TL;DR
Your starting Meta ads budget needs to cover two things: enough conversions per week for ad sets to exit the learning phase (Meta wants roughly 50 per week per ad set), and enough spend per creative to judge whether it's any good (roughly 1 to 3 times your target CPA per creative). Add those up and most serious businesses land somewhere between $1,500 and $5,000 a month to start. Below that, you're not advertising, you're donating to Meta's shareholders. The exact number depends on your margins, your average order value, and how many creatives you need to test.
"It depends" is not an answer
Ask any agency how much you should spend to start Meta ads and you'll get the same non-answer: "it depends." Technically true. Practically useless. It's the media-buying equivalent of a doctor saying "it depends" when you ask what's wrong with your knee.
So let's do what "it depends" refuses to do and actually work through the dependencies. There are really only four of them: your target cost per result, how many creatives you need to test, your audience size, and your tolerance for slow learning. Everything else is commentary.
The two things your budget has to buy
Forget everything else for a minute. A starting budget exists to purchase exactly two things: learning and creative testing. If your budget can't buy both, it's not a budget, it's a wish.
Learning. Meta's delivery system needs roughly 50 conversion events per week per ad set to exit the learning phase and stabilize. This isn't guru lore, it's Meta's own documented guidance. Do the math backwards: if your target cost per purchase is $40, one ad set needs about $2,000 a week ($40 x 50) to learn properly. Yes, really. Most accounts we audit are running five ad sets at $20 a day each and wondering why nothing ever stabilizes. Nothing stabilizes because nothing ever learns. You're running five permanent science experiments with no control group.
Creative testing. Every creative you test needs enough spend to produce a verdict. Our rule: budget 1 to 3 times your target CPA per creative before judging it. At a $40 target CPA, that's $40 to $120 per creative. Testing 10 creatives means $400 to $1,200 just for the creative test, before you've spent a dollar on scaling anything.
Add them up and the honest minimum for a real business with a $40 CPA looks like this: one or two ad sets at learning-level spend ($2,000 to $4,000/month) plus creative testing ($400 to $1,200/month). That's $2,500 to $5,000 a month. Not $150.
The $5/day fantasy
Somewhere along the way, "$5 a day to test" became folk wisdom. Let's be clear about what $5 a day buys you in 2026: about 150 to 300 impressions, depending on your CPM. That's not a test. That's a rounding error. You will learn nothing, conclude "Meta ads don't work for us," and go back to whatever you were doing before, poorer and more cynical.
Small budgets don't fail because Meta punishes small advertisers. They fail because the math doesn't work. At $5 a day and a $40 CPA, you'd need 8 days to get a single conversion. Meta needs 50 a week to learn. You are asking the system to optimize with roughly 2% of the data it needs, then blaming the system.
What the number actually depends on
Your margins and AOV. A brand selling $200 products at 70% margin can afford a $60 CPA and therefore needs a bigger testing budget in absolute dollars, but the economics work. A brand selling $15 products at 20% margin has a $3 CPA target and... honestly, Meta ads probably aren't your channel. The math doesn't care about your enthusiasm.
Your sales cycle. Short cycle (impulse buy, low AOV): you get signal fast, you can start leaner. Long cycle (high-ticket, lead gen): conversions trickle in, you need more budget to get the same number of weekly events, or you optimize for a higher-funnel event first.
Your audience size. Tiny niche audience of 50,000 people? You don't need a huge budget because you'll saturate quickly. Broad audience of 50 million? You need real spend to explore it. Budget and audience size have to match, or you're either burning money on reach you'll never use or starving an audience that could scale.
Your creative situation. If you already have 20 proven creatives from organic or another channel, your testing budget shrinks dramatically. If you're starting from zero creative, budget more. Creative is the variable most people forget to price in, and it's the one that matters most.
Honest minimums by scenario
| Scenario | Monthly starting budget | Why |
|---|---|---|
| DTC brand, $50-100 AOV, starting from scratch | $3,000 - $5,000 | Learning spend + 10-15 creatives to test |
| Lead gen (real estate, services), $100+ CPL target | $4,000 - $8,000 | Fewer, more expensive conversions; learning costs more |
| Small ecommerce, $25-50 AOV, some organic creative | $1,500 - $3,000 | Lower CPA target, existing creative reduces testing cost |
| Testing a new offer with proven account history | $1,000 - $2,000 | Account already has signal; you're testing, not building |
These are starting budgets, not forever budgets. The goal of month one is to buy information: which creatives work, which audiences respond, what your real CPA is. Month two is when you scale what month one taught you.
The question nobody asks (but should)
"What happens if I spend half the minimum?" Nothing dramatic. Your ad sets stay in learning limited forever, your creative tests never reach significance, and after two months you'll have spent $3,000 to learn what a proper $3,000 month would have taught you in three weeks. Underfunded testing isn't cheaper. It's slower and more expensive per insight.
The uncomfortable truth: Meta ads have a minimum buy-in, and it's higher than the gurus admit because "start with $5 a day" gets more course sales than "save up $3,000 first." We'd rather tell you the truth before you spend the money than sell you optimism and watch you churn.
FAQ
Can I start Meta ads with $500 a month? You can spend it, but don't expect to learn much. At that level you're buying a handful of conversions a month, which means no ad set ever exits learning and no creative test reaches significance. Save until you can fund a real test, or start with organic to build creative first.
How much of my budget should go to creative testing vs scaling? In month one, most of it goes to testing. A common split: 70% testing new creatives and audiences, 30% on whatever is working. As winners emerge, flip the ratio. If nothing is working after a full testing cycle, the problem is the offer or the creative, not the budget split.
Does a bigger budget always perform better? No. Bigger budgets learn faster and test more, but they also amplify bad creative and bad offers faster. Budget is an accelerant, not a fix. Fix the fundamentals first, then add fuel.
What is the 50 conversions per week rule? Meta's guidance for exiting the learning phase: roughly 50 optimization events per week per ad set. Below that, delivery stays in "learning limited" and performance is unstable. It's the single most ignored number in media buying.
Should I start with one campaign or several? Start with as few ad sets as possible, each funded well enough to learn. One campaign, two to three ad sets, each with enough daily budget to hit meaningful conversion volume. Consolidation beats fragmentation at the start.
When do I increase the budget? When you have a creative and audience combination hitting your target CPA consistently over 7+ days with stable delivery. Then scale gradually: 20-30% budget increases every few days, not 5x overnight. Sudden jumps reset learning and you'll be back at the start.
Written by the MetaMaxd team. We run Meta ads for brands that are tired of agency theater: no vanity metrics, no "hacks," just accounts engineered to make money. If you're not sure what your starting budget should buy, that's what our audits are for.
