This promoted content is produced by a publishing partner of Open Mic. A paid-for membership product for partners of The Drum to self-publish their news, opinions and insights on thedrum.com - Find out more
Why mastering smart machines could unlock the next programmatic evolution
January 17, 2022
While researchers might be feeling daunted by the rapid progress of artificial intelligence (AI), marketers have been long aware of the advancements in this area. It’s some time since smart tools began moving out of labs and into advertising stacks: so much so that machine learning is already sliding down the hype curve. In parallel with rapidly growing adoption, the industry has also seen ongoing uncertainty about the long-term impacts of rising AI use for brands.
Concerns don't just stem from flat resistance to automation. Programmatic is the prime example of an area that has flourished with assistance from advanced technologies, now attracting over $150 billion in global ad spend, and stands to benefit even further. But despite awareness of the potential for sharper real-time ad matching and greater relevance, recent studies show 75% of those piloting AI have doubts about whether tools can be trusted.
Overall, such anxiety stems from a classic issue: control. While marketers are keen to seize AI’s efficiency and scale-boosting abilities, they remain fearful of handing over too much power. This worry is largely unfounded — but understanding why calls for a closer look at what today’s solutions offer and how they can fuel well-managed programmatic evolution.
Keeping optimization within tight parameters
Although getting closer, hyper-sophisticated AI isn’t here yet. Development is a significant way off of achieving the general intelligence needed to replicate or surpass human cognition and decision-making; with the majority of researchers agree this milestone is likely to be reached in the next century, but others estimate it could take several hundred years. At present, initiatives widely used in advertising are more service droid than Ex Machina.
Applications typically function within narrow parameters. Producing outputs determined by the data feeding them, as well as particular objectives. All too often, their advantages have been overshadowed by speculation that dependence on variable input is among the main reasons behind continuing mistrust in AI. But it’s also vital to recognize the potential current operating dynamics provide for marketers to retain a firm grip on programmatic efforts.
Harnessing built-in controls, they can keep campaigns on track by setting specific rules for algorithms to follow. In fact, the concept of closely managed optimization is already gaining traction on multiple fronts, including instructing systems to direct bids at the audiences and inventory likely to fuel performance against core goals. While implementations so far have tended to focus on metrics that don’t give much insight into true ad effectiveness or drive informed choices — such as click-through rates — they are an encouraging start.
Leveraging emerging tools will help marketers take this approach further: adjusting media buying in line with outcome-based measures of precise impact on areas such as awareness and brand favourability. For instance, platforms supported by reinforced learning are making it possible to pre-emptively pinpoint which impressions will offer the best chance of delivering goal actions in different contexts, be that brochure downloads or social media interaction.
Armed with this ability, marketers can tell algorithms to only place bids in situations where users have a high probability of taking specified steps, ensuring AI plays by their preferred rules and maximising results at the same time.
Blending smart analysis and robust guarantees
Used alongside other control levers, these refined yet adaptable dials have considerable scope to help enhance advertising success while maintaining clarity on a persistent basis. And top of the list is applying curated mechanisms that give buyers and sellers alike more jurisdiction over transactions, in terms of prices, terms, and inventory selection.
Trading methods such as programmatic guaranteed are becoming an increasingly popular option for marketers keen to ensure total transparency about deals and trading partners, especially after ISBA findings about the “unknown delta”. By combining tightly coordinated deals with AI optimization and assessment, they can ensure total visibility into every aspect performance throughout campaigns, from buying to measurement.
Tapping AI as a means of pinpointing which media opportunities will generate the strongest ROI, ongoing measurement against outcome-centric metrics can deliver a detailed view of whether deals — and algorithms — are yielding desired results. In adding to giving marketers accurate insight they can use to quantify deal value and forge productive future partnerships, this will equip them with the granular performance data required enable smarter in-the-moment decisions about which ads to serve for the highest outcomes.
Moreover, they can also build a constantly expanding pool of knowledge about what works for target audiences that allows them to constantly improve ad success: translating data about unique preferences and performance into bespoke predictive models, which steer targeting and delivery towards ever-greater relevance and profitability.
As AI develops the power to lighten the load of more everyday tasks, apprehension around ethics and machine-led choices is inevitable. Equally, however, its ability to further streamline optimization brings tantalizing prospects for better automated targeting and performance against tangible campaign objectives. To push ahead with the next programmatic evolution, marketers must start getting to grips with the dials and capabilities that will allow them to guarantee high impact, without scarifying clarity or control.