"How much should we put into Facebook and Instagram ads?" comes up in almost every strategy call, usually right after a business owner has tried $20/day boosted posts and seen nothing to show for it. The honest answer isn't a flat number — it's a formula. But unlike search advertising, Meta's ad platform has a hard technical minimum built into how its algorithm learns, and most small budgets fail before they ever get a fair test.
What Facebook and Instagram ads actually cost in 2026
Meta pricing is quoted as CPM (cost per 1,000 impressions), which trips up business owners used to thinking in cost-per-click. Across US campaigns, average CPM sits in the $8–$14 range, but the spread by industry and objective is what actually drives your budget:
- Local services (home services, salons, fitness studios): $6–$12 CPM, often with strong performance from geo-targeted radius campaigns
- Real estate: $10–$20 CPM for listing and lead-gen campaigns, higher in competitive metro markets
- Automotive (dealers, service centers): $8–$16 CPM, with video and carousel formats typically outperforming static images
- Travel and hospitality: $9–$18 CPM, with strong seasonality — CPMs climb sharply in the 8–10 weeks before peak booking windows
- Professional services (legal, financial, B2B consulting): $15–$30+ CPM — Meta's audience is colder than search intent, so cost per lead runs higher even when CPM looks moderate
The number that actually matters for budgeting is cost per lead, not CPM. A $10 CPM with a 1% click-through rate and 3% landing page conversion works out to roughly $33 per lead before you've spent a dollar on creative testing — and that math changes fast depending on how tightly your audience and offer are matched.
The minimum viable budget: what Meta's algorithm needs to leave the learning phase
Unlike Google Ads, Meta publishes an explicit threshold: each ad set needs roughly 50 optimization events (typically leads or purchases) within a 7-day window to exit the learning phase. Below that, delivery is unstable, costs swing widely, and performance data is close to meaningless.
Work it backward: if your cost per lead runs around $30, hitting 50 leads in a week means a budget of roughly $1,500/week, or about $6,000/month, per ad set. That's higher than most small businesses expect to hear — and it's exactly why running six ad sets at $10/day each almost always underperforms one or two ad sets funded properly. Consolidation, not diversification, is what gets a Meta account out of its worst-performing phase.
An account split across too many ad sets isn't testing more — it's just making sure none of them ever leaves the learning phase.
A working-backward formula
The same logic that governs search budgets applies here, adjusted for Meta's colder audience and higher creative dependency:
- Customer value: what is a new customer worth — first purchase, or lifetime value if the relationship repeats?
- Allowable cost per acquisition: how much of that value can go toward winning the customer and still hit margin?
- Lead-to-customer rate: if 20% of leads close, your allowable cost per lead is a fifth of your allowable CPA.
- Expected conversion rate: divide your allowable cost per lead by your landing page's expected conversion rate to get a maximum affordable cost per click, then translate that against your industry's CPM and click-through benchmarks above.
If the resulting budget is below what it takes to hit 50 weekly conversions per ad set, the fix usually isn't a bigger overall spend — it's narrowing to a single offer, a single audience, and a single ad set until the volume math actually clears the learning-phase threshold.
How budgets should change over time
Weeks 1–3: creative and audience testing
Expect elevated cost per lead here, on purpose. You're finding which hook, format, and audience actually produces qualified leads before asking the algorithm to optimize toward volume.
Weeks 4–8: consolidation
Kill underperforming creative and audiences, and reallocate spend into what's working. This is also where conversion tracking quality matters most — if your pixel and conversions API aren't both firing cleanly, Meta is optimizing against incomplete data, and no amount of budget will fix a signal problem. Landing page quality compounds here too, in much the same way we cover in how much conversion rate optimization costs.
Week 8+: scaling
Increase budget in increments of 20–30% at a time, not overnight jumps — Meta's algorithm treats large sudden budget changes as a new learning event, resetting the stability you just built.
Signs your budget is wrong (in either direction)
- Individual ad sets are stuck "learning" for weeks with no exit — spend is too thin, spread too wide, or both
- Cost per lead is inconsistent week to week with no clear trend — the algorithm doesn't have enough signal to stabilize
- Frequency (average times a user sees your ad) climbs past 3–4 within a campaign flight — your audience is too small for the budget, not too large
- You can't name your cost per lead by campaign — a Meta budget without event-level tracking is a donation, not a channel
So what's the actual number?
For most US small businesses running one focused campaign, a realistic starting range is $2,000–$6,000/month — enough for one or two ad sets to clear the learning-phase threshold within the first month. Real estate and professional services, where cost per lead runs higher, often need the upper half of that range or more before Meta's algorithm has anything real to optimize against.
The number that's actually right for you comes out of your own customer value and close rate, not an industry average. Run the formula above against your numbers, or we can build the media plan with you — audience, creative, and budget sized to your real unit economics — before a dollar goes into the account.