Most first-time Meta advertisers burn their budget on the wrong objective, the wrong audience size, and creative that never had a chance. Here is the exact structure we use for clients on their first $500 — every decision explained so you can defend it to your own boss.
1. Objective: purchases or leads only
Skip Traffic and Engagement objectives — they optimize for cheap actions, not customers. If you sell online, use the Purchases objective with a Conversions API server-side event. If you sell offline, use Leads with a native Instant Form.
2. Pixel + CAPI is not optional
iOS updates broke browser-side tracking. Without the Conversions API, Meta's optimizer sees maybe 60% of conversions — and blends worse audiences into your set. Set up CAPI in your Shopify/GTM/backend before spending a dollar.
3. Audience: broad + one interest layer
One broad ad set (age + country, no interests), one interest-based ad set built from 3–5 aligned interests. Let the algorithm learn. Do not stack 20 interests — you're just building noise.
4. Creative: 3 hooks × 3 formats
Test three hooks (problem, proof, offer) in three formats (static, UGC video, carousel). Nine ads total. Feed them all into the same ad set and let Advantage+ Creative pick winners.
- →First 3s = hook
- →Captions burned in
- →One clear CTA per ad
- →Native > polished
5. Budget & schedule
$20–30/day per ad set, minimum 5 days before you judge anything. Do not turn ads off before 3 days of data. Kill only the bottom 20% on cost-per-purchase — and only after the ad set exits learning.
6. Read the report the right way
Ignore CTR. Focus on cost per purchase / cost per lead and ROAS by ad. If ROAS is bad but AOV is trending up, you have a creative problem, not a targeting problem — swap hooks first, audiences last.
7. Scale carefully
When a winner clears 3× your target ROAS for 5 days, duplicate at 2× budget — do not edit the original. Aggressive edits reset learning and burn 3 days of history.