As the percentage of ad spend being bought algorithmically increases, and the continued integration of machine learning and AI capabilities into those buying platforms becomes more prevalent, much of the buying and planning work is perceived to be handed over to the machines. However, in reality, the requirement for deep digital skillsets has never been more important.
While the granular work, previously carried out by specialist resource, has seen a reduction, and as a result some agencies and brands move to multi-platform investment/biddable specialists to run the campaigns, we still see a distinct advantage in retaining deep channel specialism to plan and buy digital campaigns.
It is crucial to fully embrace the algorithms, leveraging their full potential for smart media buying. However, in doing so, human skill, guidance and insight has only increased in importance. Here are six reasons why.
1 – Feeding algorithm, the right inputs
Previously, we have discussed the importance of feeding the optimal learning data to the algorithmic. A true specialist can understand how the engine works, how the bidding strategy optimizes and what data best fits the audience definition to drive the outcome. Those with deep channel knowledge are also able to understand the data points that will be matched within that platform, i.e., how Meta will match to CDP, what in-platform identifiers it will match with, what that audience will look like, etc.
Beyond raw data, the experts can give the engines context and content that sit outside the reach of the algorithm. For example, feeding in weather patterns that affect buying rate, store location performance, relevant social trends or other environmental factors.
2 - Understanding the machine’s optimizations
With AI driving much of the optimization and able to leverage millions of proprietary signals beyond the grasp of human optimizations, it is still imperative that we understand what is going on under the hood. Only those experts with a strong understanding of the channel can decipher the increasingly diminishing performance data the platforms return. They can join the dots, bridge the gaps and sew together to generate actionable insights, course corrections and manual optimizations.
3 - Removing short term noise from optimizations
While the digital platforms move towards a start broad and let it optimise approach, there is still a necessity for rigorous campaign monitoring, particularly in performance campaigns. While this used to be a case of many small tweaks and lever pulls, the algorithmic buying capabilities mean this can be counterproductive – too much change for the algorithm to learn and function. Performance planning therefore becomes a case of understanding what is just short-term noise in a metric; a drop in CTR due to price pressure, an increase in cost due to competitor spend. Compared to more serious change in metric and needs to be changed manually; creative, bidding strategy, etc., the best insight to undertake this job is still generated by a platform specialist, who understand the signals and nuances to make these decisions.
4 - Layer tech and tools to create a competitive advantage
As more spend goes into the major algorithmic platforms, the more brands are competing with the same tools, on the same inventory, with similar data points. One way to create an advantage is layering tech and tools into the campaign from custom algorithms, AI creative, and unique data use cases, to bespoke marketplaces. Again, a true specialist can understand what tech and tools will give the advantage and which combination of unique solutions should be leverage for the campaign objective. These experts understand how the algorithmic works, and how best to hack it with proprietary solutions.
5 – Generating actionable insights post campaign
The nature of algorithmic buying means brands are receiving less and less of the traditional performance data they are accustomed to receiving back. Much of this data now sits in the black box, such as placement performance or creative asset metrics. Therefore, channel specialists are lent on to give context on results, translate the diminishing level of performance data into insights, and create learnings and future behaviors that can be fed back into the business. As the level of data becomes simpler, the method to understand and create insight becomes more complex and requires deeper skillsets.
6 - Flattening of the channel silos
Channel specialism isn’t going anywhere, it’s going everywhere. As the rise of algorithms breaks down the channel silos, we see that the skills previously needed in one channel increasingly become required across all channels. Search skills are becoming a core function within social, commerce knowledge is expected within social campaigns, retail specialists are brought into display buying, the list goes on. Even for a campaign running in just one channel, there is a requirement for multiple specialists to create digital experiences consumers are now expecting.
These six reasons are exactly why we retain deep channel expertise within our digital teams. The knowledge, skills and insights from these experts are becoming more important as we shift towards publisher-owned algorithms. Our ability to leverage these skills and integrate them with other specialities within our dentsu capabilities – Total Search, Total Social and Total Addressable and Commerce – means we can adapt to the requirements of the Algorithmic Era and give our brands the best possible solutions.
