As the digital advertising industry continues on the long road to third-party cookie deprecation, marketers are having to increasingly look at how their available datasets can continue to help them to produce effective advertising, writes Mark Dunckley, Senior Sales Director at Scibids UK.
First-party and consented third-party datasets have a significant role to play in the cookieless future, particularly in optimising their media buying. However, activating this data can often be challenging. Marketers are faced with a plethora of obstacles, meaning their intention of utilising these datasets is often not reflected in what is actually activated within demand-side platforms (DSPs). As a result, only 52% of brand marketers believe they have managed to make full use of the customisation options offered by DSPs.
This is a cause for concern for marketers because, with media budgets facing more scrutiny than ever before, failing to unlock the true value of their data will leave them falling well short of achieving the strong results their organisations are demanding of them. Brands simply cannot afford to lose the value of the data that might remain locked up across their businesses, and need to find ways to make these disparate and siloed data sources interoperable. This will drive long-term, efficient and effective media buying.
Thanks to the plethora of systems that marketers must contend with today, data has probably never been utilised in a more inefficient manner. The result is that most marketers are not even scratching the surface of the value their data should be driving.
Marketers have to manage and navigate data from sources such as their customer relationship management (CRM) system, their customer data platform (CDP), their brand safety partners, and their attribution partners. They want to be able to utilise all of that data in their media buying, but the idea of doing that in a simple, efficient, and streamlined manner is almost impossible given the constraints of their in-house resources.
Brands and agencies need to find solutions that provide the ability to match different datasets and marry them with the DSP’s activation-level data, enabling marketers to optimise their campaigns toward real business outcomes, rather than proxies such as clicks or views. The solution to achieving this lies in the power of artificial intelligence (AI).
Every campaign run by an advertiser generates huge amounts of data. These datasets are often too large and too complex for any human being to comprehend and start to find patterns within. So, often, they will only look at a few key data points. These data points could be as simple as “publisher X is good/bad” or “channel Y is good/bad,” and assumptions are made from there. That’s how most optimisation is done in the programmatic ecosystem.
In the past, it was a lot easier, because there were far fewer data vendors and measurement providers within the ecosystem. Marketers would optimise toward the numbers in the DSP and make decisions based on what they saw there. Today, there are now so many different solutions, systems, and channels that it’s physically impossible for a human to make those decisions credibly, let alone in real-time. It’s inaccurate, inefficient, and it takes too much time. And by the time optimisations have been made, things have already moved on, so in many cases, these optimisations are often incorrect or, at best, imperfect.
This has helped to give rise to the wider use of AI within programmatic. Rather than a nice-to-have, it’s become an absolute necessity.
AI technology is capable of making precise decisions based on a level of data that no human could have comprehension of. It can act as an integrator, bringing together disparate datasets and finding patterns between them to make optimisations in real-time. Moreover, these learnings being applied in real-time means that a campaign is always working toward delivering the best possible outcomes for a business, rather than only being applied to future campaigns.
This doesn’t mean that the machines are taking over. They are instead removing the laborious, time-consuming, thankless tasks, and freeing up teams to do better work. Opening up the door for them to produce more value-driving, strategic, and creative projects – which previously may have been sacrificed in favour of data and technology.
If businesses can use technology for good, freeing up time and resources to do more creative, strategic work, everyone wins.
Finding the right solution
Having identified this, agencies have been exploring the possibility of producing their own AI solutions. They know that with all the various data points that their clients have, there is no way for them to utilise all of it within marketing activity. The idea of doing that in an efficient, streamlined, simple manner that can be replicated is almost impossible. Especially when any data that is used is heavily watered down by the time it gets to the point of activation.
However, within many agencies, there is often a lack of time and resources available to produce the sophisticated solutions they need to activate AI-powered media buying at scale. So, identifying external integrators is crucial in ensuring they are not left behind.
The once basic mechanism of running advertising has become quite challenging to deliver now, and that’s where AI will have a massive impact. It’s become too difficult and complex for a human to make the decisions that will impact campaigns in the best possible way.
It’s not that the existing solutions, companies, and people working within the ecosystem are doing a bad job – they will all continue to have a valuable role to play. It’s just become so complex, and AI is here to help run it all.
For media buying to work at its most efficient and effective, all data sources have to be working in-sync. And the only way to make this a reality is for agencies and brands to begin implementing sophisticated technology that can begin to bring this scatter graph of data together.