Remix and remake a proven ad

1. The problem this fixes

Most people who struggle to advertise spend too much time making new ads, and not nearly enough time doubling down on the ones that already work.

What makes that mistake feel reasonable does not hold up.

You tire of your own ad first. A story is told of a company head who walked past the same print ad, pinned to a wall, every morning for three months, and finally asked when they were going to stop running it. The answer: it had not run once. He had seen it a hundred times because he worked there; the public had not seen it at all. You make campaigns for the people you sell to, not for yourself. Once an ad works, run it until it stops working, not until you are bored with it.

"New" does not mean starting from zero. New means new to the viewer, and buyers have a far lower threshold for what counts as new than you do. A fresh version of a winner is new to them and is still, plainly, more of what already worked. When one of those versions beats the one before it, the ad has got better. You get close to the old results, or better ones, without wearing the audience out.

The method on this page is described as doing more, doing it better and doing something new, applied to advertising: it produces volume, makes that volume better, and generates new winners. Choosing among those moves across a whole business is SOP 40 — Choose the next growth move.

2. Why the effort splits eighty and twenty

The reasoning is the power law, also called the Pareto principle: 80 percent of advertising results come from 20 percent of winners, and only 20 percent of results come from 80 percent of your effort. Stated again at the close: 80 percent of sales come from 20 percent of ads.

In an ideal world every ad you made would be like the 20 percent that bring 80 percent of the results, and the claim is that doing so gives 5x the output. If you put 80 percent of your effort into variations of ads that are already proven, you will make more winning ads than if you start from scratch every time.

A thought experiment makes the point. Two advertisers each make 100 ads. One makes 100 brand-new ones. The other takes his top five ads from the month before and makes 100 versions of those five. The one who put his effort into copying his winners wins the next month, without question, and the method is said to be likely to make you a lot of money — at least, it has done so in one case.

The split is not a mood; it is where your hours go:

Share of your effort	What it goes on
80 percent	Remixing and remaking the ads that already won (steps 2 and 3 below)
20 percent	Testing totally new ideas, looking for the next winner (step 4 below)

3. The four steps

The condition on all of it: this works only if you have already made ads. You cannot improve ads that do not exist. The moment you have even a single winner, you can start.

Step	What you do
1. Identify winners	Find the ads that outperform: your top 20 percent, your top 10 percent, your top 5 percent, and so on. The more ads you make, the more winners you have to choose from, and that is where the large output comes from
2. Remix winners	Keep the raw footage of a winner and change the edits — everything done after filming (section 4)
3. Remake winners	Record a new ad on basically the same winning script, with a real-world variable changed rather than a digital one (section 5)
4. Try new ideas	Spend the remaining 20 percent of your time testing totally new material. Keep running variations of the previous winner until a winner comes out of the new material; then put that one through steps 2 and 3 as well

Why remixes come first. They are the fastest to make, and they also wear out the fastest. Get them running quickly to keep your ads performing, which buys you the time to do step 3.

Why remakes come second. Real-world changes take longer, but the ads they produce can keep you going for months, and in some cases years.

A new winner does not retire the old ones. You keep using them; you simply have one more winner to make winners from.

Flag: vary the winners, or rework the runners-up? Section 2 puts 80 percent of the effort into variations of ads that are already proven. A second view, from an ads practitioner: when five ads launch and one stands out, it is common for the second and third places never to be looked at again; go back through your past spend, take those, and rework them with what you know now, which usually has a better chance than more versions of the ads you already run. This page does not settle which reading is right.

4. Remixing: the same footage, edited again

To remix, you need a winning ad. Remixing can be done straight away on the winners you already have, especially with AI, and it is the fastest route to money. Think of the list below as a master checklist of changes you can make to one winning ad. The remixes are listed from easiest to hardest:

Order	Remix	What you change
1	New speed	Play the ad slightly faster: 1.1 to 1.2 times normal speed, without making the voice in the ad sound strange. You can also slow it down. Sometimes a faster or slower version converts better; if it only converts the same, it still converts like a top performer, which makes it one
2	New filters	The look of the ad: black and white, sepia, higher contrast, or a filter in the style of a social platform, so your video looks unlike the ad you started from
3	New background or border	Swap the background color, or add a thin trim around the edge of the ad (especially with images) and vary its colors. The closing example swaps a white backdrop for a neon-green gradient border
4	New fonts or captions	The typeface of the ad's text or headline. The closing example sets the first three words, or the whole subtitle, in bold comic-style capitals
5	New headline	Different overlay copy. On some platforms you can place a headline above the ad; or put one across the top third of the ad itself. Keep the rest the same
6	New medium	Take a winning hook or headline from a video and make it a standalone still image
7	New format	A version native to each placement on each platform: square, vertical and so on. The closing example re-cuts a vertical selfie ad into a square 1:1 version, and says native placements improve performance a great deal
8	New meat	Keep the winning hook and splice in a different middle section you have already recorded. This one needs editing skill. The closing example keeps the winning hook and splices in a different case-study clip as the body
9	New effects	Add or change motion graphics, stickers and animations. The ones that matter most are the early ones, in the hook, because that is what most people see first. The closing example drops a motion sticker over a revenue number the moment it shows on screen, or shows visual proof right after a hook built on a number

Vary all of them. Once you have a winner you can make a very large number of versions by changing combinations of these variables, and since every one of them is done on a computer, you get many more ads without extra recording time. The instruction is to use them all and test them all.

The cost of remixing. Remixes tend to wear out faster than remakes, mainly because they tend to look more like the original. That is good, because the original worked, and bad, because a close copy has a shorter shelf life than an ad that looks more different.

