Is your store leaking revenue?
Find out exactly where you're losing sales — takes 2 minutes.
The spend-based formula for weekly ad volume, the new-versus-proven split, and why quality still beats the number
How Many New Ad Creatives Should You Launch Per Week?
There's a formula for this. Most stores guess.
Launch one new ad creative per week for every $10,000 a month in ad spend, rounded up, with a hard floor of 2 per week. A store spending $30,000/month launches 3 new creatives weekly; a store at $8,000/month still launches 2. This is the exact output of the Creative Volume Calculator, a public Meta-ads planning tool built on spend-tied creative cadence.
A store spending $100,000 a month launches 10 new ads a week. A store spending $10,000 a month launches 1 — but the floor kicks in, so it's really 2. Always round up.
That's the easy half. Here's the part almost nobody teaches: not every one of those ads should be a brand-new concept.
weekly = max(2, ceil(spend / 10000))
One ad per week per $10,000 in monthly spend. Round up. Floor of 2. Simple enough to do in your head before your next budget meeting.
But the number of new concepts inside that weekly count changes with spend. And getting that mix wrong is why most stores either run out of fresh ideas or burn budget testing things they already know work.
What Percentage of Weekly Ads Should Be New Concepts vs Proven Winners?
At low spend, almost everything should be new. At high spend, almost none of it should be.
The share of weekly ads that should be brand-new concepts starts near 100% at low ad spend and decays toward a 20% floor once monthly spend reaches $200,000. The rest of the weekly volume goes to iterating proven, already-winning concepts. A store at $20,000/month runs close to 80% new; a store at $200,000/month runs mostly iteration.
The logic: at low spend you haven't found your winning concepts yet, so you're still searching. At high spend you have several proven concepts carrying real revenue, and the highest-ROI activity shifts from finding new ideas to squeezing more life out of the ones that already work.
The formula behind the split:
newPct = max(0.20, 1 − 0.80 × (min(1, spend/200000))^0.6)
newAds = round(weekly × newPct)
proven = weekly − newAds
You don't need to run this by hand. Here's what it produces at common Malaysian and Singaporean DTC spend levels, converted to MYR at roughly RM4.7/USD:
| Monthly spend (USD) | Monthly spend (MYR, approx.) | Ads/week | New concepts | Proven/iterated |
|---|---|---|---|---|
| $10,000 | RM47,000 | 2 (floor) | 2 | 0 |
| $20,000 | RM94,000 | 2 | 2 | 0 |
| $50,000 | RM235,000 | 5 | 3 | 2 |
| $100,000 | RM470,000 | 10 | 5 | 5 |
| $150,000 | RM705,000 | 15 | 5 | 10 |
| $200,000+ | RM940,000+ | 20 | 4 | 16 |
Source: the Creative Volume Calculator's published formula, reproduced with the tool's exact output.
Two things in that table look like errors and aren't.
At $10,000 and $20,000 a month, the 2-ad floor plus integer rounding pushes both ads into the "new" column, so the stated proven percentage produces zero actual proven ads that week. That's the floor overriding the ratio, not a bug.
At $200,000+, the new-ad count drops to 4 while it was 5 at $150,000. That's the model snapping to whole ads as spend crosses the decay curve — you can't launch 4.2 ads.
For most Shopify stores in Malaysia and Singapore, this means sitting somewhere in the $10,000 to $50,000/month band, where the reality is: 2 to 5 new ads a week, and for a while, all of them are new concepts because you haven't found your proven winners yet. If your team has no in-house editor, that 2-ad floor alone is a full-time job most weeks — plan headcount or a creative partner around it, not around the bigger numbers further down the table.
Why Does the New-Concept Share Decay as Spend Grows?
Because at scale, you already know what works. Making more of it beats guessing again.
The new-concept share decays because low-spend accounts are still searching for a winning formula, while high-spend accounts already have proven concepts carrying most of their revenue. Iterating a known winner is lower-risk and higher-efficiency than launching a brand-new idea, which is why the model shifts weekly volume toward iteration as spend — and the number of confirmed winners — increases.
An iteration is a meaningful variation on something that's already proven: a new hook on a winning script, a different opening 3 seconds, a new creator delivering the same core angle. It's not a headline swap — the source framework is explicit that post-2025 algorithm changes mean a "meaningful" adjustment matters, not a cosmetic one.
The payoff of iteration: you extend a winning concept's life, reach new pockets of the same audience, and reduce creative fatigue — without the risk of a totally new idea. A known concept, iterated, is the lowest-variance, highest-ROI creative work you can do.
This is also why brands with big budgets don't look like they're constantly reinventing themselves. WebMedic has audited Meta ad accounts where a store believed it needed more creative variety, when the real gap was that it hadn't identified which of its existing concepts were actually winning yet.
