How Artificial Intelligence (AI) is changing ad creative optimization

As you might know, AI has been used for multiple things in many industries and adtech is not an exception. However, AI models for ads have been almost exclusively used for targeting and audience segmentation. For example, in programmatic advertising AI is being used to gather data from users that visit websites and these are used to segment audiences. But the marketing industry is still leaving a huge part of the potential impact untouched: optimizing the creative content of each ad.

‍

AI for optimizing ad creative

‍

If you’re working on a Marketing campaign you surely have to decide on which assets will work better. But how do you do it? You probably use traditional testing, such as A/B tests — or maybe you just stick to your pure intuition. When does the problem appear? You might have to test many assets, but you just don’t have the time and money to do that and you need to have everything ready in less than a week. All those creative decisions you’re doing without data could mean a huge risk for all the precious investment your company has spent.

‍

Besides, after you spent all that time and money deciding on your creatives, you didn’t find a way to improve your decision-making process for the next campaign. It’s simple: traditional tests or surveys take too long and cost too much. You can’t use it for most ads.

‍

Where does AI come in to save the success of your campaign? Through algorithms, the technology can understand the cognitive impact of visuals across memory, attention, and conversion, and is also able to predict them and make recommendations to improve the assets in seconds, without wasting all your marketing budget.

So how does AI change the game of ad creative optimization?

‍

1. It is more accurate.

‍

AI allows systems to process an immense amount of assets, thus discovering deep patterns of cognition and behavior that are impossible to identify with human analysis. The more data an AI system has, the more it improves its accuracy. It’s difficult to admit, but we, humans, aren’t good at predicting what a consumer will first see from an ad or which scene will be more remembered.

‍

On average, humans predict correctly which asset is better 52% of the time — those are approximately the same chances as flipping a coin.

‍

Memorable’s models are continuously tested against humans to measure their prediction accuracy. How precise are we talking? We average a 0.03 absolute error at predicting memorability and 97% of accuracy in saliency. Now, that’s accuracy.

‍

2. It is scalable.

‍

Because of the use of AI in advertising, we are able to test hundreds or thousands of different ads and see which one is the better option, at the same time and without having to do it manually. Do you have to review hundreds of assets for the next campaign? Don’t panic, with AI it’s no problem.

‍

Upfield, a major plant-based foods company, decided to test all 5 iterations of an ad with Memorable’s AI platform until it reached 21% - 25% result improvements — all of this before spending a single dollar in media or showing it to a single consumer. Best of all: they didn’t have to wait to get the results, so they just seamlessly integrated Memorable as a step in their iterations!

‍

Traditional methods like surveys or A/B tests don’t have the ability to reach that scale of tested assets: it’s a completely different game in terms of time and money. With live testing, companies spend weeks and an average of 30% to 50% of their media budget before they can even see which assets work better. None of these mechanisms enable testing several assets and much less using data to inform decisions throughout the production process.

‍

This is because those methods rely on a brute-force approach that doesn’t leverage past data: for every new ad, they need to collect information again, as if we knew nothing about past experiences. Good AI systems leverage every bit of past data to learn the patterns associated with good and bad performance.

‍

While most ads are optimized using fewer iterations in a long time with lower results, AI allows the really committed marketers to test and optimize every creative decision until the ad reaches the peak of its performance.

‍

3. Faster than ever.

‍

AI platforms allow you to have instant predictive results and recommendations on how to improve the assets in less than 2 seconds. Testing in seconds means that marketers and creators don’t need to change their ways of working: they can now test every of the existing design iterations, and work with data instead of intuition from early on.

‍

To get key information while using traditional alternatives you need to create both ads, test them separately and wait until each ad has been exposed to enough people before concluding which one works better — and you might have to wait around 3 to 4 weeks till you have visible results.

‍

With an AI alternative, you can see those results instantly and close campaigns in much less time, plus inform every creative decision along the way. Why waste weeks on testing which ad will have a better performance when you can test many ads in no time and make as many changes as you like without showing them to your customers? Besides, with Memorable you won’t even have to pay more. Just test all iterations, in no time.

‍

Now you will have more time to think about the strategy of your brand or what your next innovation will be.

Leave it to the robots

‍

Creative is the key to uplifting your sales, that’s why it’s crucial that it becomes infallible. And now you can do it with data instead of intuition—without increasing costs or times.

‍

The tech world solved data-driven ad placement a while ago. Today, Memorable brings to marketers a solution to solve data-driven ad creative using data. And it’s stronger than any other method out there.

‍

Wondering how much better we do it? Book a demo now and see why the global marketing leaders are already using our technology to optimize their creative.

‍