5. Remaking: the same script, recorded again

Remakes happen during recording: you change the real-world parts of the ad, so you have to record again (or, with AI, generate new subject matter). That takes more effort, which is why remaking comes second, but this is where a winner can keep performing for years. If you really want to advertise like a pro, run remixes and remakes at the same time.

Each remake keeps what made the ad win in the first place. The remakes are listed from easiest to hardest:

Order	Remake	What you change
1	New clone	Nothing on purpose: record the same ad again in a new session. No recording session comes out identical to another — time of day, weather, lighting and sound all leave small differences, enough for it to convert like a new ad
2	New props	If you used no props, add some; if you did, use new ones or take them away. Costumes, hairstyles, jewelry and accessories count. The change is large, but the winning script stays, so the ad stays likely to convert. The closing example swaps a stack of cash for a silver platter holding sign-up sheets
3	New examples	Use a different example as either the hook or the meat — change the names and numbers inside the same script. In the case given, an ad that opened on the previous month's lowest performer on a leaderboard was recorded again the next month with the name and the number swapped, the rest left alone, and it worked again
4	New setting	The same script in a different place: a beach, your car, your office, a sidewalk instead of indoors. A different place makes the same ad new again
5	New talent	Find, coordinate and direct a different person to deliver it. Ideally, pick people who look like your buyers: if you sell to midwestern men aged 18 to 25, make more versions with different men of that description; if you want more women, use more women. Even a second person of your own demographic will at the very least look different. This is called a wildly underused variation
6	New combination	Several changes at once: a new prop with a new post-production effect, or new talent saying the script under a new filter or headline. The closing example stacks a new setting, a new prop and a new caption style

Why this is leverage. You put the most effort into the thing with the highest likelihood of winning, so your time earns more.

6. When you are starting out, and what the loop looks like

If you are just starting, run as many different ads as you are able to until one wins. The moment one does, put it through sections 4 and 5 and make more winners. Eighty percent of your energy goes on variations; the other 20 percent goes on a completely new ad that beats the original winners. The variations keep you competitive now; the new ads keep you competitive for good.

Three ideas sit underneath the loop:

• Enough volume takes luck out of it. Produce many ads, find the few that work, put 80 percent of your effort into remixing and remaking them, and spend the other 20 percent hunting for the next winner. One winner is enough to start; the variations keep it going.
• Winners are rare, so get everything out of them. When an ad finally clicks, run it hard until it dies, then run it again as variations: new colors, headlines, props, speeds and fonts; new settings, talent and examples. Each small twist costs a fraction of inventing a new winner and gives steady returns, and steady winners take the swings out of advertising.
• The frameworks work for any business. That idea belongs to SOP 114 to SOP 117.

The timing given: remixes you can use tonight; remakes you can film this weekend.

7. Where the instructions do not line up

The five-times figure. The claim is that making every ad like the winning 20 percent gives 5x the output. Check the arithmetic: if a fifth of the ads bring four-fifths of the results, each winner brings four times the average, so a full set of winners gives four times the results — or five times as many winning ads. This page does not settle which reading is right.

What a clone keeps. A clone is "same everything", recorded again; in one example, a clone keeps the gestures, script and background but changes clothing and lighting. Read the first way, a change of clothes is a new-props remake, the next rung up. Read the second way, a clone may change clothes and light. This page does not settle which reading is right.

Testing many variations at once. This page says to vary all the remixes, combine them freely, and test them all. SOP 41 — Run one test a week, section 2, says to run one test per week per platform, because testing several things at once means you never learn which one worked; its section 5 applies the rule per pathway, and SOP 41 is written for getting more leads from advertising you already run without adding volume. Neither page says whether its rule reaches the other's case. Followed this way, one winner turns into many live variations quickly, and you may not know which change carried the result. Followed SOP 41's way, you learn exactly what each change did, and a winner turns into variations far more slowly. This page does not settle which reading is right.

8. What this page does not decide for you

• The measure that makes an ad a winner. The cut-offs given are top 20, 10 or 5 percent. The remix notes speak of versions that convert better, but no single number is named for ranking winners. Not established on this page.
• When a remix has worn out. Remixes wear out faster than remakes; beyond running an ad until it stops working, no sign or period is given for calling one finished. Not established on this page.
• What to spend testing the variations. This page gives no figure for a test budget.

A target cost for each ad, with rules to cut or scale it, is SOP 275, section 7, and its test budget is section 5.

9. The checklist

Question	The answer
Who runs this	You, or whoever makes and runs your ads
When to start	As soon as you have a single winning ad
Where most effort goes	80 percent into remixes and remakes of winners
Where the rest goes	20 percent into totally new ideas
The four steps, in order	Identify winners, remix winners, remake winners, try new ideas
Remixes, easiest first	Speed, filters, background or border, fonts or captions, headline, medium, format, meat, effects
Remakes, easiest first	Clone, props, examples, setting, talent, combination
Which wears out faster	Remixes
Which lasts longer	Remakes: months, in some cases years
When to stop running a winner	When it stops working, not when you are bored of it
What "new" means	New to the viewer
Which remix, clone or test reading is right	This page does not settle which reading is right
What makes an ad a winner	Not established on this page

10. What this page does not cover

The twenty ad frameworks to use when you have no winner yet are SOP 114 to SOP 117, five to a page. Choosing between doing more, doing it better and doing something new across a whole business is SOP 40. The testing cadence is SOP 41 — Run one test a week. Recording hooks and bodies separately so they can be recombined is SOP 34 — Front-load the opening seconds. Running your best free post as an ad is SOP 38.

Ad budgets and bidding are not covered on this page.