Does your ad account look like a content mill instead of a testing system? Find out where your creative process is leaking budget — take the free Revenue Score. 3 minutes. Free. No pitch.
How Many Concepts Do Big Ad Accounts Actually Run?
Fewer than you'd think. The ad count is high; the concept count is not.
Large DTC brands running Meta ads at scale operate roughly 10,000 active ads at any time, built from only 10 to 50 total creative concepts, heavily iterated. The concept is the repeatable structure — avatar plus format plus storytelling pattern. The ads are the variations built on top of it.
This is the distinction the weekly formula depends on. A concept is something you could write an SOP for: a specific avatar talking through a specific format about a specific angle. An iteration is a controlled variation inside that concept — same story, different lens.
Confuse the two, and "launch more creative" turns into a content mill: dozens of near-identical ads that all test the same idea instead of testing new ones. The formula's new-vs-proven split exists specifically to prevent that — it forces you to ask whether a given week's ad is genuinely a new concept or an iteration of one you already have.
Does Quantity Matter More Than Quality in Ad Creative Testing?
No. The formula sets a ceiling, not a target you must hit regardless of quality.
Ad creative quality matters more than volume on Meta's current algorithm, which rewards ads that perform, not ads that exist. If a brand can only produce 3 genuinely strong ad concepts in a week at $90,000/month spend, launching those 3 outperforms padding the week to 9 with mediocre filler. Roughly 1% of ads typically drive the majority of account revenue.
This is the override that makes the formula useful instead of dangerous. Meta's current ad-matching approach (commonly referred to by practitioners as "Andromeda") doesn't reward ad count — it rewards ads it can confidently match to an audience because the creative signals clearly who it's for. A mediocre ad gets almost no exposure regardless of how many you upload next to it. Practitioners in the space put it bluntly: "decent gets no exposure."
So treat the weekly number from the formula as your ceiling for the week, not your quota. If your team can only produce fewer high-quality ads, produce fewer high-quality ads. The formula tells you how much room you have to fill profitably — it doesn't obligate you to fill all of it with weak ideas just to hit the count.
How Should You Group New Ads When You Launch Them?
By concept, not by budget convenience. And never mix them into an existing ad set.
Group new ad creatives into ad sets by concept — genuinely different concepts go into separate ad sets, while closely related variations of the same concept can share one. The practical sweet spot is 4 to 8 ads per ad set, flexing up to 12 only when every ad is genuinely strong. Never add new ads into an already-running ad set; launch a fresh pack instead.
Mixing new ads into an existing, already-optimized ad set resets its learning and drags down ads that were already performing. A new pack of 4 to 8 ads gets to compete on a level playing field, and the algorithm can redistribute spend within that pack as ads spike, plateau, or fatigue — without disrupting anything already running elsewhere in the account.
How Long Should You Wait Before Judging a New Ad?
Longer than most stores are patient for. 7 to 14 days, depending on spend.
Wait 7 to 14 days before judging a new ad's performance: 14 days if daily spend is under $1,000, 7 days if it's $1,000 or more. Judging creative earlier than this window produces false negatives, because Meta's delivery system needs that time to find the ad's stable audience and spend level before performance data is reliable.
This is where most stores sabotage their own testing. An ad that looks weak on day 2 can still be finding its footing — pulling it early doesn't just waste the ad, it starves the account's learning of a full data cycle. The operational cadence that works: launch on a fixed day (Monday is common), let the window run, analyze on a fixed day afterward, and brief the next pack from what you find. Consistency in the cadence matters as much as the math in the formula.
Does This Creative Testing Formula Apply to TikTok Too?
Largely, yes — the principle is platform-agnostic even though the exact numbers were built for Meta.
The creative testing formula is platform-agnostic; only the source data (Meta's Andromeda-era ad delivery system) is Meta-specific. A competing model used by some TikTok- and UGC-focused creative teams runs a flat 6 new concepts plus 2 hook iterations per 8-ad weekly pack, regardless of spend level.
There's a second, simpler model some practitioners use instead: a flat 6 new concepts plus 2 hook iterations per 8-ad weekly pack, regardless of spend level. It's described as roughly an 80/20 split between new concepts and testing new hooks on a winner, though the literal math (6 of 8) works out to 75/25.
Both models solve the same problem — how much of your weekly ad output should be exploration versus exploitation — but they reach different answers. The spend-based formula scales your new-vs-proven mix with your budget, which fits a store whose spend is actively growing or shrinking. The flat 6+2 model is simpler to run operationally but doesn't adjust for account size.
For a Shopify store spending $10,000 to $50,000 a month — the band most MY and SG DTC brands sit in — the spend-based formula is the better fit, because at that spend level you're still mostly in concept-discovery mode and the formula reflects that automatically. The flat 6+2 model becomes more useful once an account is mature and stable, and a team wants a fixed weekly workload rather than a moving target.
