12 Winners a Year: Why You Only Need One Great Ad a Month to Double a Brand

It's a Friday afternoon and you're three coffees deep, scrolling the ads manager, and there it is. One ad set, green across the board, spend climbing on its own without you nudging it. You didn't touch it. It just took off. That feeling, the one where an ad starts eating budget and the CPA actually drops while it does, is the whole game. And most brands go entire quarters without it because they're measuring the wrong thing.

Here's the reframe I want to hand you, because it changed how we run creative completely. You are not trying to make good ads. You're hunting for rare ones. The kind that don't nudge the account 5 or 10%, they jump it 100 or 200%. You'll find one of those maybe once a month per account. And one a month, it turns out, is plenty.

What a winner actually is

Most people define a winning ad as one with a nice ROAS. I used to as well. It's the wrong definition and it'll quietly cap your growth.

Here's the trap. You launch a fresh creative on a small test budget, say A$20 a day, and it posts a lovely 4x. You get excited. You drag it into the main campaign that's spending A$2,000 a day. And it does nothing. Just sits there, barely spending, refusing to take off.

That happens because at A$20 a day the ad was only ever catching the lowest-hanging fruit, the people already warmed up by everything else in the account. Of course the ROAS looked good. It was skimming. Drop it in with the big dogs and it can't compete, because it was never tested against real volume.

So I threw out the old definition. A winning ad isn't one with a good ROAS at small spend. A winning ad is one that takes the majority of the spend in the account and either pushes the CPA down so you can spend more, or holds performance steady while it scales. That's it. Does it eat budget and keep the economics intact? Then it's a winner. Does it post a pretty number on A$30 a day and stall the moment you scale it? That's not a winner, that's a tease.

Once you hold that definition, you stop chasing 5% improvements. You start hunting for the ad that doubles the account. Different prey entirely.

The maths nobody runs

Here's the part that reframes the whole job, and I'd encourage you to actually do this sum for your own brand rather than nod along.

A real winner, by the definition above, roughly doubles the business when it lands. Not every ad. The rare one, the once-a-month one. So picture a brand sitting at A$100k a month in revenue.

January, a winner lands. The account roughly doubles its working creative and you're now running closer to A$190k for the month once it's scaled. February's mostly spent wringing that one dry. March, another winner. Now you're building on the new base, not the old one. Over a year, if you find even one genuine winner a month and you actually scale into each, you are not adding 12 small wins. You're compounding off a rising number each time.

That's the bit people miss. Twelve winners a year sounds modest. It isn't, because each one builds on the last. The brand that finds one real winner a month and rides it will pull miles ahead of the brand pumping out forty mediocre ads a week and scaling none of them.

To put it another way: the job is not volume of ads. It's frequency of winners. And you only get a winner a month if you're testing with enough discipline to actually surface one and brave enough to kill everything that isn't.

The 70/20/10 cadence

So how do you spend your creative hours to find one of these a month? The allocation I keep coming back to is 70/20/10. I first heard it framed by Alex Hormozi and it's how we run a production week.

If you've got a full day to make ads, here's where the hours go.

• 70% on what already works. Take your current best performers and make more of them. Same concept, same structure, the proven stuff. This is the bread and butter and it's where most of the day goes. Not glamorous. It's the work.
• 20% on adjacent ideas. One step removed from a proven concept. Not a tiny tweak, but not a wild swing either. Something you've started to see a flicker of signal on and want to push further. A new angle on a winning hook. A different proof point under the same headline.
• 10% on wild swings. The genuinely weird ideas. The "my mate reckoned this'd be funny" stuff. Most of it dies. But your next out-of-nowhere winner usually comes from this 10%, so you can't skip it. You just can't let it eat the whole day either.

Get the 70 done first, then the 20, then the 10 if you've still got energy. Most brands invert this. They spend 80% of their creative time on wild new swings, hoping to knock one out of the park, and almost no time milking what's already proven. That's backwards, and it's why their winners are accidental instead of regular.

What this looks like on a real ecom calendar

Let me make it concrete, because 70/20/10 is easy to agree with and hard to actually run.

Say it's a homewares brand and you're planning April creative. The 70% is your evergreen workhorses: the comparison static that's been holding for two months, the founder talking-head that keeps converting, the review-led carousel. You're producing fresh cuts of all three. New backgrounds, new headlines, same skeleton.

The 20% is adjacent. The comparison static works, so you try the same layout against a different competitor objection. The founder video lands, so you script one in the same voice but built around a new use case.

The 10% is the swing. A meme-style static. A trend you've seen blow up on TikTok organically, repurposed cold. A format you have no evidence for but a hunch about.

Then a seasonal moment lands, say an EOFY push or a long-weekend sale, and the mix shifts hard toward the 70 for a fortnight, because a sale is no time for experiments. You ride the proven angles with the discount layered on. The wild swings can wait until the calm after.

Wringing a winner dry

Last piece, and it's where most of the compounding actually comes from. When you find a winner, you do not let it fade and quietly die. You abuse it.

One winning concept should spin out a hundred variations. Black and white version. Different filter. New hook on the same body. Different headline. Different subhead. Different opening three seconds spliced onto the same back half. The reason this works is that the way I scroll is not the way you scroll, so the more variations of a proven idea you put out, the more different people it catches, and the longer it holds before it fatigues.

A neat trick here: slice the first three seconds off your best winner and paste it onto a dozen other ads. You'll often get ten or more usable creatives off one good hook, for almost no extra production. The concept did the hard work. You're just refitting it.

The brands that scale aren't taking endless new swings and praying. They find the rare winner, then squeeze everything out of it before moving on.

Where to from here

So pull up your account this week and ask one honest question. In the last 90 days, how many ads actually took the majority of your spend and held the economics, versus how many just posted a nice number on a small budget and stalled?

If you can't point to one a month, the problem usually isn't your creative volume. It's the testing discipline underneath it, the kill rules, the willingness to scale into a winner instead of admiring it. That's exactly the sort of thing a Signal/Noise Audit is built to surface: where your winners are dying before they get the budget to prove themselves. If you'd find it useful to have someone trace that for your account, that's what it's for.