Ad creative that converts: how to create ads that work.
Ad creative converts when it grabs attention in the first seconds, makes clear what the person gains, gives a reason to believe and asks for an action. Beauty helps, but the message decides. And since nobody gets it right the first time, the method is to produce variations and let the numbers choose.
Fabrício MartinsFounder of Martin Marketing & Design
In short
- What creative is: the ad itself: image, video, copy and call to action.
- What weighs the most: the hook. Without it, nobody gets to see the rest.
- How to get it right: variations of the same idea, tested with a small budget.
- When to replace it: when cost per result goes up and click-through rate goes down.
What ad creative is and why it weighs so much
Creative is everything the person sees in the ad: the image or the video, the copy, the headline and the button. On social media, it does most of the work.
The reason is simple. The platforms have automated almost everything else. Today the system itself finds the audience and adjusts the bids. What it does not do is create the message. Two advertisers with the same budget and the same audience can have very different results just because of the piece.
That is why treating creative as "the artwork for the post" wastes the strongest lever in the campaign. It is strategy.
The anatomy of creative that converts
Pieces that look very different usually share the same structure underneath. There are four parts, in this order:
- Hook. What makes the person stop scrolling. It lasts one to three seconds.
- Promise. What they gain. A benefit, said the way the customer would say it.
- Proof. The reason to believe: a result, a testimonial, a demonstration.
- Call to action. What to do now, with a clear verb.
When an ad performs badly, the problem is almost always in one of these parts. Lots of people see it and nobody clicks? The promise or the call to action is missing. Nobody gets past the start? The hook is weak. Lots of clicks and few sales? The ad’s promise does not match the page.
The hook: the first seconds
In the feed, a person decides in an instant whether to keep looking. The hook is what wins that instant. Four types work often:
| Type of hook | How it starts | Example |
|---|---|---|
| Problem | Names the customer’s pain | "Your site gets visits and nobody asks for a quote?" |
| Result | Shows the destination | "A full calendar for the whole month, without depending on referrals" |
| Contrast | Puts the before next to the after | The same page, before and after the fix |
| Demonstration | Shows the product in use | The service happening, with no introduction |
What does not work as a hook: the company logo, an opening sting and lines like "Hi, how are you?". The brand appears, but after attention has been won.
With video, remember that many people watch with the sound off. The hook needs to work in the image and in the text on screen.
Formats and when to use each one
| Format | Good for | Watch out |
|---|---|---|
| Single image | A direct offer and a short message | Too much text in the artwork is unreadable on phones |
| Carousel | Several products or a step-by-step | The first image needs to convince on its own |
| Short video | Demonstrating, explaining and building trust | The first seconds decide everything |
| Video that looks like content | Testimonials and real use | It needs to look true, not staged |
There is no champion format. There is the format that suits what you need to show. A service that is hard to explain calls for video. A promotion with a price calls for a direct image. When in doubt, produce the same idea in two formats and compare.
All of them need to be born vertical, filling the whole phone screen. Adapting a horizontal piece afterward always leaves the result looking like a leftover.
Video that looks like real content
One of the fastest-growing formats is the video that looks like it was made by a customer, not by the brand. A person talking to their phone camera, showing the product or telling their experience. It is known as UGC, for user-generated content.
It works for two reasons. It does not look like an ad at first glance, so it gets past the filter of people who have learned to ignore advertising. And it shows a real person, which builds more trust than a heavily produced piece.
A simple script for this kind of video:
- The situation. The problem the person had, in one sentence.
- The discovery. How they found the product or service.
- The use. The experience, shown on screen.
- The result. What changed, with one concrete detail.
- The invitation. What to do to get the same.
The limit is the truth. A made-up testimonial or an exaggerated result misleads the customer and can get the ad account shut down. Use real accounts, with permission, and make it clear when it is advertising.
The copy that goes with the piece
Ad copy carries less weight than the image, but it can win or lose the click. A few practical rules:
- The first line is the only one guaranteed. The rest is hidden behind a "see more".
- Talk about the reader. "You" appears more than "we".
- One ad, one idea. Three benefits competing for attention add up to less than one well explained.
- Use the customer’s words. The ones they use on WhatsApp, not the ones in the corporate brochure.
- End with the action. What happens after the click.
The ad button also communicates. "Learn more" asks for less commitment than "Buy now". Choose the one that fits the moment of whoever will see it.
How to produce variations without multiplying the work
Testing does not mean creating ten pieces from scratch. It means building the same idea in swappable parts. A simple way:
- Three hooks. One problem, one result, one demonstration.
- Two formats. Image and video.
