Tuesday evening saw another fascinating Performance Firestarters, this time on the Future of Programmatic, which has obviously had a not insignificant impact on media over the past few years and will no doubt play an even more significant role in its future. For this event, we wanted to understand more about where programmatic is headed and as always with Firestarters we had three exceptional speakers to help us navigate this question.
Graham Field, Head of Programmatic at MediaCom, began with some context setting, defining programmatic from a media perspective as: 'The automated method of deploying media where data is utilised, to optimise delivery at an event level', and describing four pillars critical to the quality of deployment - technology, inventory, data and expertise (on this latter one he made an interesting point about the lack of skills currently, client and agency side, in the area - good people are hard to come by).
He then set out a maturity timeline for programmatic starting with digital display, which is well established and more embedded in client understanding, then mobile, which is commonplace but still with teething troubles (like tracking capability), and sonic/audio which is evolving due to the rapid adoption of music streaming. His maturity model then broadened out to include out-of-home which is emerging at pace (trading desks for OOH will soon be within agencies), but brings new challenges such as the metrics we use (e.g. how do we count a OOH 'impression'). Publishing is adopting its own form of programmatic for the sale of print inventory, but the scale opportunity is of-course TV. Graham made a good point here about the rationale and incentives for change - as long as TV is inflationary and doing well, there is perhaps little incentive for change but the pressure to trade more programmatically will increase and of-course TV budgets are huge.
Graham then talked about data ('talked about a lot more than it is used') and some tricky questions around use of client first party data (mapping it to a 'usable unit' to enable better tracking and deployment) and what this could mean for agency/client relations. And he finished off by emphasising the importance of trust in the whole ecosystem - how they work with media owner partners who are attempting to monetise quality content but in the context of a lot of middle men taking money out of the market for example, and why we should never forget the human touch to how we deploy programmatic.
David Carr, Strategy Director at DigitasLBi, followed with a talk that began by framing how some people see programmatic as like Soylent Green (like an apocalyptic food of the future). This is not helped by the language and plethora of acronyms that surround it, or the 'plumbing diagram' slides that are typically used to describe it in talks. Instead, he said, we should think of programmatic as personalisation at scale.
So it's about taking identity with context to enable relevance, often combining first and third party data to understand the 'job to be done'. Deployment is currently too focused on playing the numbers game, and not enough on creative optimisation - we see the same examples of personalisation being used - romeo reboot, share-a-coke, Diesel decoded, but there is greater opportunity to shape and personalise content programmatically. A bit like this Cook Along Kitchen Experience from the BBC R & D unit which is about planning an entire production around flexible, customisable media. So David's challenge was to think beyond advertising and think about utility, and how the relevance that is oft touted in the context of targeted messaging might be used to create better, easier customer experiences, and enhanced systems that can self-improve using data.
Our last speaker, Julie Jeancolas, the Head of Agencies (Media Buying Solutions) at Google, began by talking about how difficult it is for anyone to predict the future (like Gartner predicting in 2010 that Symbian would be the largest operating system in the world) but also how the growth of programmatic has been one of those hockey-stick growth patterns, now predicted to become 16% of all UK media spend in 2017. She had a nice take on the key drivers of that growth: reduced cost and risk (the ability to automate, test and learn, with little up-front commitment); greater transparency (better brand safety controls, viewability and attribution); and improved quality (of data, inventory and ad formats). But, she said, the future will need all three of these things to work well together, a harmony between people, data and tech, and how the opportunity comes to life when you apply the data in smarter ways (she used KeepTruckin as an example, which I hadn't seen before - a service that dramatically increases efficiency in supply chain through automation of manual processes and real-time tracking of vehicles), which will become truly interesting in the age of AI. Greater application of AI in programmatic will enable myriad combinations of different data points which, when combined with a unified ad tech platform, cross-team data sharing (something that already doesn't happen enough) can drive real-time personalisation at scale, leading to the inclusion of a greater number of signals and new possibilities.
In the panel discussion afterwards there was a lot of focus on the bigger picture of what this could mean for the structure and working of agencies, and where clients are at. It was all an interesting counterpoint to last week's Firestarters where there was more talk championing a move back from ultra-efficient targeting. The speakers in this session were keen to point out that personalisation and programmatic does not always have to be about specific targeting of messaging, it can also be about utility and about discovery. My thanks to our excellent speakers, to Google as always for hosting, and to Scriberia who as usual did a great job of visualising all the talks - you can see the full visualisation here.