Running ads that sell
Do this first
You want Meta ads you can test properly and judge on real sales.
Fix tracking before you spend
- Use the Conversions API and track meaningful events. For COD, send confirmed or delivered orders. Read the lesson
Work out breakeven before launch
- Count returns and work out breakeven ROAS. Judge delivered sales or qualified leads. Read the lesson
Check Meta settings before launch
- Check placements, limited spending, Related media and Enhancements. Recheck after publishing or duplicating. Read the lesson
Keep structure simple
- Use 3 to 5 ads per ad set, one angle each. Exclude existing buyers. Read the lesson
Test different reasons to buy
- Test distinct reasons to buy. Expect roughly one winner per 11 to 27 concepts. Read the lesson
Build every ad as hook, body and CTA
- Film hooks, bodies and CTAs together, then recombine them. Read the lesson
Make the CTA say five things
- State the action, benefit, timing, effort and next step. For COD, explain delivery and payment. Read the lesson
Judge and scale with rules
- Scale winners about 20% at a time. Judge spend against your target cost per sale. Read the lesson
Questions you might have
Which placements should I turn off?
- Many reels cut Audience Network, Threads and Messenger without quality evidence. Check your real results. Full answer
Broad or interest targeting?
- Test both for 7 days using delivered orders. It's unresolved. Full answer
How many creatives should I test?
- Expect 11 to 27 concepts per winner. Avoid near-identical variants. Full answer
Why do results drop when I double budget?
- Extra spend reaches less interested buyers. Increase about 20% and add angles. Full answer
How do I set up a Reply to keywords ad?
- Open Conversations at ad level, enable Reply to keywords and add your message. Availability varies. Full answer
How do I protect my page and pixel from bans?
- Own the assets. Share through partner access and keep backup admins with 2FA. Full answer
Watch these reels
@akshitraja: Meta's Andromeda retrieval engine shortlists a few thousand ads per
- How Meta shortlists ads. Watch
@sevenfiguresmp: A study of 578,000+ Meta ad creatives and $1.29B in spend found only
- Most creatives lose. Watch
@evanseech: A checklist of Meta ad settings the creator turns off by default
- Meta settings to check. Watch
@nicoaguirre.td: The Meta pixel works but reports incomplete data because of iOS
- Fill gaps in pixel data. Watch
@dakota.hermes: The angle matrix
- Find different reasons to buy. Watch
@hormozi: 50-5-3 ad method
- Hook, body, CTA. Watch
Avoid these
Fake news-headline or fake text-thread ads
- Fake formats risk misleading-content violations. Why
Health claims like '15 kg in 30 days'
- Copy formats, not health claims. Claims face regulation and rejection. Why
Turning campaigns on and off nightly
- Use scheduling with a lifetime budget. Why
Hundreds of ads a week on a small budget
- Test fewer, distinct concepts. Small accounts have lower winner rates. Why
Trusting big results shown on demo data
- The 800x claim used demo data. Why
The reels
Get the next lesson in your inbox
When I find a reel worth acting on, you get the short version: what it says and what to do. Free, rare, one click to leave.
Rather have it done for you? Work with Tanish