There is no single right ad budget, and anyone who gives you one without asking about your business is guessing. The right number falls out of your own math: what a customer is worth to you, how often ads turn into customers, and how much you're willing to pay to win one. Get those straight and the budget picks itself.
Here is how to work it out, starting with the shortcut and then the version that actually holds up.
The quick rule of thumb
A common starting point is to put a percentage of your revenue toward marketing, and a slice of that toward paid ads specifically. Established small businesses often land somewhere around five to ten percent of revenue on marketing overall, with newer businesses chasing growth spending more, and paid ads taking a portion of that.
It's a fine sanity check, but it's backwards as a decision tool. It sets your budget from what you earned last year rather than from what an ad is actually worth to you now. Use it to catch a number that's wildly off, not to set the number.
The version that holds up: work back from a customer
The reliable way is to start from the value of a customer and work backwards to what you can afford to spend winning one.
- Start with what a customer is worth. Not just the first sale, but what they spend over the time they stay with you. A client on a monthly plan is worth far more than one project.
- Decide what you'll pay to acquire one. If a customer is worth a lot over time, you can afford to pay more up front to win them. A common comfortable target is spending no more than twenty to thirty percent of a customer's value to acquire them, though it depends on your margins.
- That's your target cost per acquisition. Multiply it by how many customers you want this month, and you have a budget grounded in your economics instead of a guess.
The beauty of this is that it scales with the value of what you sell. A business selling high-value services can afford a far bigger budget per customer than one selling a cheap product, and this math shows it plainly.
Budget enough to actually learn
Whatever the math suggests, there's a floor below which you learn nothing. Ads need a minimum volume of data before you can tell what's working from what's noise.
A good rule: spend enough to get at least one conversion per week, ideally a few. If your target cost per acquisition means you'd get one sale a month, you'll be staring at random results with no idea what to change. In that case, either narrow your targeting so each click is more likely to convert, or accept that you're running a small, slow test and give it longer.
Treat your first sixty to ninety days of ad spend as paid research, not as the campaign that has to pay for itself. You're buying data about your real cost per lead. That data is worth the spend even if the first leads aren't.
Put the money where intent is highest first
Not all ad spend is equal. Someone typing "emergency electrician near me" is ready to buy. Someone scrolling a feed is not. When your budget is small, start where the intent is highest, usually search ads for people already looking for what you sell, and add awareness and retargeting later once the high-intent channel is proven.
Spreading a small budget thin across five platforms is the fastest way to learn nothing on any of them. One channel, done properly, beats five done weakly.
Scale by results, not by feel
Once a campaign is reliably bringing in customers below your target cost, increase its budget gradually, around twenty to thirty percent at a time, and watch whether the cost per customer holds. Push too hard too fast and you often reach lower-quality audiences and your costs climb.
And be just as willing to cut. Any campaign that isn't paying after a fair test should lose its budget to the one that is. The businesses that win at ads aren't the ones that guessed the perfect number on day one. They're the ones that measured honestly and moved the money toward what worked.
What you can ignore
Ignore what your competitor spends, because you don't know their margins, their close rate, or whether their ads even make money. Ignore the pressure to be on every platform. And ignore any agency that reports on impressions and reach without tying a single number back to leads or sales. If the spend can't be traced to customers, it can't be judged, which is exactly why tracking has to be in place before the first dollar goes out.
The short version
Don't copy a flat number. Start from what a customer is worth, decide the share of that you'll pay to acquire one, and multiply by how many you want. Budget at least enough to get a conversion or two a week so you actually learn something, put it into the highest-intent channel first, and treat the first couple of months as paid research. Then scale what works and cut what doesn't.