How to set an ads budget that can actually work

Advertising is the one marketing channel where a budget can be objectively too small. Not less effective — unable to function. Below a certain spend the platform cannot gather enough signal to optimise, and you end up paying full price for untargeted delivery, which is the worst version of the product.

Paid ads appear in 548 of the 27,169 briefs we hold, and a recurring pattern in them is a budget that has been set by what feels affordable rather than by what the maths requires. That is understandable and it is also the most common reason a campaign is judged a failure when it never had a chance to be anything else.

Work backwards, not forwards

The wrong question is how much can I spend. The right one is what is a customer worth, and the chain from there is short:

1. What is a new customer worth to you — not the first sale, the whole relationship.
2. What share of enquiries become customers? If you do not know, find out before spending anything.
3. That gives what an enquiry is worth. It is usually higher than people expect, which is the useful part.
4. What share of ad clicks become enquiries? Five per cent is a reasonable starting assumption for a decent landing page.
5. That gives what a click is worth to you, and therefore what you can afford to pay for one.

Now compare that figure to what clicks actually cost in your market. If your affordable cost per click is comfortably above the going rate, advertising can work for you. If it is below, no budget makes it work and the answer is a different channel — or a change to the conversion rate or the customer value, which is a business question rather than an advertising one.

The learning threshold

Both major platforms optimise by observing conversions. Both need a meaningful number of them within a rolling window before that optimisation becomes reliable — a few dozen per campaign per fortnight is the commonly cited shape of it, and it is a floor rather than a target.

This has a hard consequence: your minimum viable budget is your cost per conversion multiplied by that threshold. If a conversion costs you fifty units and you need thirty a fortnight to leave the learning phase, your floor is three thousand a month. Spending a third of that does not produce a third of the result; it produces a campaign permanently in learning mode, which is the state where costs are highest and performance is least predictable.

The landing page is part of the budget

Paid traffic multiplies whatever your landing page already does. If the page converts poorly, advertising buys you a faster, more expensive demonstration of that fact. It is genuinely common for the highest-return action available to a campaign to be fixing the page rather than raising the bid.

The reason this gets skipped is that the page feels like a fixed cost already paid, while the ad spend feels like the variable being managed. That is backwards: the page's conversion rate is a multiplier on every unit of spend for as long as the campaign runs, so a one-off improvement compounds across the entire budget.

How long before you can judge it

Give a campaign enough time to accumulate the conversion volume the threshold requires — a fortnight at minimum, a month for anything with a considered purchase. Judging in week one is judging the learning phase, which is guaranteed to look bad and tells you nothing.

Equally, do not let a campaign run for six months on the grounds that it needs time. If the affordable cost per click calculation said the channel does not work for your economics, more time will confirm that rather than change it. The calculation comes first for exactly this reason.