Most budget arguments in Meta Ads are really scope arguments wearing a number. Someone says “we have RM5,000 a month”, and two people at the table picture different things. One pictures RM5,000 landing in the auction. The other pictures RM5,000 covering the agency, the ads and the videos.
That confusion is expensive, and it happens before a single campaign goes live. An account funded at half of what the plan assumed will underperform for reasons that look like poor media buying and are actually arithmetic.
The second split, how the media money spreads across audiences, is where the popular advice goes wrong in the opposite direction. Fixed ratios get passed around as rules: seventy-twenty-ten, sixty-thirty-ten. They are reasonable descriptions of a mature account and poor instructions for a new one.
This guide from IZI Digital Marketing works through both splits in order, what the total has to cover, then how the media portion moves as the account matures. It ends with the five settings that decide whether your intended split is the one Meta actually spends.
Before the detail, a useful practical view on changing budget without disturbing what is already working.
When and How to Use the 20% Scaling Rule
Source video: How and When to Use the 20% Scaling Rule for Your Facebook Ads!
PART 1 · DIAGNOSE
The Three Budgets Hiding Inside One Number
IN BRIEFA Meta Ads budget contains three separate budgets, media spend, management fee and creative production. Splitting the funnel before splitting these three is the most common sequencing mistake, and it is closely tied to what a Facebook Ads package should cover in writing.
Meta bills only one of the three. The other two are yours to arrange, and whether they are funded properly decides how much the first one can achieve.
- Media spend. The money that reaches the auction. This is the only portion Meta charges you for, and the only one that buys reach directly.
- Management fee. What the agency or in-house buyer costs. Fixed monthly, a percentage of spend, or a hybrid, each behaves differently as you scale.
- Creative production. Shooting, editing, designing and writing the ads. Routinely assumed to sit inside one of the other two, and routinely funded by neither.
Creative is the one that quietly breaks. An account can carry a healthy media budget and a competent buyer, and still stall in month three because nothing new has been produced since launch. The auction keeps asking for fresh material, and the budget line that would supply it was never opened.
BENCHMARK BRIEFING 1 OF 4
How the Funnel Split Changes as an Account Matures
IN BRIEFThere is no single correct funnel split. A new account needs almost everything at the top because it has no warm audience to retarget; a proven account can afford a heavier middle. The tiering in our Facebook Ads packages follows the same logic.
| Audience layer | New account | Proven account | Scaling account |
|---|---|---|---|
| Cold prospecting | 85–90% | 65–75% | 70–80% |
| Warm engagement | 0–5% | 10–20% | 10–15% |
| Retargeting | 10–15% | 10–15% | 10% |
| What sets the ceiling | No audience to retarget yet | Retargeting pool size | Creative supply |
| Typical account age | Months 1–2 | Months 3–6 | Month 6 onward |
Illustrative model by IZI Digital Marketing, built on Meta’s documented learning phase requirements and audience-building mechanics. Indicative ranges for planning, not measured market averages.
Read the fourth row before the percentages. Each stage is capped by something real, an empty retargeting pool in month one, then the pool’s size, then how fast new ads arrive. The percentages are consequences of those constraints, which is why copying a mature account’s ratio into a new one produces a retargeting ad set with nobody in it.
Not sure which column your account is actually in?
Send us read-only access and we will tell you which stage the data supports, and what the split should be this month rather than in theory. See how a Blueprint session runs
PART 2 · DESIGN
The Floor Below Which Splitting Makes Results Worse
IN BRIEFEvery split divides your budget into smaller pools, and each pool has to collect enough conversions to learn. Below a certain total, the correct number of ad sets is one, a constraint we work through with every Facebook Ads services client before touching structure.
Meta states that an ad set stays in the learning phase until it gathers around 50 optimisation events since the last significant edit. That figure applies per ad set, not per account, and it is the arithmetic that governs how many ways you can safely split.
Work backwards from it. If your cost per enquiry is roughly RM40, one ad set needs about RM2,000 to clear learning at a sensible pace. Three ad sets need three times that. Divide a budget into pools that each collect a handful of conversions a week and every pool ends up guessing. The account then performs worse than it would have with a single undivided ad set.
