How Much to Spend on Ads: Three Numbers, Not a Guess

Ask ten small business owners how much they should spend on ads and you get ten numbers, all of them guesses. Some copy a competitor. Some pick a round figure that feels safe. Some spend whatever is left after salaries. Almost none of them start from the only place the answer actually lives: what a customer is worth, and what one currently costs to win.

The budget is not a marketing decision. It is arithmetic on three numbers you already have, or can find in a week. Once you have them, the "right" budget stops being a debate and becomes a range you can defend.

The three numbers

What a customer is worth. Not the first invoice, the realistic total: the average booking value, times how many times a typical customer comes back, minus the cost of delivering it. A travel desk whose customers book twice a year for three years has a very different ceiling than one that sells single tickets to strangers. Be honest and conservative; an optimistic number here poisons everything downstream.

What a customer costs today. Last quarter's ad spend divided by the customers it produced. If you cannot compute this because nobody tracked which customers came from ads, that is the first fix, and it is worth more than any budget decision, because you are steering blind.

How many you can actually serve. The number the follow-up articles keep circling back to: if your team handles forty enquiries a month well and eighty badly, budget that produces eighty is money spent on leads you will lose.

The arithmetic, plainly

A customer worth ₹12,000 in margin over their lifetime, costing ₹3,000 to acquire, is a machine that turns ₹1 into ₹4. At that ratio the rational question is not "how little can we spend" but "how much can we spend before the ratio breaks or the team drowns." A customer worth ₹3,000 costing ₹3,000 is a machine that turns ₹1 into ₹1, and no budget fixes that; the fix is the cost side (the search terms report, the landing page, the follow-up speed) before another rupee goes in.

Most small accounts sit between those two, and the budget conversation should start by locating where. The ratio decides whether to scale, hold, or stop and repair.

Starting from zero, when there is no history

A new account has no cost-per-customer yet, so the first budget is explicitly a learning budget: enough to buy roughly thirty to fifty conversations over four to six weeks, because below that the numbers are noise. For most service businesses that is a few hundred rupees a day on one channel, not a grand launch across three. The goal of month one is a trustworthy cost per conversation, not profit, and calling that a loss misreads what was bought.

Setting the ceiling and the floor

Two guardrails turn the arithmetic into a standing policy. The ceiling: never let cost per customer exceed about a third of what the customer is worth, so one bad month does not wipe the margin. The floor: never drop below the spend that keeps campaigns out of perpetual learning, which for most small accounts means not slicing a modest budget across many campaigns, and not pausing and restarting every slow week. A steady modest budget outperforms a generous one switched on and off.

Revisit monthly, not daily

Put the three numbers on the same weekly sheet as everything else and re-run the arithmetic once a month. When the ratio improves because the account got cleaner, the budget can grow, in the measured steps that keep campaigns stable. When it worsens, the answer is diagnosis, not a budget cut or a budget hike. Budget follows the ratio; it never leads it.

FAQ

Is there a rule of thumb like "spend some percent of revenue"?

Those benchmarks exist and they are nearly useless for a single small business, because they ignore your margin, your repeat rate and your capacity. Two businesses with identical revenue can justify budgets that differ five-fold. Do the three-number arithmetic instead; it takes an hour.

My customer value is hard to estimate. What then?

Use the first booking's margin alone as a deliberately low floor. If the ads make sense even on that number, they make sense. Refine upward once you have a year of repeat data.

Should the budget change for festival season?

Yes, planned in advance rather than reacting mid-season, and only if the capacity number moves with it. Demand doubling while the team stays the same produces expensive unanswered leads; the budget should rise with the people who can answer.