AI models rise as the cookie crumbles

As Apple and Google lead the crackdown on third-party tracking cookies, the future of advertising technology lies with using artificial intelligence to power a more holistic approach to serving ads.

Advertising underpins the open internet, funding most of the free content that people consume around the world. Tracking cookies present an important way for marketers to get more bang for their buck, by ensuring they deliver relevant and engaging ads to their target demographics.

This is set to change, with Google declaring its Chrome browser will stop supporting third-party tracking cookies in 2022, replacing them with a group profiling system. Meanwhile, Apple is also significantly curbing the use of tracking cookies on its mobile devices.

A recent IAB report warns that publishers could lose up to $US10 billion in ad revenue when third-party cookies are disabled, due to the limitations of their ad personalisation options. According to Google’s research, most publishers could lose 50 to 70 per cent of their revenue if they don’t reconfigure their approach to advertising and data management by 2022.

Without the insights provided by third-party tracking cookies, online impressions can be worth significantly less to marketers. Not just because they have far less information with which to create a relevant experience, but also because they can’t measure whether the advertising was effective.

Much needs to change in terms of making sure that it is still possible to create relevant experiences and to create value through online advertising.

Third-party tracking cookies gather data to better understand individual consumers within their target demographic. With the coming changes, marketers instead need to think more about context and ‘probabilistic’ data. This allows them to better understand the different types of consumers without uniquely identifying them.

To do this effectively requires trawling through vast amounts of data points and signals in search of insight. This is a perfect job for artificial intelligence.

With the current ‘programmatic’ online advertising model, marketers use data such as cookies to identify consumers that they wish to target wherever they go on the internet. This requires marketers to dive into the granular detail of a campaign when selecting audiences and deciding how much they’re prepared to pay to show advertising to those consumer segments.

This is quite a time-consuming task that needs to be continually monitored and updated; otherwise, marketers risk wasting money on creating irrelevant experiences.

Artificial intelligence is much more efficient and effective when it comes to identifying strong and weak signals for targeting advertising, such as characteristics and patterns of behaviour. Using AI, the model can constantly update itself as new behaviours emerge, enabling it to be more responsive to change than relying on people alone.

This becomes even more important when marketers are relying on context and probabilistic data rather than cookies.

‘‘The idea is not for artificial intelligence to completely remove people from the process,’’ says Konrad Feldman – co-founder and chief executive of global advertising technology company Quantcast.

Instead, the use of artificial intelligence frees marketing people from the more mundane number-crunching tasks, so they can add greater value elsewhere.

‘‘Google’s announcement about deprecating third-party cookies is focusing the minds of the industry,’’ Feldman says. ‘‘We’re moving to a privacy-first, consent-first world and that is a good thing.’’

Audience Planner is Quantcast’s live consumer insights product and part of the Quantcast Platform. It aggregates data from more than 100 million web destinations and grants advertisers and publishers a more comprehensive view of how consumers spend most of their time online, allowing them to more effectively allocate budget.

Quantcast’s artificial intelligence engine, Ara, then uses this insight as a blueprint to identify, target and grow a much larger lookalike audience across all corners of the internet.

Traditionally, marketers were required to specify in fine detail all of the inputs that should be followed when executing a campaign. AI allows marketers to specify the outcomes they require and the constraints for their strategy, after which the AI finds and continually modifies the inputs.

Artificial intelligence is not designed to take over and push marketers to the sidelines, Feldman says. Instead, it’s a powerful tool that reduces the amount of time marketers need to spend managing the system – freeing them up to do more marketing.

‘‘It also reduces barriers to entry, as programmatic advertising has been difficult to buy,’’ he says. ‘‘We have simplified this; so you don’t need an army of specialists with deep technical expertise.

‘‘We think that marketers should spend more time marketing and less time managing the intricacies of martech.’’