Smarter Ads Through Artificial Intelligence

“Hey Google, optimize my website for maximizing conversions and revenue. And while you’re at it, please do something about the user experience as well.” 

Okay well, we’re not quite there yet, but the last few years have seen artificial intelligence make a serious entry in the realm of online advertising and ad tech. The likes of Google Assistant and Alexa are becoming daily drivers in our lives promising more personalization, increased productivity and predictive entertainment. 

Years of insights on visitor behavior and consumer analytics are now being fed through advanced AI engines and sophisticated machine learning algorithms in a bid to afford improved conversions, an optimized user experience across digital assets and most importantly--better ROIs.

Today, digital product managers and online publishers have access to a growing number of metrics and KPIs for gauging performance. As such, the excess of data can be overwhelming. The sheer number of dimensions demanding optimization make it challenging for individuals and teams to stay focused on larger goals. Growing competition and common access to information that was once confined to the arcane circles of specialists has everyone thinking about SEO, user experience and engagement. Online entrepreneurs have become so obsessed with quants that they’ve stopped relying on good old-fashioned intuition to design rich user experiences for humans.

Resultantly, the last few years have seen new intelligent tools and platforms that are enabling online publishers and webmasters to automate most of their testing and optimization activities. Solutions offering A/B and multivariate testing capabilities have existed for a while, however, the process of designing experiments and test cases is by and large a manual one and has its limitations. 

The ability to test ads on websites, software and mobile apps from multiple networks and virtually unlimited permutations of placements, formats, sizes with minimal intervention has been the need of the hour. The new wave of ad-based automation solutions is driven by sophisticated deep learning algorithms and artificial intelligence paradigms catering to the growing need of smart ad tech. Constantly studying and learning from the behavior of online visitors and product users, these tools take on most of the grunt work which has hitherto cost much time and caused serious misery to digital professionals.

Most publishers are leaving a lot of money on the table due to unoptimized ad placements and sub-par user experience. Being a publisher myself, I’ve had the chance to personally try out a couple of AI-based tools over the last few months on my websites.

Let There be Ads…and there are, well…ads

A welcome addition to the repertoire of website owners that use ads for monetization, a flick of a switch can now actually trigger ads across websites and constantly test for the best placements, formats and sizes. These platforms also potentially overcome a phenomenon called ‘ad blindness’ which affects many websites running ads. Over time visitors become unreceptive to ads in their familiar locations.

Proliferation of AI

Companies such as Google and Amazon have been the harbingers of artificial intelligence. They have managed to successfully fuse AI in their products and services that have become a staple in our everyday lives. it’s no surprise that Google’s AI-first approach has made its way to the online advertising space that it very much pioneered and turned into a multibillion-dollar cash cow. Google search is also powered by its AI-based RankBrain that can better handle never before searched queries much better. In fact, RankBrain’s accuracy is now better than Google’s human engineers by a factor of 10% across different criteria—such as predicting the no. 1 result for a particular search query.  

While Google’s parent company Alphabet has made many bets (Other Bets) in a bid to diversify, it’s still predominantly an online advertising company with 87% of its revenues (2017) coming from this space. The online advertising industry which is slated to hit 335 billion by 2020 will always be extremely important for Sundar and his team of merry people. Interestingly Google is not the first one to the party bringing AI-based ad platforms to the public. 

Meet The Ezoic Platform

The most noteworthy platform and one of the pioneers in intelligent ad testing is Ezoic. It comes with a very sophisticated feature-set allowing digital product managers and publishers to test and improve their ad RPM by testing an unlimited number of permutations. It works really well especially if you’re using multiple advertising platforms. For instance, if you’re using a combination of AdSense and Media.net, Ezoic will work out the best performers across different locations on your website. The most important goals of Ezoic and an area where it has managed to achieve success is maximizing RPM and revenue per session. Ezoic claims that by focusing on the best performing ads, a better overall user experience for visitors is achieved through maximizing revenue through a minimum number of advertisements. A typical approach adopted by many webmasters is to increase the number of ads they’re displaying on their website to improve income which can be counter-intuitive especially after Google lifted its three ads (display and link units) per page policy. It requires some initial setup, but is relatively easy if you’re using WordPress. Through its advanced machine learning algorithms, the Ezoic platform tries to understand visitor engagement and browsing patterns.

Google Auto Ads

More recently, Google AdSense rolled out its Auto Ads featured to majority of its partners. Over two million publishers use AdSense making it one of the largest ad networks. Auto Ads promises an enhanced user experience and improved revenues by the flick of a switch. For many new publishers where the primary goal is to create quality content without getting into technical details, this is a great option. Auto Ads not only automate the process of inserting Google ads across a web property but also tests countless permutations involving placements, formats and sizes. This gives many content creators the freedom to start monetizing from get-go without worrying about manually testing different combinations. The only caveat with Google’s platform is that it works on its own ad networks and inventories. If you rely exclusively on AdSense, this is a great option. Considering that Google is rolling out new ad units all the time, Auto Ads is great news for AdSense members. 

Advanced publishers may need to approach automated ad testing platforms with a grain of salt. The level of convenience afforded by these platforms comes at the expense of lesser control and lack of visibility. Relying on these platforms will not provide you with much insights into what’s working and what’s not. You may not be able to ascertain the best possible ad strategy for your digital properties leaving you with a solution that is more like a Blackbox providing no information about its inner workings. 

Both Ezoic and Auto ads promise to take the hassle out of testing endless combinations of ads that will yield maximum RPM and revenue for publishers. Many publishers who’re new to ad monetization may be leaving a lot of money on the table due to insufficient testing. When I started publishing and trying out ads for monetization, it took me a fair amount of trial and error to find a winning combination. However, some of the changes in terms of ad placements and formats literally quadrupled my revenue. 

AI-based ad optimization solutions are ushering a new era of intelligence which will prove to be immensely useful for both publishers and advertisers. Over time, we will see more of such platforms in the market and a rise in their adoption resulting in a vastly simpler learning curve in achieving success in online ad monetization.