The most expensive mistake in small-business advertising isn't overspending — it's spending an amount that can't possibly work, learning nothing, and concluding "ads don't work for us." Budgets fail at both extremes: too small to generate evidence, or too spread out to dominate anything. Here's a first-principles way to set a number you can defend.

Start from the customer, not the wallet

Before any budget talk, you need two numbers only you can supply: what a new customer is worth to you (including repeat business — a client who returns for years is worth far more than one invoice), and what percentage of enquiries you close. From those, work backwards: if a customer is worth a certain amount and you close, say, one enquiry in three, then an enquiry is worth roughly a third of a customer. Your entire paid strategy is buying enquiries for meaningfully less than they're worth. Our free ROI calculator runs this arithmetic for your numbers in two minutes.

The evidence threshold

An ad budget's first job is to buy information: which keywords, which ads, which pages produce enquiries. Information requires volume — a few hundred clicks is a reasonable planning floor before patterns are trustworthy. So price a click in your market (Google's free Keyword Planner shows bid ranges for your exact keywords and location), multiply by a few hundred, and that's roughly the minimum monthly commitment for a real test. In a cheap-click industry that's a modest number; in legal or insurance it's substantial. If the arithmetic says you can't afford the evidence threshold in your vertical, don't run diluted ads — put the money into organic search and your Google Business Profile instead, where small budgets compound rather than evaporate.

Concentrate, don't spread

The universal small-business budget error is spreading: a little Google, a little Meta, a little boosting posts, none of it decisive. Channels have minimum effective doses. One campaign, one tightly defined service, one city, one landing page — funded properly — beats five underfunded fronts every time. You expand after something works, funded by the thing that works. The 70/20/10 shape is a useful discipline once you're established: 70% to proven channels, 20% to promising tests, 10% to wild bets. On day one, it's 100% to one honest test.

Geography is a budget decision

A radius is a spending instruction. Target a whole region when you serve one city and you're buying clicks from people who will never become customers. Draw your campaign geography around where your customers genuinely come from — tighter is almost always more profitable at small budgets, because the same money buys deeper presence in front of fewer, righter people.

Kill criteria, decided in advance

Write down, before spending, what result keeps the budget alive: usually a maximum acceptable cost per enquiry, held over a defined period. Then honour it. Campaigns limp along for months on "it's building awareness" — awareness is what advertisers say when they can't show enquiries. Paid advertising is the most measurable channel that exists; refuse to let it be judged on vibes. This is also the standard to hold any agency to, including us: our ads management reports against cost-per-lead you can verify in your own account, and a campaign that can't justify itself monthly should be turned off, not defended.

The budget conversation with an agency

If you hire help, the budget conversation reveals the agency before any work does. Healthy signs: they ask what a customer is worth to you before quoting anything, they separate ad spend from management fee in writing, and they're willing to name a budget below which they'd advise against ads in your vertical. Warning signs: a management fee priced as a percentage of spend with no performance reference (it rewards them for your budget growing, not your results), reluctance to give you owner access to your own ad account, and "trust the process" in place of a defined cost-per-enquiry target. You should own the account, see every number, and be able to leave without losing your history — our own management is structured that way precisely because the industry's defaults aren't, and because a client who stays out of choice is worth ten held by contract.

One more variable worth planning deliberately: cash-flow timing. Ads charge you now for customers who may pay you in weeks (service businesses with quotes) or months (long sales cycles) — so a budget that's affordable on paper can still strain a small business's month. If your sales cycle is long, size the test budget so you can hold it steady for the full cycle plus a margin; stopping a campaign right before its first deals close is a common and entirely avoidable way to record a false negative. Seasonal businesses should invert the instinct to spread spend evenly: bank the budget and concentrate it into the weeks your customers actually decide.

The honest summary

Work out what an enquiry is worth to you, price the clicks in your actual market, fund one concentrated test past the evidence threshold, and pre-commit to the number that means stop. That's the whole discipline. Everything else — bidding tricks, scripts, dayparting — is optimisation on top of a budget that either can produce evidence or can't.