DECISION BOX · WHO CONTROLS THE SPLIT
| Setting | Who decides the split | Main risk | Best fit |
|---|---|---|---|
| Campaign budget (Advantage+) | Meta, in real time | Money crowds into one ad set | Cold prospecting at volume |
| Ad set budgets | You, fixed per set | Spend stuck in weak sets | Protecting a retargeting layer |
| Campaign budget with minimums | Meta, inside your limits | Limits set too tight to work | Most Malaysian SME accounts |
Verdict: Choose the third row unless you have a specific reason not to. Meta’s Advantage+ campaign budget supports ad set minimums and maximums, which keeps a retargeting layer funded without freezing the rest of the campaign into a split you guessed at.
BENCHMARK BRIEFING 2 OF 4
Where Each Ringgit of a Monthly Meta Ads Budget Goes
IN BRIEFIn a healthy arrangement, media spend takes the largest share, management sits second and creative production is a real line rather than an afterthought. The same proportions shape how we scope a digital marketing package.
| Budget line | Relative share | What happens if it is underfunded |
|---|---|---|
| Media spend |
Largest |
Ad sets never leave learning |
| Management fee |
Second |
Account checked, not managed |
| Creative production |
Third |
Fatigue from month three |
| Landing page and tracking |
Fourth |
Clicks that never convert |
| Tools and reporting |
Smallest |
Decisions made on guesswork |
Illustrative model by IZI Digital Marketing, built on the working structure of a Malaysian Meta Ads retainer and Meta’s documented conversion tracking requirements. Relative weights only, not billed figures.
The third and fourth rows are the ones businesses cut first, because they are the only lines where cutting produces no immediate visible change. Both bite later, and both are cheaper to fund than to repair.
PART 3 · DESIGN
Testing Money and Scaling Money Are Not the Same Money
IN BRIEFRing-fence a testing portion, roughly 15% to 25% of media spend, and judge it on learning rather than on return. Mixing the two makes the account look inconsistent and pushes buyers to stop testing, which is exactly the wrong response. Which numbers to judge each on is covered in the Meta Ads metrics that actually matter.
Testing spend and scaling spend answer different questions. Scaling spend asks how much more of a known result you can buy. Testing spend asks what the next known result will be. Judged by the same yardstick, testing always loses, it is supposed to, because a test that only ever confirms what you already knew was not a test.
The practical move is to name the portion out loud. When 20% is declared as testing money at the start of the month, a losing test is a finished experiment rather than a mistake somebody has to defend in the report.
BENCHMARK BRIEFING 3 OF 4
How the Split Should Move Over Six Months
IN BRIEFA budget split is a schedule, not a setting. Testing share starts high and falls as winners emerge; retargeting share can only rise once there are enough people to retarget. A split that has not moved in six months is a split nobody has reviewed.
| Period | Testing share | Retargeting share | Decision due this month |
|---|---|---|---|
| Month 1 | 40% | 0% | Confirm tracking is correct |
| Month 2 | 35% | 5% | Cut the clearly weak angles |
| Month 3 | 25% | 10% | Name the winning angle |
| Month 4 | 20% | 12% | Raise budget in steps |
| Months 5–6 | 15–20% | 10–15% | Refill the creative pipeline |
Illustrative model by IZI Digital Marketing, built on Meta’s documented learning phase behaviour and the sequence in which retargeting audiences accumulate. Indicative pattern, not guaranteed timings.
The last column carries more weight than the percentages. Each month has one decision attached, and a split that changes without one of those decisions behind it is drift rather than management.
PART 4 · DEPLOY
Setting the Split So Meta Actually Follows It
IN BRIEFAn intended split and a delivered split are different things. Five settings decide whether the money lands where you planned, campaign structure, budget type, ad set limits, the increment you scale in, and the review date.
- Separate campaigns by job, not by audience. Prospecting and retargeting belong in different campaigns so a shared budget cannot quietly drain one into the other.
- Pick campaign or ad set budgets deliberately. Meta’s guidance on campaign budgets and ad set budgets sets out the trade-off, automatic distribution against fixed control.
- Set minimums on the ad sets you want protected. Usually the retargeting layer, which loses on raw efficiency early and still earns its place.