- Two calls to action. One direct, one softer.
With these parts, there are already twelve possible combinations from a single round of creative work. There is no need to run them all. Start by varying only the hook, which changes the result the most, and keep the rest the same.
Golden rule: change one thing at a time. If two pieces differ in everything, you find out which one won, but not why.
This way of working needs organization. Name each piece by what it tests, such as "problem hook, video, direct call to action". Without that, a month later nobody remembers what was compared.
How to read the result
Likes and comments do not pay the bill. The numbers that say whether creative works are different:
| Number | What it shows | If it is poor |
|---|---|---|
| Retention in the first seconds | Whether the hook holds | Change the opening |
| Click-through rate | Whether the promise interests people | Rewrite the promise or the call to action |
| Cost per lead or per sale | Whether the piece pays back | Check the landing page as well |
The third one is what decides. An ad with lots of clicks and few sales is not a good ad. It is an ad that promises one thing and delivers another. To do the full math, see how to calculate ROAS.
Give the test time. Deciding on two days of data is flipping a coin. Wait until the piece has enough views and results before declaring a winner.
Fatigue: when to replace the piece
All creative wears out. The same people see the same piece several times, stop reacting and the cost goes up. The signs:
- Click-through rate falls week after week.
- Cost per result rises without anything else having changed.
- Frequency, the number of times each person has seen the ad, is high.
- Comments appear from people saying they have seen it too many times.
The answer is not to throw away what was working. It is to refresh in parts: the same script with a different hook, the same promise with a different image, the same video with a different opening. The winning idea continues, with a new face.
Anyone advertising all year needs a production routine, not one campaign per semester. A new batch of pieces every few weeks keeps the account healthy.
Converting without taking the brand apart
There is a real tension between performance and identity. The piece that converts best today may be the one that looks least like the brand. The way through is to define what is fixed and what can vary.
Stays fixed
- The brand’s colors and typography.
- The way of speaking.
- The sign-off at the end of the piece.
- The company’s core promise.
Can vary
- The hook and the opening.
- The format and the length.
- Who appears in the video.
- The offer and the call to action.
A recognized brand converts with less effort, because part of the trust is already built. A set of brand guidelines with ad templates resolves much of this tension.
Creative ideas by type of business
- Local service. The before and after of a real job, the professional answering the customers’ most common question, a testimonial recorded on site.
- Online store. The product in use, the box being opened, a size comparison and the detail the catalog photo does not show.
- B2B service. A client result in one sentence, a clip from a meeting or behind the scenes, a spreadsheet or dashboard showing the problem being solved.
- Course or event. A clip from a class, a testimonial from someone who has taken part and what the person will be able to do afterward.
- Clinic or practice. The professional explaining a procedure, the treatment space and answers to questions about preparation and recovery, always within the advertising rules of the profession’s regulatory body.
The common thread: show what is real. Generic stock imagery is the fastest way to look like everyone else. A simple photo of your work, your team or your customer carries more information than the perfect image a hundred other companies also bought.
In regulated fields, such as health and law, check the rules before producing. Promises of results and before-and-after images may be prohibited.
Common mistakes
- Opening with the logo. Nobody stops scrolling because of a logo.
- Saying everything at once. A piece with five messages gets none across.
- Text that is too small. If it cannot be read on a phone, it does not exist.
- Using the same piece everywhere. Each placement calls for a format.
- Judging by taste. The team’s favorite piece is rarely the one that sells the most.
- Sending the click to the wrong page. The ad promises and the landing page needs to deliver.
Frequently asked questions
Three to five variations of the same idea are a good start. With fewer, you have nothing to compare. With many more, the budget spreads out and no piece gets enough views to show a result.
Not always. A professional finish conveys trust, but it does not replace a clear message. Often a simple video shot on a phone beats an expensive piece, because it feels more real. The ideal is clarity first and polish second.
You can reuse the idea, but not the piece. In Google search, the ad is text and answers a query. On social, it is an image or video and needs to interrupt. Each channel calls for its own version.
It is not mandatory, but it usually helps in service businesses, where trust in the person matters. When the owner does not want to appear, real customers, the team or a demonstration of the work play the part.
When the numbers ask for it. High-budget accounts wear pieces out in a few weeks. Smaller accounts take longer. Watch click-through rate and cost per result: a drop in one and a rise in the other tell you it is time.
Conclusion
Creative that converts is neither luck nor inspiration. It is structure: hook, promise, proof and call to action. And it is method: vary one part at a time, read the right numbers and refresh before it wears out.
If producing pieces is the bottleneck of your campaigns, get to know Martin’s ad creative work.
