Digital Essays 2017: Why big data needs big AI – by Charli Ursell, senior director of digital planning, PHD

We’re experiencing a period of rapid transformation right now, with technology becoming increasingly (and almost invisibly) a part of us. This merge not only closes the gap between humans and machines, but it also generates vast amounts of data with every action we take.

With up to five tech layers on every media plan, marketers can tap into millions of data points to inform planning approaches, buying and optimisation strategies, thus enabling them to measure outcomes with greater accountabil- ity. But let’s take a step back for a minute. When we consider how extensive this data pool has become – together with how we have decided to use it – how much of its full value is actually being realised?

We’ve come a long way in the past decade, with a constant stream of new tech and the connected buzzwords, but artificial intelligence (AI) is a different proposition altogether. Having

talked about it for years and with a few products now available, you could be forgiven for thinking you’ve seen and done AI. Yet there is so much more to it and we’ve barely scratched the surface of its real potential.

The arrival of big data took the agency world by storm. We moved from demographic targeting and media results (remember CTRs?) to being able to implement specific targeting and serve specific ads, for example, to people already in the purchase mindset. This transition happened in the space of just a few years and, following initial data-driven success stories, most clients now expect fully-fledged data-led solutions as the norm rather than the exception.

However, the traditional agency structure has begun to crack under the volume of data and the enormity of the analysis required to make full use of it. It has become increasingly clear that all this data was never meant for humans but for machine learning (ML) instead.

Even though we are expanding our team of data specialists, it’s estimated we extract little more than 10 per cent of the full value of the data available to us. We need to rely on automated systems powered by a combination of AI and ML to extend our reach, push further and dig deeper. This is an essential step forward to add even more value to our clients’ bottom line.

Today, you may be focusing on your video ad’s viewability score or discussing the need for new creative variations, but consider what you could get from working with sharper intelligence. You could be investing that same amount improving your site’s speed by a few seconds because you’d immediately know that it would double your conversions and increase your profitability. This isn’t science fiction and we’re achieving just that today by using machine learning and automation to analyse greater volumes of data and generate more meaningful and usable insights.

Thanks to these early wins, we’ve already outlined PHD’s long-term automation roadmap in the region. We’ve also evolved the capabilities of our teams, creating tailored partnerships between human intelligence and technological power for each client and objective. This merge will lead to AI featuring more and more prominently in the way we operate.

While these changes are necessary to our continued transformation, there is still a vital element that often gets overlooked in all of this: the consumer. As is so often the case with data, we can become distracted with the many learnings and opportunities we glean from it, and lose sight of our main purpose: to provide value to the end consumer.

What’s more, consumers have started to realise that all the information they provide freely, such as names, email addresses or credit card details, is worth more than just the convenience it offers them. You only need to look at the rise in ad-blocking tech adoption across the globe to understand this shift. Consumers increasingly want to decide who has earned the right to speak to them and when, which will technically mean that they will regain the ownership of data about them from advertisers and media.

With the implementation of protection legislation such as GDPR (general data protection legislation) next year, consumers will become even more cautious and look for a fair value exchange in the data they provide. We’ll be expected to comply with strict new guidelines in how we obtain and use this data, offering more transparency and a chance to build and retain their loyalty in the process.

As consumers become more selective, they will also learn to rely on automation and ML in the form of virtual personal assistants (VPAs). Instead of giving their data to different intermediaries, consumers will simply feed their VPA their preferences, which, over time, will learn and anticipate what their user wants. In this context, advertising will be directed less to humans and more to machines acting as their gatekeepers.

While we may not be there yet, it’s clear we as an industry need to consider how we access, analyse and use this data in a more effective way. As an agency, we are already exploring ways to facilitate a fair exchange for the benefit of both brands and consumers. We’ve only just begun to unleash the potential of AI and there’s much more to come. Success depends on how we manage this transition.