend separate, same as every tier above. No funded-project row, deliberately: below this you buy projects, here you buy capacity.
Maintenance and growth are two different line items
They get merged constantly, which is why people think their website costs too much. Maintenance keeps what you have working: updates, backups, uptime, security patches, small fixes. Growth goes and finds customers: content, SEO, ads, conversion work.
Maintenance is closer to insurance. You don’t evaluate it on leads generated, any more than you evaluate your locks on burglaries prevented. Growth is evaluated on leads. The maintenance half breaks down on the care plans page; the build half in the pricing calculator and the cost guide.
What if half your revenue lands in one quarter?
Split it into a floor and a push. The floor runs all twelve months regardless. The push is timed to land before your season starts.
Budget off percentage-of-revenue and you spend most when customers have already found you, least when the decision is being made. Backwards. If your busy stretch is April through July, publish your content in January. Your ads can run late. Your content cannot.
Whatever flexes in the off-season, it isn’t maintenance. December is exactly when you want a plugin conflict found quietly, by somebody watching for it, instead of on your best Saturday.
What do businesses overspend on at every stage?
Three things, in rough order of cost:
- Ads on a site that doesn’t convert. Traffic is the amplifier, not the signal. Fix the phone number in the header, the form that goes nowhere, and the page that never says what you charge. Then buy clicks.
- Rebuilds that should have been repairs. Plenty of tired-looking sites need four new pages and a speed pass, not a teardown. Rebuild when the platform is holding you hostage or the structure can’t hold what you sell now. Boredom is not a reason, and we’d rather sell you the four pages. The calculator prices both.
- Subscriptions nobody owns. The CRM, the scheduling tool, the email platform, the AI thing from a webinar. Each is small. Together, often a few hundred a month nobody budgeted. Once a year, open your card statement and cancel anything nobody has logged into since spring.
One column never gets cut, at any revenue: hosting, backups, security updates, and whoever is accountable for them. A tested backup turns a bad morning into twenty minutes. No backup turns it into rebuilding from screenshots mid-season. Cancel the scheduling tool. Keep the care plan.
How do you know at year end whether the budget worked?
Four numbers, none of them in your dashboard by default: leads by source, cost per lead by channel, revenue closed from web leads, and how many jobs you turned down because you were full. That last one is the good problem, and it changes what you buy next year.
Then a gut check. Can your site do something in December it couldn’t do in January? If not, you still got what maintenance buys, a site that stayed up and stayed patched. But the year ran on defense only, and our plans include time for fixes and improvements. If nobody spent yours, send the list and we’ll book it.
Common questions
Is the seven-to-ten percent rule ever useful?
As a ceiling, yes. If your itemized budget lands above a tenth of revenue and it’s not a deliberate land-grab year, something on the list is a want wearing the costume of a need. As a starting point, no, it hands you a number without telling you what to buy.
What if I can only afford one thing this year?
A site that clearly explains what you do, what it costs, and how to reach you, kept alive by a maintenance plan. One purchase, not two. An unmaintained site has an expiration date somebody else picks. That floor is cheap on purpose: Launch is $2,400 flat, smallest care plan $300 a month.
Do I have to pay for maintenance?
Somebody has to do it, every month, forever. Updates, off-site backups restored once to prove they work, security patches, uptime checks, and the small fixes that stack up. Real work, real consequences when it slips, and it never slips on a convenient week. A plan is how it stops being an emergency: ours run $300 for Care and $650 for Grow, month to month with no contract, and Dedicated from $5,000. At 7am on a Saturday, it’s our morning.
Can an AI builder replace this whole line item?
It replaces the blank page, which is not nothing. It doesn’t structure your services, write pricing you’ll defend on the phone, keep the thing patched in month nine, or pick up when it goes down in June. We use AI in our process too. Good assistant. Not a studio.
The part that matters more than the number
The biggest lines here buy something you own: the build, the pages, the design. Your domain, your code, your content, no platform holding the keys. Other rows are rent. Hosting, ad spend, and software buy access only while you keep paying. Fine, as long as you know which is which. A defensible budget is one where at year end you can point at what you bought outright.
So, concrete next step. Run the pricing calculator, it builds a real range in two minutes, nothing to schedule. Then send me that range plus what you’re spending now, and I’ll tell you which line I’d cut, which one I’d add, and which one is quietly costing you jobs. I usually have an opinion, and it’s worth more to you now than next January.