AI Takes Over Ad Creativity

07/31/2020
by
Sasha Schriber
ai takes over ad creativity

Paid performance marketing is a true art, and one has to be experienced enough to make the right assumptions from the start. Usually, no one wants to spend too much money on experimenting, without an acceptable return on this investment in the end. If the first assumptions were wrong or keeping low-performance rates, a human marketer has to be able to optimize the ad at a very quick pace.

There are already many places where AI can be applied in online marketing, and specifically in the paid performance marketing (PPC) field. The life cycle of a paid ad, or the "ad journey", i.e. an ad that was created with the aim to be published on Google, Facebook, Instagram, LinkedIn, Twitter, Pinterest or any other platforms that provide paid targeted advertisement as a service, can be described as the following:

Ad creation - where the ideation process happens in a human's head (business owner or a marketer). It's the ad title, ad image or video and a description text to be produced. It is also suggested to work with multiple versions of ad creatives to achieve expected results faster by spending less money overall.

Ad budgeting and placement - deciding on how much you will invest and on which platform should you place the ad. It is a known fact that less is more and focusing on 1-2 platforms is generally better than focusing on too many at once. Of course, one needs to know which platforms are best for your specific product or service.

Ad optimization - watching and modifying your ad performance day by day or hour by hour, based on multiple parameters: age and gender of your audience, keywords and interests, cost per view, cost per click, which platform performs better and many others.

Ad placement is somewhat already explored, however, it does not pose a serious challenge for technology - it’s clear that the knowledge of ad placement and its compliance with any platform guidelines could definitely be taken over by a machine. The last part of the process (optimization) has been already severely taken over by tools existing on the market: the process of targeting the ad to the right customers, and optimizing it during the life cycle of the ad campaign can be easily delegated to a system that has way more knowledge about all types of businesses and campaigns than you or any human marketer. Cross-platform optimization, i.e. reshuffling the budget between more than one platform is already mastered by us at Nanos and works even for smallest ad campaign budgets starting from $50 which are to be spread over several platforms and days.

What is really important to note that so far ad creation has been barely touched by researchers and AI. Why? It’s just too difficult, even for a human to come up with a likeable and visually-appealing ad which has a catchy text as a title and as a description. The ad should be also grammatically correct and contextual, i.e. relevant to the website or a landing page where it is going to drive traffic to.

But guess what? Last month we at Nanos.ai received results from a user study conducted with 20+ participants. The questions were rather simple: here is a short description of a product or service and here are two ads about it generated from the description. Participants had to pick the ad that they liked better and assess if it was grammatically correct, likeable, and contextual. Overall, 5 sets of textual ads and landing pages were presented. We asked a few simple questions: did they like the ad in general? Was it grammatically correct? Was it relevant to the information provided in the product description?

What participants didn't know was that one of the ads was generated by a human marketer and the other - by Nanos machine learning algorithm. We were very curious to see the results, as, by default, we assumed that participants would choose ads written by a marketer. However, most preferred ads generated by the machine. We were surprised, but also very excited at the same time. It seems that ad creativity, in particular the text, can now be solved without major human intervention. What does it mean for you, as someone who wants to place an ad online? You don't need to think very much about what to put in your ad anymore and on top of it, you can instantly create multiple versions of the same ad and test those on any platforms you want.

This is a significant improvement in the workflow of many marketers and agencies who currently spend enormous amounts of time coming up with and iterating ad content prior to publishing it online. This includes making sure that the ad resonates with its intended target audience, that it addresses the potential need of the viewer. Imagine instantaneously creating an unlimited number of ads, briefly checking them for sanity reasons, and instantly turning them in for testing or removing the ad fatigue from the old ads.

This is just one of the breakthrough milestones showcasing how AI can work hand-in-hand with a professional marketer and help save time and money. With AI becoming more and more intelligent and being able to detect any potential ad issues faster, it offers a promising look into the future of marketing. So, are you ready to hand in your ad creation to an AI?