How much budget a campaign needs before it learns
A campaign budget is not decided by what is left over, it is worked backwards: how many conversions each ad set needs before the platform can optimise, what you expect each one to cost, and over how many weeks. That gives the floor the test sits on and the daily figure you actually type in. If your budget lands under it, the part that matters is further down.
What the arithmetic rests on
The threshold you are aiming at. Meta publishes 50 per ad set in a seven day window as its guidance for leaving the learning phase. If your network, your account or your event asks for something else, change it.
Your estimate, not a fact. It is the weakest number in the whole sum and the one that moves it most.
Each ad set learns on its own. Splitting a budget across more boxes does not multiply it, it divides it.
How long it runs before you decide anything on what you see.
The number already in your head, or the one the client said. Below it is held against the floor in conversions per week, which is the only comparison that means anything.
What the test costs
Per ad set per week
400 €
Your cost per conversion times the threshold you set. It is what one ad set has to spend to reach that many conversions inside a week.
Per day, whole campaign
171,43 €
The figure you type into the platform: one ad set week times the number of ad sets, divided by seven. Over thirty days that is 5142,86 €.
The whole test
4800 €
One ad set week times the ad sets times the weeks. It is what the trial costs, not a monthly budget.
What your budget buys
19.4 conversions
Conversions per ad set per week that the figure you had in mind funds, spread over thirty days. The threshold you set is 50.
Every line here hangs off the cost per conversion, and that number is your estimate until the campaign hands you its own. Run the sum again with the real figure at the end of the first week, because that is the only one worth anything. And be clear about what this is: what the test costs, never what it returns.
Your budget against the floor
You land under it, and no bid fixes that. Either cut the number of boxes that have to learn, or lower the cost of the conversion you measure, or give it more time knowing what that does and does not buy.
Fewer ad sets
1 ad set
How many your budget funds up to the threshold, at the same cost per conversion. Concentrating the spend on one is usually worth more than spreading it over five where none of them learns.
A cheaper conversion
3,11 €
The cost per conversion your budget does reach. You get there by measuring something that happens earlier in the funnel, a sign up instead of a purchase, in exchange for that conversion being worth less.
More time
3 weeks
How long each ad set takes to accumulate the threshold at your rate. Watch this one: the threshold is counted per week, so a longer test does not pull the campaign out of the learning phase, it only builds enough volume to read something at the end.
- The cost per conversion you type in. It is an estimate and everything else hangs off it, so the sum is worth exactly what that guess is worth until the campaign corrects it.
- The learning threshold, which is platform guidance and not a law. Every network counts its conversions its own way, and editing the ad set, the bid or the creative sends the campaign back to the start.
- The split. Spend is divided equally across the ad sets here and delivery is assumed to be steady, when in practice one ad set eats the others and the platform does not spend the same amount every day.
- The outcome. This page says what the trial costs, not what you will get out of it, and nothing on it predicts sales or return.
Frequently asked questions
What is the minimum budget for a campaign?
There is no universal figure, there is a sum, and it comes out of three things that are yours: how many conversions each ad set needs before the platform can optimise, what you expect each one to cost, and how many ad sets you will have running at once. Meta publishes 50 conversions per ad set in a seven day window as its guidance for leaving the learning phase, and that is the only threshold written down here. On another network, or with another conversion event, the number is different and you have to supply it.
Can I start with a small budget?
You can spend it, which is not the same as learning from it. On a budget that produces few conversions a week, the ad set stays in the learning phase, delivery keeps exploring instead of sharpening, and the differences you see between one ad and another fit inside the margin of error. The spend is real from day one; the reading is not real until there are enough cases behind it.
What if I cannot reach the floor?
There are three ways out and none of them is raising the bid. The first is cutting ad sets and concentrating the same money where it does reach the threshold. The second is measuring a conversion further up the funnel, a sign up or a form instead of a purchase, which costs less and therefore accumulates sooner, accepting that the conversion is worth less. The third is running the test for longer, with one warning: the threshold is counted per week, so more weeks will not pull the campaign out of the learning phase, they only give you enough accumulated volume to read something at the end.
Can I launch the campaign from GoFeed?
No. The ads dashboard reads Meta, Google, LinkedIn, TikTok and Pinterest into the same table, with their spend, impressions, clicks and conversions, so the client report comes out of one place. Campaigns are created and edited on each platform, never here, and Google is read only. Every profile gets it, whether you run clients, your own brands or your company marketing, from the second tier of its plan upwards.
Fourteen days to decide
Connect one account, schedule a week and see whether it fits. If it does not fit, cancel before day 15 and nothing is charged.