How much budget you need for Meta ads to get your first leads
A starting budget is not "whatever you can spare," it comes from two numbers: how many conversions a campaign needs per week to get through learning, and what one lead costs you. A campaign needs roughly fifty conversions a week, so the weekly budget is fifty times your expected cost per lead.
The question "how much should I budget for ads" usually comes up before launch, when there is no cost-per-lead number yet. You can still calculate an answer instead of guessing one.
Where the minimum comes from
Meta's algorithm learns from conversions. While there are only a few, it does not yet know who to show the ad to, and it spends the budget trying things out. The benchmark Meta itself gives is about fifty optimization events per ad set per week. That is not a "for best results" suggestion, it is the threshold below which optimization performs poorly.
From there the formula is simple: weekly budget equals fifty conversions multiplied by the expected cost per conversion.
| Expected cost per lead | Weekly budget | Monthly budget |
|---|---|---|
| $5 | $250 | ≈ $1,000 |
| $20 | $1,000 | ≈ $4,000 |
| $50 | $2,500 | ≈ $10,000 |
| $100 | $5,000 | ≈ $20,000 |
The numbers in the table are arithmetic based on the learning threshold, not a promise of results. Your own cost per lead will be its own number, and you can only learn it from your own data.
What to do if you do not have that much
A common situation: you do the math and realize your budget is half of that. There are exactly three options, and none of them is "launch as-is and hope."
- 1.Optimize for a cheaper event. Not a purchase, a lead. Not a lead, the start of a conversation. Cheap events accumulate faster, so the campaign gets through learning.
- 2.Narrow everything except the audience: one offer, one ad set, one region. A budget spread across five ad sets never reaches the threshold in any of them.
- 3.Count in two-week blocks instead of one. The fifty-conversion threshold is a weekly figure, but a campaign moving slower still learns, just more slowly and with more variance.
Three mistakes that eat up a starting budget
Editing every day
Every significant change sends the campaign back to learning. Edit daily and the campaign never gets out of it, and the whole budget goes to trial and error.
Five ad sets on a small budget
It seems like five hypotheses will get tested faster than one. In practice the budget splits five ways, no ad set accumulates enough conversions, and by month end nothing is clear.
Turning off the campaign on day three
In the first few days, cost per lead is almost always higher than it will settle at: the campaign is still exploring. Judging by those days is like judging annual revenue by a single Monday.
How to tell whether the budget is actually working
Not by the cost per lead in the first few days, but by whether it stabilizes. Chart cost per lead by day over two weeks. If the spread narrows, the campaign is learning. If the number swings just as wildly as it did on day one, learning is not progressing, and the problem is either the budget or the event you are optimizing for.
How many conversions per week does a Meta campaign need?
The benchmark Meta itself gives is about fifty optimization events per ad set per week. Below that threshold, the algorithm does not have enough data, and optimization performs noticeably worse.
Can I run Meta ads with a $300 monthly budget?
Yes, if you optimize for a cheap event and keep it to a single ad set. For an expensive event like a purchase, that budget will not be enough to get through learning, and the money will go to exploring the audience instead.
Why is cost per lead higher in the first few days?
During the learning phase, the campaign explores audiences and formats, so it pays more for impressions. You cannot judge results by the first few days: watch whether the spread in cost per lead narrows by the end of the second week.