Budget capping sounds like the responsible choice. Set a ceiling, protect your spend, sleep at night. But most founders who ask about this have it backwards — the cap isn't protecting their budget. It's locking their campaign in learning mode.
Two things get called "budget capping"
Sorting out the terminology first saves a lot of frustration.
Type 1: A daily budget that's too low
If your daily budget is below a functional threshold, Meta can't gather data fast enough. The learning phase needs 50 conversions per ad set within 7 days. If your budget only generates 2–3 conversions a day, you won't hit that target. The campaign stays stuck on "Learning" or flips to "Learning Limited."
The minimum viable daily budget is your target CPA multiplied by 1–2. If your CPA goal is ฿500, you need at least ฿500–1,000 per day per ad set during the learning window.
Type 2: The cost cap bid strategy
This is the one that causes more grief. Cost cap tells Meta the maximum average CPA it's allowed to hit. If the market clears above that price on a given day, Meta doesn't bid. Full stop.
Set a ฿200 cost cap when the market's clearing at ฿350 that week, and your ads barely run. Impressions collapse. Conversions disappear. Learning phase never clears.
Why cost cap stalls learning
During learning phase, Meta's running experiments. It tests which users in your audience actually convert, what placements work, which creative drives the click. To do that, the algorithm needs to bid across a wide range of impressions — including some that feel expensive.
Cost cap cuts that exploration short. You're telling Meta "don't pay more than X" — but discovering that a slightly pricier user converts 3× better requires bidding above X. That signal disappears when you cap it.
Three signs this is happening to you:
- "Learning Limited" status for more than 10 days
- Daily spend is consistently under 50% of your budget
- Fewer than 10 conversions per week despite a sizable audience
The trap nobody mentions — cost cap inside CBO
Here's where it compounds. If you're running CBO and you've also set cost caps on individual ad sets, Meta can't allocate intelligently across them. Each ad set is constrained from bidding freely, so the algorithm doesn't know which one would win if it actually got real budget.
The visible result: a CBO campaign that consistently underspends because Meta can't figure out where to put the money.
Minimum budget by target CPA
| Target CPA | Minimum daily budget/ad set | Why |
|---|---|---|
| ฿200 | ฿300–500 | Needs 7+ conversions/day to reach 50 in 7 days |
| ฿500 | ฿750–1,000 | Higher cost per conversion = slower data accumulation |
| ฿1,500+ | ฿2,000–3,000 | Run Lowest Cost first; don't add cost cap until you have real data |
If your daily budget is under 2× your target CPA, don't set a cost cap yet. You're just slowing the data collection that would eventually let you set a sensible cap.
When cost cap actually belongs in your account
Cost cap is a scaling tool, not a testing tool. Use it after:
- The campaign has cleared learning phase (status shows "Active," not "Learning")
- You have 50–100 real conversions at Lowest Cost to anchor your CPA baseline
- You're increasing budget and want to stop CPA from drifting as you scale
Start cost cap at 1.2–1.5× your observed CPA from Lowest Cost — not your target CPA, your actual one. Setting it tighter than that is the fastest way to break a campaign that was working.
Quick reference
| Situation | What to use |
|---|---|
| New campaign, no data | Lowest Cost (no cap) |
| Learning Limited >10 days | Increase budget first, remove or loosen cost cap |
| 50+ conversions, scaling up | Test cost cap at 1.2–1.5× your observed CPA |
| CBO + want to control CPA | Use minimum spend floors per ad set, not cost cap |
What to do next
If your campaign is stuck in Learning Limited, start with one calculation: daily budget ÷ target CPA. If that number is under 1, the budget is too low — everything else is secondary. Generate a blueprint in AdBlueprint and check the Budget strategy section. It calculates your minimum viable budget automatically and tells you whether your setup can realistically exit learning phase, without the guesswork.
- [insert link to "CBO vs ABO: which Meta budget strategy actually wins?"]
- [insert link to "Meta Ads learning phase: what it means and when to intervene"]