How AI is transforming mobile advertising: Insights from industry experts
The rise of artificial intelligence (AI) has had a profound impact on almost every industry – and marketing is no exception
This article explores these topics with key insights from experts at Remerge - a team of mobile marketers who specialize in programmatic in-app retargeting campaigns for some of the world’s biggest mobile apps.
How AI entered the advertising industry
In the early days of computing, even the most primitive of programs were employed to analyze and aggregate customer data to inform human decision making. The turn of the millennium soon gave rise to recommendation engines, and within a decade, programmatic advertising had come into play – where algorithms are used to automate the targeting, bidding, buying and fulfillment of online ad placements on behalf of brands.
Today, the use of AI in advertising has skyrocketed – and it now plays a prominent role in the mobile advertising space. Not only has it enhanced traditional online marketing practices, but it has paved the way for entirely new methods and approaches. While AI automates and simplifies many processes, it also comes with limitations – so it’s important for advertisers to understand the growing prominence of AI in order to distribute ad budgets effectively and ensure a return on ad spend (ROAS).
Here are some ways in which AI is being used today in mobile advertising and how these use cases can impact your brand.
1. A/B testing in connection with AI
As the shift toward online marketing arrived in the early 2000s, A/B testing became a true data-driven science – and today it’s common practice across many online channels. Before long, mobile devices became increasingly popular for consuming content, so some UX designers and marketers began to specialize in mobile experiences. Through extensive A/B testing, they had to learn what worked on mobile – from the placement of CTA buttons on mobile web pages, to the best-performing ad formats for mobile-specific channels like in-app advertising.
AI has become an important part of A/B testing (particularly in the context of programmatic mobile advertising), because it’s much faster and more efficient than manually running multiple experiments. In the past, you may have created variations of ad headlines for Facebook ads or app and web campaigns. The platforms facilitating these campaigns use AI to test the best-performing ad copy. Not only that, but they can learn which headlines are most effective among certain subgroups.
When A/B testing variations of your mobile ads, try to only change one element at a time. If you test two ads but the second one has a different headline and a different visual, you will only learn which ad performed better. You won’t know if it was the copy or the design work that made the impact. It’s the smaller, individual adaptations that give a stronger sense of why certain changes make a difference.
It’s worth noting that because mobile display screens are much smaller in comparison to desktops, ad headlines on mobile have a lower character count, so marketers must test and adapt their copy to fit within these restrictions. Mobile displays are also vertically oriented, so ad creatives must be formatted specifically with mobile in mind. It’s also makes sense to A/B test different creative formats to establish whether video ads perform better on mobile than banners for example.
2. How AI is used to optimize programmatic advertising efforts
Across the apps and websites you visit, you’ll inevitably have seen some form of advertising. These ads aren’t there by chance, but are in fact dynamically, and very strategically placed by intelligent algorithms. Within milliseconds of your visit, apps and websites will harvest your data along with information about the available ad slots they have, then share these details with a third party called a Supply Side Platform (SSP).
The SSP packages this information into a ‘bid request’ which they offer up via a virtual auction. Then, Demand Side Platforms, known as DSPs, (companies who buy and fulfill ad space on behalf of brands), will examine the bid request, determine if the ad opportunity is relevant for any of their clients, and then place a bid for the opportunity to show you (as the user), one of their client’s ads. The value of a user and their likelihood of converting, alongside factors such as budget, market landscape and campaign objectives, all influence how the DSP makes the bids. The highest bid of all the competing DSPs will win the ad placement.
These entities and processes are the driving force behind programmatic advertising – and almost every step of it is governed by artificial intelligence and the various subsets of this discipline, such as machine learning – where data is used to recognize patterns and make predictions. AI’s application in programmatic advertising includes analysis of datasets to help build out customer segmentation strategies and refine their targeting. AI can also be used to optimize ad placements by identifying the best times to show the ads. It can even improve bidding strategies to generate cost effective traffic, and it can help in detecting ad fraud when apps and websites use bots to create fake ad engagement.
Speaking on AI use cases within programmatic in-app advertising, Jacky Ma, director of data science at Remerge explains, "Our application of AI is multifaceted." Ma breaks down Remerge's current projects and how AI plays a pivotal role in them:
Predictive modeling and bidding decision:
Rooted in the more traditional understanding of AI, we harness its power in predictive modeling, which heavily informs our bidding decisions. Our primary use of AI centers on price optimization. By predicting a bid request’s value and its market competition, we can strategically position our bids. To gauge the impact of our changes, we also employ intricate causal models that enable us to ascertain the exact impact of our alterations. Our other ongoing projects delve into how different contexts influence conversion behaviors, and how the market landscape varies across app publishers.
Language learning models (LLM):
Moving away from traditional AI, a modern trend revolves around LLMs. These are machines capable of understanding and responding in human languages. While they don’t tie directly into our core product, they present robust use cases, especially in automating tasks that were traditionally manual. As part of our exploration, we’re piloting projects that automate document creation and analysis.
Holistic AI implementation:
Recognizing the transformative power of AI, we’ve established a dedicated working group. This team is responsible for overseeing AI’s incorporation across various facets of our business, from tools and training to altering our operational and business models.
3. How generative AI is being used for mobile ad creatives
The recent rise of generative AI has been a hot topic for creators and marketers. Being able to quickly turn written prompts into highly-personalized, top quality ad creatives has plenty of advantages. However, there are still challenges with AI-driven ad personalization on mobile.
Nick Barnett, principle solutions engineer at Remerge explains: “I think the main issue will be privacy. Both Android and iOS are working on ways to obfuscate an individual user’s data so that advertisers can't identify who they are sending an ad to. I expect them to fully solve this problem soon, so I think realistically AI will end up more as a tool for generating ads rather than dynamically personalizing them. Making creatives is time consuming and having AI do it for you will be a lot faster and easier.”
Elaborating further on the benefits and challenges that come with AI and user privacy, Dan Jones, Remerge’s global director of creative operations explains: “One key advantage, recognized by design experts and industry leaders, lies in AI's ability to empower creative strategies used by design teams and marketers. By leveraging AI's capabilities, creative efforts can be enhanced and amplified to elevate ad performance. As we navigate the complexities of a privacy-conscious advertising environment, the strategic utilization of AI becomes paramount.
"This adaptive approach resonates across diverse audiences, addressing the limitations of individual user tracking while upholding privacy standards. This pivotal shift lays the groundwork for effective and ethically aligned advertising. By embracing this powerful synergy and strengthening our creative teams with AI, brands unlock the potential to craft compelling stories that foster engagement, empowerment, and captivation on an unprecedented scale. Within Remerge, we are pioneering AI solutions to generate impactful creative strategies, enhance video performance and scale creative production for impactful outcomes."
While generative AI can help with scale and production speed, AI-generated ad visuals should always be checked carefully before distribution – at least for the foreseeable future. Pan Katsukis, CEO of Remerge explains: “It’s unlikely a top brand would want to create and show AI-generated ads for its products or services without any approval, so there will still be limits, boundaries, and a need to put processes in place that will protect their brand image.”
Executive summary
Overall, AI’s looming presence in the modern age of advertising is undeniable. It plays an essential role in many aspects of mobile marketing, from A/B testing and programmatic advertising technologies, to creative strategy and development. As with any new terrain, it’s becoming ever-crucial for today’s digital marketers to understand the risks and rewards that come with the use of AI, and the way they use it to advertise their brands.