This article originally appeared in the French publication Viuz from an interview held with MediaMath CMO Joanna O’Connell at Cannes Lions. It has been translated into English below.
What does MediaMath do and how do you differentiate from your competitors?
MediaMath was founded in 2007 and, since the beginning, our goal has been to revolutionise traditional marketing in a mobile and data-driven world. We provide marketers with the best tools to take advantage of this new world. Marketers should now be able to engage with their consumers wherever they are. Programmatic is process automation and real-time decisioning using data. Programmatic started with ad exchanges back in 2003, but, by nature, algorithms and Machine Learning can be applied to all marketing.
A key point in our differentiation is to build transparent products for our users. Transparency should be for prices, placements, value chain and data. Transparency is crucial as it enables marketer to learn, to be smarter. We think that black box solutions are not ideal for marketers, who need to learn day after day, to make the right decisions.
Another point of differentiation is that we are only buy side; our clients are buyers, only buyers. This is a key element for us. We are different from some of our competitors which serve both buyers’ and sellers’ goals. All our products and innovations serve buyers’ needs, which makes us more credible towards our clients and partners.
Fragmentation / consolidation: What are the market trends?
Fragmentation and consolidation will continue simultaneously. New start-ups and more traditional companies will diversify in this market. We live in an evolving world, and that’s particularly true for the world of marketing. There are now connected fridges, which send and receive data, which makes them media. Another example: TV. Addressable TV is becoming more and more popular, with applications such as Apple TV or Roku that connect TV and change the market. This creates complexity for marketers.
Another fragmentation aspect is about usage. In a fragmented and cross-channel world, cookies have become a problem. Tracking and understanding the user’s behaviour across devices has become very challenging. The ability to track both cookie and cookieless scenarios using both probabilistic and deterministic approaches is the way forward to get accuracy and scale in identity management.
Are marketers and agencies ready to embrace programmatic?
In the US market, at the beginning of programmatic, agencies were reigning and were the experts. Advertisers came with a brief and agencies took the decisions regarding technologies, publishers and optimisations. With programmatic development and its growing impact on the business, advertisers now have more power. Brands want to internalise more and more of this expertise.
This has many implications. Advertisers become more and more embedded, would like to become specialists and recruit teams. It’s particularly true for big advertisers, and I think they can achieve this. The questions is: To what extent do they want to internalise and with which level of complexity? The necessary condition, of course, is to invest in resources.
Between two extremes (complete internalisation and externalisation), there are many intermediary solutions. For instance, marketers can be involved in choosing their technology providers, but still have their campaigns managed by agencies. Agencies are continuing to develop programmatic excellence and will be key strategic consultants for brands.
What does mobile change? How do marketers tackle this new device?
There is important growth in mobile, as well as video, and there is huge potential. More than 50 percent of impressions we see in the world are mobile. Display impressions (without video) only stand at 38 percent.
Two elements are key. First, we need to respect mobile ad units and leverage their unique assets (ie: geo-location). Then, we should not forget user experience. The mobile strategy should be consistent with the overall marketing strategy. It’s not only about delivering the right message, at the right moment to the right user, but also with the right format. The mobile ad should not damage the user experience. Sometimes an ad has to be disruptive, but it should not intrude into the user’s life. Overall, mobile advertising is efficient, but there is still room for improvement, such as better integrated ad units, to be even more powerful.
How can Machine Learning algorithms enhance advertising?
At MediaMath, we have been developing our algorithm “The Brain,” which is based on Machine Learning, for several years. We have an entire team dedicated to its development. On one side, The Brain is dedicated to marketers’ goals, such as increasing ROI. On the other side, The Brain estimates the price of each impression based on the competitive dynamics of the marketplace as a whole. Both sides determine the best impressions to bid on, as well as what price to bid.
Our work on Machine Learning is central. In a complex environment, with real-time decisioning, it is necessary to understand what the right signals are and how valuable they are. Those algorithms need to be smarter over time and should be able to learn by themselves. One application is multi-touch attribution. Marketers shouldn’t use the “last-click” model anymore and should evolve to a more complex model enabling the understanding of the true impact of each channel.