“Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Hagen Wenzek, founder of Freestyle Consulting.
The question has been around for years. Why don’t we see an improvement in price performance of digital media trades even though we have more electronic systems to enable interactions between trading desks and publishers?
Too often, media-buying teams are either not using the systems or not using them as planned. Both are bad, but the latter is worse. In a media plan the properties and audiences to be reached through digital display are specified and sent to the trading desk for execution. But instead of sending the whole budget to reach an audience via the trading desk, teams might split it up or engage directly with one or more exchanges, potentially causing one to bid against oneself and drive up prices for no good reason.
Many options exist for overcoming this type of noncompliance, but there is one that few discuss: automation. We could automate the problem away.
To solve this, we could make systems even better so practitioners like using them. Or we could implement better monitoring, complemented by education or enforcement, but that would take time and has been tried before, without much success.
And then there is automation.
The Case For Automation
Automation is done for many reasons. There is, for example, always a gain in efficiency. Automation reduces the number of people needed for a task, and as cloud computing and software as a service (SaaS) drive down IT costs, the advantage over HR costs continues to rise.
There is also the challenge to cope with increasing speed and complexity. Applications to optimize cross-channel media plans are already widely used, but the execution of that plan might be too complex or needs to happen too fast to be done by hand.
Last but not least, automation can ensure compliance. Coding a process into software removes it from the hands of practitioners altogether. That means the process will be executed exactly as designed since nobody can interfere anymore, except when planned for.
Yet with each alternative approach to automation there are multiple hurdles to overcome, each of which makes it a hard thing to do. Not everybody is up for that, especially if maintaining the status quo is so much easier.
How To Automate
One way to automate is by using tools that take over various activities within a process. If media-spending data is fed into a business intelligence system like Datorama, nobody needs to touch the data to transform, clean and visualize it. Then the account strategy team can execute the strategy planning process more rapidly or with greater precision. That is probably the easiest way to embrace automation as it is just a transition from using other tools, such as Excel.
Buying a system or using SaaS that runs a whole process is another obvious way. Somebody else defined a best practice for a process and put it into software. What is left is to enter parameters, provide the input and a transaction is executed automatically, like buying social media ads via UNIFIED. On an even more generic level, leveraging real-time bidding and other programmatic buying options are nothing else than an automation of various processes.
Lastly, you can use a workflow-management system to implement your custom process and run that (semi) autonomously. Instead of planning a digital campaign and sending out requests for proposals (RFP) by hand, Mediaocean’s Prisma product allows the plan to drive the connected workflow to automatically issue and follow up on the RFP. Or you could use a generic workflow system that can capture any process and deploy it as an app via Force.com.
Beware Of Conflicts
Each of these alternatives will help make things faster, simpler and more efficient and compliant. However, conflicts of interest often prevent any real progress, especially when the goals call for more efficiency, as well as reductions in time and complexity.
Most agency contracts today still calculate compensation based on a cost-plus basis, so the more people and time it takes, the more can be charged. And the more complex the job, the less inclined the marketer will be to put pressure on costs, let alone try and do it themselves. Agencies like IPG Mediabrands have tried to shift the compensation model to pay for performance, but it is an uphill battle.
Achieving compliance, however, is different. It is nothing but good management. You make a plan and manage its execution. If automation ensures the plan gets executed as designed, nobody can really object. Furthermore, you can already achieve benefits against that goal of compliance before the workflow is fully automated, because you can temporarily deploy resources to do re-engineered process steps manually and remove them when the software is finished, in the same vein as a Mechanical Turk.
Interest in programmatic trading has soared in recent years. Plenty of tools and systems have become commercially available, and uncounted projects to implement them have started, some with excellent results.
But progress is slow. Just because the systems exist doesn’t mean they are broadly used. In the early 2000s, it was the same situation in supply chain management. Back then enforcing process compliance through e-procurement systems transformed how purchasing departments worked. When compliance was made the objective, everything else materialized. Costs decreased dramatically and there was no turning back.
The same is true today in marketing. There is no turning back.