How to Set Your Ad Budget: Goal, Margin and Test Cycle
Setting an ad budget by framework rather than a fixed figure: how goal, profit margin and test cycle length fit together, step by step.
rabbitclip teamPublished: 5 min read
Short answer
An ad budget is not something you pick as a figure first and judge afterwards; it is a decision that comes out of thinking through the goal, the margin that goal leaves, and how long a test needs to run to see a real result. 'How much should I spend a month' has no single right answer; the right question is how much value one new customer actually leaves in the business.
Most small businesses run this the wrong way round: pick a budget first, then look for a goal that fits it. It works better in reverse: the goal and the margin get clear first, and the budget follows from those.
This piece sets out the three things that build a budget decision, without quoting a figure.
An unclear goal leaves the budget arbitrary
The first step is to be clear on what action the ad is actually buying: a store visit, a form fill, or a direct sale; each carries a different cost and needs a different amount of patience before judging it.
For a car dealership, the goal is usually not a direct sale but a showroom appointment; without knowing how often an appointment turns into a sale, there is no way to tell whether the ad spend is actually profitable. Without a clear goal, every pound spent is really an experiment with an unclear question attached.
Margin sets the budget ceiling
How much a new customer is worth depends on the profit margin that customer leaves, not on revenue; margin is the figure to look at, not turnover. An aggressive ad budget on a low-margin product turns into a loss quickly; the same budget on a high-margin, repeat service such as a spa membership can stay profitable over a longer period.
The margin calculation should not stop at the first sale; how often a customer comes back over time matters too. For a flooring manufacturer, a one-off project sale and a recurring maintenance contract do not carry the same margin, and a budget decision cannot ignore that difference.
How long a test cycle should run
Judging a channel before enough data has been collected is the single most common mistake. The right test length varies by channel; a search campaign can show a signal in a few weeks, while a brand-awareness social campaign needs a longer cycle before its effect shows.
- The goal action (sale, appointment, form, call) needs to be clearly defined
- The test budget should run for at least one full decision cycle (e.g. the time between an appointment and a sale)
- Budget should not be cut or raised on early results before the cycle completes
- Do not test more than one variable (creative, audience, bid) in the same period
How to grow the budget over time
Once a test cycle is complete and cost sits where it should against the goal, budget growth should happen in steps; a channel performing well on a small budget does not mean it performs the same way at ten times the spend. Every increase starts a new learning period.
For a residential property management firm, monthly budget was reviewed every quarter, and every increase was kept incremental rather than a sudden jump. That protected the channel's efficiency while keeping the budget from running away.
How to split budget across channels
Once the total budget is set, the split across channels should follow which channel does which job at which stage; search ads mostly capture demand that already exists, social ads mostly try to create new demand. Comparing the two against the same expectation leads to the wrong conclusion.
For a law firm, most of the budget went to search ads because the customer was already looking for a solicitor; social ads kept a small share, used only for brand awareness. The split changes with the nature of the goal; there is no fixed percentage rule.
How seasonality affects the budget
In some businesses, demand is not steady across the year; the budget should not stay steady either. Raising the budget during high-demand periods and pulling it back during quieter ones tends to give a better result than a fixed monthly figure.
For an umrah and hajj tour operator, demand rises noticeably during certain months; when the budget was planned and raised for those months in advance, wasted spend during the quiet period dropped.
Running a fixed budget without noticing seasonality means missing an opportunity during high demand and wasting spend during low demand.
Who should approve the budget
Budget decisions should sit with one person; when more than one person signs off separately, a campaign tends to stop and start repeatedly, which disrupts the learning process.
For a flooring manufacturer, budget approval passed through three people; that three-way approval process led to the campaign pausing for two weeks in some months. Once approval moved to one person, the campaign ran without interruption.
Once approval sits with one person, that person needs clear knowledge of the goal and the margin; otherwise decisions get made fast but wrong.
An ad budget is not a guess; it is a calculation where goal, margin and test cycle meet. Rather than picking a figure first and searching for a goal to match, getting these three things clear first sets both the budget and the patience needed on the right footing from the start. If you would like to build this framework for your own business, a discovery call is a good place to start.
FAQ
Should ad budget be set based on turnover?
No, it should be set based on profit margin, not turnover; high revenue with low margin turns a budget into a loss quickly.
How long does a test budget take to show a result?
It depends on the channel; search ads can show a signal in a few weeks, while brand-awareness campaigns need a longer cycle.
When should the budget be increased?
Once the test cycle is complete and cost sits where it should against the goal, and even then, in steps.
Can more than one channel be tested at once?
Yes, but each channel needs its own cycle and its own budget; results should not be mixed together.