- Raise budgets in steps of around 20%, not in jumps. Large increases restart learning, and an ad set relearning is an ad set spending at its least efficient.
- Put a review date in the calendar. Monthly is enough. The split is a decision with a shelf life, and unreviewed splits fossilise.
Step four is the one that gets skipped when a campaign is performing well and the temptation is to double it overnight. The account then spends a week rebuilding delivery it had already paid for.
Want a second opinion on your budget maths before the next cycle?
We will work through the five settings against your account and tell you which ones are undoing your intended split, in the same line-by-line way we work through an SEO audit checklist.
BENCHMARK BRIEFING 4 OF 4
Which Budget Split Mistake Costs You What
IN BRIEFBudget mistakes differ in how quickly they show up. Over-splitting a small budget hurts within a fortnight; an unfunded creative line hurts in month three. Knowing which is which tells you what to fix first.
| Mistake | How it looks in the account | What it costs | When it shows |
|---|---|---|---|
| Too many ad sets, too little budget | Everything stuck in learning | Unstable cost per enquiry | Week 2 |
| Retargeting funded before there is an audience | Ad set barely delivers | Budget sitting idle | Month 1 |
| No testing share declared | Same three ads all quarter | Fatigue with no successor | Month 3 |
| Creative line not funded | Nothing new shipped | Rising cost per result | Month 3 |
| Scaling in large jumps | Learning phase restarts | A week of weak delivery | Immediately |
Illustrative model by IZI Digital Marketing, built on recurring patterns in Malaysian Meta Ads accounts and Meta’s documented learning phase and budget distribution behaviour. Indicative timings, not measured outcomes.
Fix from the top of the table down. The first two are structural and can be corrected this week; the middle two need a scope conversation; the last one only needs patience.
THE VERDICT
Split by Constraint, Not by Ratio
The advice you will find most often gives you a ratio and leaves you to apply it. That is backwards. A ratio is the output of a budget decision, and copying someone else’s output means inheriting constraints you do not have.
Work in order instead. Separate media, management and creative. Check whether the media portion can feed the number of ad sets you want. Set the funnel split according to what your retargeting pool can actually absorb. Ring-fence a testing share and say so out loud. Then review it monthly.
Scale matters less here than most people expect. Social media is used by 53.2% of Malaysian establishments that have a web presence, according to DOSM’s Usage of ICT and E-Commerce by Establishment survey, so a great many Malaysian advertisers are working the same auction with the same constraints. The ones who do well are rarely the ones spending most. They are the ones whose split can be explained.
FAQ
Frequently Asked Questions
1. How should I split my Meta Ads budget between cold and retargeting?
Start at roughly 85% cold and 15% retargeting, then let the retargeting pool decide the rest. It depends on how much traffic you have already generated, a new account has nobody to retarget, so an early retargeting ad set simply fails to deliver. Revisit the ratio monthly as the pool grows.
2. What is the minimum Meta Ads budget worth splitting at all?
If your total media spend cannot give each ad set enough weekly conversions to leave the learning phase, run one ad set. It depends on your cost per result: at RM40 a lead, one ad set wants around RM2,000 a month. Splitting below that makes every pool worse rather than sharing the work.
3. Should the management fee come out of my ad budget?
No, treat them as two separate lines from the start. It depends on how the agency quotes, and some quote an all-in figure, which is fine as long as the media portion is stated. Comparing an all-in number against a media-only number is the most common costing mistake in this market.
4. How much of a Meta Ads budget should go to testing?
Between 15% and 25% of media spend, declared before the month begins. It depends on how fast your creative fatigues; higher spend burns through angles faster and needs the upper end. The important part is naming it, so a losing test is not treated as a failure to explain.
5. How often should I change my Meta Ads budget split?
Review monthly, adjust when a decision justifies it. It depends on stage, early accounts shift more as testing gives way to proven angles, mature ones settle. Budget increases themselves should be gradual, since large jumps restart learning and cost you a week of efficient delivery.
Want your Meta Ads budget split checked against your actual numbers?
Send us your current spend, cost per enquiry and campaign structure. We will tell you how many ad sets that budget can genuinely support and where the split is working against you.