Most small Facebook ad accounts we open look the same: eleven campaigns, each with $5 a day, each targeting a different hunch, all paused and restarted so often the algorithm never learns anything. The money didn't fail. The structure did.
Meta's delivery system is a machine-learning engine that needs volume to optimise. Split a small budget eleven ways and none of the pieces gets enough conversions to exit the learning phase — so every piece performs like a coin toss. The counterintuitive rule of small budgets: fewer campaigns, fatter budgets, longer patience.
The account shape that works under $2,000 a month
- One campaign with your real objective. If you sell online, that's Sales (purchase-optimised). If you're a service business, Leads. Resist "traffic" and "engagement" objectives — they buy you clicks from people who click things, which is its own strange audience.
- One or two ad sets. Broad targeting (age, location, little else) plus, at most, one interest or lookalike stack to test against it. Meta's broad delivery has become genuinely good; over-targeting a small budget mostly pays extra for narrower reach.
- Three to five ads inside each ad set. This is where testing belongs — creative is the targeting now. Different hooks, not different fonts: a customer-problem opener, a straight product demo, a social-proof angle, a price/offer angle.
That's the whole account. It looks embarrassingly simple next to an agency's 40-campaign build, and it will usually beat it at this scale.
Budget maths you should do before spending
Work backwards from your conversion. Meta's system wants roughly 50 conversion events per ad set per week to optimise properly. Almost no small account achieves that on purchases — that's fine, but it sets expectations: you're always partially in learning, so judge results in weeks, not days.
The more useful arithmetic is worth-a-lead. If an average customer is worth $400 and you close one in five enquiries, a lead is worth $80. Under that ceiling, scale. Over it, fix the funnel before touching the ads — usually the landing page, which does more to change cost-per-lead than any audience tweak.
Creative: the rules that survive contact
- The first second is the ad. Feed scrolling is fast; open on the product, the result, or the problem — never a logo card.
- Make it native. Polished TVC-style ads generally underperform something that looks shot on a phone in your actual shop. This is a budget advantage, not a compromise.
- Say the price if you dare. Ads that state price pre-filter clicks; your cost-per-click rises and your cost-per-customer usually falls. On small budgets, paying only for informed clicks is the better trade.
- Refresh on fatigue, not on schedule. When frequency climbs past ~3 and results sag, swap creative. Until then, leave winners alone.
The testing rhythm
One change per week, maximum. The sequence that gets to a working account fastest:
- Weeks 1–2: launch the simple structure, broad targeting, 4 ads. Touch nothing.
- Week 3: kill the clearly worst ads (spend with no conversions), duplicate the best hook with a new opening line.
- Week 4: judge the account on cost per conversion across the whole month. Decide: scale (raise budget ~20%, no more), fix the landing page, or fix the offer.
Restarting campaigns, editing budgets daily, and duplicating everything each Monday all reset learning. The most profitable button in Ads Manager is the one you don't press.
What Meta ads can't fix
An offer nobody wants gets cheaper distribution, not resurrection. If four weeks of clean structure and honest creative produce nothing, the message — not the medium — is the suspect. Cheaper to find out in four weeks than four months.
FAQ
Boosted posts or Ads Manager? Ads Manager, always. Boosting optimises for engagement — reactions from people who react to things. Same money, wrong outcome.
Do I need the pixel / Conversions API? Yes, before your first dollar. Without conversion signal, Meta optimises blind. On Shopify, the official app sets up both in minutes.
Is $10 a day even worth it? At $300 a month with the one-campaign structure, yes — for lead generation with a decent customer value. For low-margin ecommerce, save until you can put ~$30–50 a day behind one campaign.
Want the account built like this, with creative to match? Send it to taskblitz — ad setup and the landing page it points at are both tasks we do.