Either way, once a winning creative surfaces on TikTok or Meta, the UGC and iteration principles are the same: a proven concept gets new hooks and new deliveries, not abandoned for the next shiny idea.
| Model | Weekly structure | Scales with spend? | Best for |
|---|---|---|---|
| Spend-based calculator | max(2, ceil(spend/10000)), new% decays with spend |
Yes | Growing or fluctuating budgets, $10k–$200k+/month |
| Flat 6+2 | 6 new concepts + 2 hook iterations per 8 | No | Stable, mature accounts wanting a fixed weekly workload |
Where Does Ad Creative Come From Each Week?
Two sources: your own account's winners, and what competitors are already proving works.
New weekly ad creative comes from two sources: your own account's top-spending ads that hit your KPI with strong incremental attribution, and competitor concepts pulled from a public ad library. Repetition in a competitor's active ads is a signal — they wouldn't keep paying to run it if it weren't still working for them.
Once a judgment window closes, the next week's briefs come from two places. First, your own top-spending ads that are hitting your KPI and holding up on incremental attribution — not just the ones with the lowest surface-level cost per result. Second, competitor concepts pulled from a public ad library: find a real competitor, review their active ads, and note which concepts they keep running. Repetition from a competitor is a signal — they wouldn't keep spending on it if it weren't working for them.
Neither source replaces the other. Your own winners have the highest hit rate because they're proven in your exact account. Competitor concepts add freshness when your own pipeline is running dry. If you're weighing product video versus UGC for the next batch, brief both — the formula doesn't care which format wins, only that the concept behind it is real.
If your brand also produces creator content, the usage rights question matters before you can iterate freely on someone else's face and voice — build that into your production cadence, not as an afterthought once a concept starts winning.
What's the Business Case for Following a Creative Volume Formula?
Consistency. Most accounts don't fail from too little creative — they fail from creative that arrives in unpredictable bursts.
A spend-tied creative cadence gives Meta's delivery system a predictable weekly rhythm instead of the irregular bursts most stores produce — zero new ads some weeks, ten launched at once in others. Launching in large irregular batches makes every ad in that batch compete for the same learning phase at once, which is why a fixed weekly formula outperforms ad-hoc launches.
A store that launches ads only when someone gets around to it has no data rhythm: some weeks nothing new goes live, other weeks ten ads launch at once and cannibalize each other's learning phase. A formula-driven cadence — even an imperfect one — gives Meta's delivery system a predictable weekly rhythm to learn from, and gives your team a fixed, plannable production workload instead of a scramble.
None of this replaces watching your actual marginal ROAS as you scale creative spend. The formula tells you how many ads to make. Your margin math tells you whether making more of them is still profitable at your current spend level.
Treat the whole model as one practical operating system you can adopt and adjust — not a fixed law. If your team's real quality ceiling is lower than the formula's weekly count, respect the ceiling. The quality override exists for exactly that reason.
Frequently Asked Questions
How many new ad creatives should I launch per week?
Launch one new ad creative per week for every $10,000 in monthly ad spend, rounded up, with a floor of 2 per week minimum. A $30,000/month account launches 3 new ads weekly; anything below $20,000/month still launches 2 due to the floor. This is the published output of the Creative Volume Calculator's spend-based formula.
What percentage of my weekly ads should be brand-new concepts?
The new-concept share starts near 100% at low spend and decays to a 20% floor once monthly spend hits $200,000 or more. A $20,000/month account runs close to 80% new concepts; a $150,000/month account runs roughly 33% new and 67% iteration of proven winners.
How long should I wait before judging a new ad's performance?
Wait 7 to 14 days: 14 days if your daily spend is under $1,000, 7 days if it's $1,000 or higher. Judging sooner produces unreliable data because Meta's delivery system hasn't finished finding the ad's stable audience and spend level yet.
How many ads should go in one ad set?
The practical sweet spot is 4 to 8 ads per ad set, flexing up to 12 only if every ad is genuinely strong. Never inject new ads into an already-running ad set — launch a fresh ad set instead, since mixing resets the learning phase for ads that were already performing.
Is more ad creative always better for Meta or TikTok ads?
No. Meta's current ad-matching system rewards ad quality over ad count, and roughly 1% of ads typically drive the majority of account revenue. A brand that can only produce 3 strong ads in a week should launch 3, not pad the week to 9 with weaker filler just to hit a target number.
Keep Reading
Free Download
We audited 10 Shopify stores doing $1M–$20M. Here are the 15 leaks every one had.
Real screenshots (names blurred). Most operators have 8 of the 15 and don't know it.
From 10 audits across SG, MY, AE — fashion, beauty, electronics, food.
PDF in your inbox in 30 seconds. 1 email. No follow-up unless you ask.
