The past few years may seem to have been defined by artificial intelligence’s entrance into the mainstream, but the reality is that the technology has been around for decades.
This is especially true in the world of digital advertising, in which AI-based and machine learning tools have long been implemented to automate a wide variety of processes, including the optimisation of media buying.
With humans incapable of analysis at the scale and speed necessary to power quality media outcomes effectively, the algorithms have been trusted for years to accelerate efficient media buying through demand-side platforms (DSPs).
This has fostered the idea that AI for media buying, as well as digital advertising more broadly, is a ‘black box.’ A system where the inputs and operations aren’t visible to the user. In the case of many ‘off-the-shelf’ algorithms available through DSPs, this is true. Often these types of algorithms also pursue metrics that are proxies of business objectives.
Many advertisers may be satisified with this trade-off of control for performance under the assumption that there needs to be significant expertise in-house to finely hone AI’s capabilities. However, thanks to the pace of technological development, this assumption is today incorrect.
AI is powering a greater degree of performance in media buying, with customisable algorithms increasingly available that offer advertisers heightened levels of control, transparency and precision over their media outcomes. Enabling advertisers to optimise their media buying to their business-specific KPIs, these algorithms are helping to shed some of the mystique behind AI for advertising and usher in a new golden age of efficient and quality media.
Optimising toward what matters
Advertisers are well aware that they should be creating bidding strategies that align with real business outcomes. However, the fragmentation of the media ecosystem combined with the amount of data that requires analysis means that many do not have the resources to execute these strategies manually. This led to the rise of ‘off-the-shelf’ algorithms provided through the DSPs, which were sufficient for a time but offered little in the way of optimisation.
This need for human control was recognised in recent years by the major DSPs, which opened up their APIs to enable brands to bring their own plug-in AI to the bidding process. The result is that marketers today have access to the most advanced capabilities of the DSP in the form of custom bidding.
Custom bidding algorithms enable advertisers to manage their media plans efficiently and effectively, allowing them to make optimisations that are specific to the business. Critical components of the media plan – including quality CPMs informed by attention, viewability, frequency, budget, and brand safety etc., which are unavailable through the ‘off-the-shelf’ algorithms – can be managed at scale. The AI also progressively learns, which maximises efficiency for advertisers.
These types of algorithms also make use of non-user-specific signals and semantic data, as well as enabling marketers to bring their own first-party data. This is of great importance in this era of privacy-friendly ad decisioning.
For media buying, the technology means marketers are no longer forced to sacrifice control in favour of performance, or vice versa. Customisable algorithms open the door for marketers to find a beneficial balance between human oversight and machine-led efficiency, unlocking the full potential of their data-driven marketing campaigns.
AI’s promise and perils
This is just one of many areas that AI technologies are transforming digital advertising. With increased efficiency, dynamic creative freedom, and superior campaign optimism, there are plenty of opportunities available from leveraging AI technologies, but it is important for marketers to remember the perils. For example, the rise in sophistication of ad fraud schemes, the proliferation of made-for-advertising sites and the spread of mis- and disinformation (deepfakes for example) have boomed in part due to the greater accessibility of AI.
Marketers need to actively engage with the technology to understand what AI means for their media and how to best leverage it. If these efforts and education continue, there should no longer be concerns over the so-called “black box” nature of the technology. Instead, AI is giving back control to marketers at a time when they need it most, helping to drive toward better transparency and superior business outcomes.
About the Author
Nick Reid is SVP and Managing Director EMEA at DoubleVerify. DoubleVerify is a leading software platform for digital media measurement and analytics. Our mission is to make the digital advertising ecosystem stronger, safer and more secure, thereby preserving the fair value exchange between buyers and sellers of digital media. Hundreds of Fortune 500 advertisers employ our unbiased data and analytics to drive campaign quality and effectiveness, and to maximize return on their digital advertising investments – globally.
Featured image: Adobe
