The creative decides the cost. So test the creative.
Facebook and Instagram campaigns for demand you have to create — built around testing creative fast, because creative is what decides the cost.
On Meta, people aren’t searching for you. They’re scrolling, and your ad has to earn a stop. That single difference changes everything about how the account should be built.
Which means the biggest lever is not the audience. Meta’s targeting is good enough now that broad audiences often beat narrow ones. The lever is the creative — the hook in the first second, the offer, the format. Two ads to the same audience can differ by three or four times on cost per result.
So I build accounts that can test creative quickly and kill losers early, rather than accounts with forty ad sets and no clear read on any of them.
Everything in the programme.
Account and pixel audit
What's actually firing, what's double-counting, and whether the numbers in your account bear any relationship to your real enquiries. Usually the first thing that needs fixing.
Conversion tracking set up properly
Meta Pixel plus Conversions API where it's needed, with events that match real business outcomes — an enquiry, a booking, a call — not a page view someone decided to call a lead.
Campaign structure that can be read
Fewer campaigns, more budget per test. A structure with forty ad sets and RM30 each tells you nothing, because nothing ever reaches statistical significance.
Creative testing plan
Multiple hooks and formats tested systematically against one another, with a written rule for when something gets cut. Creative is the biggest cost lever on Meta and most accounts never test it deliberately.
Retargeting that isn't annoying
Separate messaging for people who already know you, frequency capped so you're not following the same person around for three weeks after they've decided no.
Reporting a human can read
Spend, results, cost per result, and the trend. Not reach and engagement. If a number doesn't help you decide something, it isn't in the report.
How the work actually runs.
Fix the measurement first
Before a single ringgit moves, I make sure the events firing in your account match real business outcomes. Optimising against a broken signal is worse than not advertising — you'll scale the wrong thing confidently.
Start narrow, spend deliberately
One campaign, a small number of tests, enough budget per test to learn something. Most small accounts fail because the budget is spread so thin that nothing ever produces a readable result.
Cut fast, scale slowly
Losing ads get killed on a written rule, not a feeling. Winners get more budget in steps, because moving too fast resets the learning phase and costs you the performance you just found.
Report what you'd act on
Every month: what was spent, what came back, cost per result, and what I'm changing next. Including the tests that failed, because those are half the information.
Results from this service.
No case study for this service yet.
Honest answer: I haven’t run this for a paying client yet. What I do have is the work on my own site — hypothesis, method and numbers, published in full.
Priced before we start, not after.
You get a written scope and a fixed number before any work begins, so there is no invoice at the end that surprises you. If the job turns out to be bigger than we thought, I tell you at the point I find out — not when the bill arrives.
Common questions about Meta Ads.
What's the minimum I should spend?
For a local business in Malaysia, around RM1,000–1,500 a month is the point below which testing stops being meaningful. Under that, there isn't enough data to tell a good ad from a lucky one, and you end up paying for guesses. I'll tell you honestly if your budget is too small to be worth managing rather than take the fee.
How long before it works?
You'll see data within days but not a verdict. Meta needs roughly a week per test to leave the learning phase and produce numbers worth reading. Give it a month before judging, two before scaling. Anyone promising results in the first week is describing luck.
Why does my current agency's report look good when my phone isn't ringing?
Usually because the report measures reach, impressions and engagement — numbers that always go up when you spend money, and that tell you nothing. Ask for cost per enquiry instead. If they can't produce it, the tracking isn't set up, which is the actual problem.
Do you write the ads and make the creative?
I write the copy and direct the creative — hooks, angles, formats, what to test against what. For video production or photography I'll tell you what's needed and work with whoever shoots it. I'm not going to pretend to be a videographer.
What can't you do yet?
Large-scale ecommerce catalogue and dynamic product ads, accounts spending five figures a month, and anything needing a full server-side tracking build. I work best on local service businesses and small ecommerce where the goal is enquiries or bookings. If your account is bigger than that, I'd be learning on your money and I'd rather say so.
Articles on Meta Ads.
First articles publishing this week.