By Romain Rouquié, UK Head of Customer Engagement, Scibids
With the upcoming deprecation of third-party cookies and changes to other third party identifiers, advertisers will need to adapt and find new types of automation to help process the large but disparate pools of data they have in their hands.
Artificial Intelligence (AI), particularly within the ad decisioning phase, is able to perform certain tasks better than any human can. For example, it can crunch huge datasets all day, every day, and come up with the best approach based on that data.
It can push new models into the Demand Side Platform (DSP), affect how the DSP bids, improve campaign performance, and enable advertisers to start ingesting their first-party data or different sources of truth into their activity.
AI complements human input
Within the decisioning phase, AI is able to enhance what smart trading teams are doing, and complement the human skillset, which should still be leading on campaign planning, strategy, setup, and ideation.
Powerful AI and smart trading teams provide an ideal combination for delivering on business outcomes for brands.
Without the use of AI, advertisers and agencies can struggle to access the most advanced features that DSPs have available, because even the smartest data scientists can take a long time to create one algorithm for one campaign.
On the other hand, sophisticated AI can provide a turnkey solution that works across multiple DSPs, and be scaled across multiple campaigns in different territories for different KPIs.
Humans are a key part of the process
Nonetheless, the inclusion of AI shouldn’t take away from the importance of human input within the advertising process. In fact, it’s just as important for smart trading teams to continue being involved.
If a campaign setup is poor, and things haven’t been designed in the right way from the initial human input, then the AI is going to go in the wrong direction and fail to deliver on what you require it to. The AI should exist to provide more options, changing the way that traders think about setting up campaigns.
A lot of campaigns tend to restrict certain criteria in order to drill down into what the advertiser is looking to achieve – whether that’s the audience, the inventory, or the setup. AI, however, should be treated as the opposite, preferring to have the widest possible playing field to work within. Rules can still be applied around parameters such as brand safety and viewability, but the role of AI, in general, should be to scale campaigns, utilising its ability to search into areas that a manual trading approach wouldn’t allow.
Advertisers seeking a more algorithmic approach to their campaigns should always be looking to widen their scope and do more.
Proving the worth of AI
Although AI can provide the answers that advertisers are looking for, the transition toward using more automation takes a change in mindset, and an understanding of what the AI is there to do and how it can do it.
As the AI is able to bid the most optimal price for each impression, it offers the possibility to find pockets of performance outside of your optimised set up. To take advantage of this capability, you should start by removing all restrictions put in place solely for optimization purposes. This can mean allowing to target new devices, demographics, audiences, or URLs outside of your inclusion list. Every element previously excluded because poorly performing can be put at the service of the AI to generate better outcomes.
A/B testing can validate that the AI is driving better performance than the current benchmark, even within the current campaign setup. However, a smarter way to test the AI is to compare your traditional approach against an enhanced set up ensuring a wider pool of bid opportunities.
Step by step, you’ll start to see the value of utilising AI that way and start to become more comfortable with the idea of loosening the restrictions you have in place for your campaigns.
Could you take away some of the restrictions that are still in place to widen the pool and scale the campaign, while keeping the KPIs consistent? Can you ingest first-party data or third-party sources of truth into the models to enrich them and optimise them directly toward outcomes?
As a trader, powerful AI provides you with more options. You’re no longer restricted by the manual approach, and can really start thinking differently. And that’s why AI is perfect for smart traders who like to think outside the box.
*Scibids is a client of Bluestripe Communications, owned by Bluestripe Group, which owns NDA.
