Matt Nash, UK Managing Director at Scibids examines the obstacles to AI’s integration from complexities surrounding the technology to lack of data science capabilities, and why it’s crucial marketers understand which AI to invest in if they want to deliver the most successful campaigns and maximise ROI.
Not so long ago, the mere mention of Artificial Intelligence (AI) sent chills up marketers’ spines with fear that this futuristic tech could displace jobs across the industry. Yet now, 42% of UK programmatic spend uses algorithmic decisioning or artificial intelligence.
The benefits it brings to digital marketing such as operational efficiencies, actionable insights, privacy-compliant targeting and crucially, revenue gains, have made AI a tech worth investing in but there are still significant barriers to greater adoption of AI.
So, what is holding the marketing industry back?
In short, as with any powerful new technology, there are always going to be initial reservations largely stemming from misconceptions. Indeed, AI faces bigger perception challenges than most tech concepts.
For a start, it is a long-standing source of creepy sci-fi plots in pop culture, from 2001: A Space Odyssey to the Terminator series. Then, as AI became more recently a real technology with increasing applications, those horror stories were replaced by a more grounded concern that the technology could start to replace humans in carrying out certain jobs.
Meanwhile, at an operational level, what we see from some business leaders is more of a resistance to disrupting their current marketing model. For some marketers, this means a fear that their business just isn’t big or sophisticated enough to incorporate AI into their campaigns, or that they lack the data science expertise to leverage it effectively. Others worry that it would require them to throw away everything they’re doing already and start from scratch.
Overcoming these attitudes is a matter for the vendors of AI technology to ensure all parts of the digital ecosystem have the education they need to feel confident in adopting AI in their programmatic campaigns. By proactively breaking down these misconceptions surrounding AI, we can dissolve these barriers and help brands achieve performance levels only possible with the technology.
Not all AI is alike, know the difference
Since AI has become a bit of a buzzword, thanks to a range of industries talking about it, it is understandable that some marketers might see it all as the same.
The reality, though, is that AI for marketing comes in many shapes and sizes. No two marketing strategies are identical, so we shouldn’t expect the same AI model to work for everyone. The technology has many different functions and also varies widely in sophistication. There are off the shelf products that may be suitable for some advertisers, but the more advanced, bespoke offerings are designed to work within a business’s existing stack, minimising upheaval while maximising KPIs.
Better targeting without cookies
When the industry finally says goodbye to cookies, customisable algorithms will play a vital role in targeting. Without access to traditional identity-based targeting, brands are going to need AI-powered methods in order to keep delivering their ads to the right audiences.
Most customisable algorithms put into demand-side platforms (DSPs), have been programmed manually by a human, and hit 1-2 scripts a week (or month). While these are good for learning, the algorithms that have been optimised by AI will be pushing scripts two or four times a day, which is necessary for achieving scale in this increasingly complex data ecosystem.
Indeed, DSPs are now able to draw from a wide range of data signals relating to the context of a given impression, which inform how valuable it is to an advertiser.
Because these signals don’t depend on the user’s identity, this approach to targeting marks a big step forward from cookies in terms of privacy.
The volume of data involved would be unmanageable for a human, whilst an AI is able to crunch huge datasets and come up with an answer in an instant. Where the ‘intelligence’ really comes into play is in the AI’s ability to learn as it goes, continually adjusting its algorithm as a campaign progresses. Rather than measuring success in terms of views and clicks, which correlate poorly with metrics like sales and brand consideration, a strong AI can be built around a brand’s own KPIs.
A brilliant new colleague
AI is sometimes perceived as a whole new way of doing things, and often as something that can only be leveraged by large companies or those with the data science know-how. But despite its impressive power, it is not about to replace humans, nor is it above our control.
Instead of viewing it as a synergistic upgrade on human intelligence, think of an AI as a member of your team, with very specific skills. While an AI can bring capabilities to the table that humans can’t match, we remain much better at tasks such as campaign creation, strategy and ideation. Moreover, there are highly sophisticated AI technologies out there that can be seamlessly integrated into your tech stack with little logistical disruption or necessary technological expertise.
As the capability of AI grows in the coming years, our jobs will look different, but marketers will find themselves more empowered, not less. Make no mistake, AI will be a vital part of your tool kit in the foreseeable future; and the sooner you start, the sooner you’ll see the benefits.
UK Managing Director