Ad Tech Needs to Learn Plain English

So many technology startups focus on a thin slice of the value chain. This is interesting, yes, and perhaps even lift producing. But it is not significant in and of itself to move the needle for a brand. One of the reasons MediaMath has scaled to the level we have is because see one of our key roles as systems integrator, aligning with multiple specific technology providers based on which make the most sense based on given brand’s objectives. So, that means multiple attribution partners, multiple exchanges, multiple rich media companies, etc. The second key role is optimizing against the super-set of media, data sources, and tools out there to make the whole greater than the sum of the parts.

Our industry seems like a government agency to some buyers, with all our acronyms and the byzantine fragmentation of our world. How many Digiday readers can even explain what they do to someone not in the industry? That’s a real problem. We’re addressing this by focusing on both simplicity and transparency at the same time. We found it important that these two move in lockstep. Simplification is important, because marketing successfully in digital requires the careful coordination of domain experts, quants and statisticians, marketers, and technologists. Getting these four groups to show up in one place, let alone work in one organization and align around a vision is hard. So you need to simply and codify the complexity required to do marketing well by creating abstraction layers that focus less on that what and more on the why – delivering better results in the process. However, you can’t do this in a black box. Sophisticated marketers need to have transparency to the inputs and outputs of their marketing. They need to have control and understanding of their supply chain. They need to build a marketing system that allows them the ability to test and learn. Some use it, most are just glad to work with a partner that provides full transparency, and rely on the speed of execution that simplicity part of the equation provides.

There is a lot of teeth gnashing about the complexity of digital marketing and the alphabet soup of logos and acronyms. I think it’s overblown. And let’s not forget, despite our foibles, our industry continues to grow pretty healthily. Yes, it’s hard. Yes, you need folks who are expert at assembling a complex set of best-in-class and perhaps niche specialist resources to get the job done. But lots of industries have this problem. Go take a walk through your kitchen and log all of the brands. Koehler and Viking and Meile and GE and KitchenAid and American Standard, etc. Thousands of logos get assembled every day. Plumbers, electricians, lighting designers, all conspire to work together to get thousands of homes built per day. Just get a good architect, which is the agency, and a good builder, which would be your ad tech provider, and the fact that there are lots of logos in your marketing structure will matter less.

Audience buying is a major step forward for this industry and enables brands to reach their customers and potential customers wherever they might be. It also allows buying platforms to drive much greater efficiency than the traditional media plan. Our approach is to bring in a number of best-in-class data providers to give our clients the greatest reach and match rates. Interoperability is key. But let’s not forget context. Context is integral to how audience segments are defined in the first place, of course. The right audience and the right context – together – make the best environment for a brand message. From a recent study we did on the billions of impressions we see flowing through our system every day, publisher mix  — i.e. the context in which an ad appears — was a primary predictive factor of campaign success.

Maybe there will be a perfect model that describes marketing effectiveness but it will take an infinitely long time to design. It will know what happens when your brand moves one shelf up in Walmart at the same time as digital marketing ramps up and TV GRPs are down, across all audiences. Absent this perfect model, we have abstractions and proxies. Some are really easy to track but not very good measures of intent – those are clicks. Some are easy to track, pretty weak indicators of intent, but are widely accepted which makes for an efficient planning and buying process —  those are the GRPs. Some are fairly easy to track, better indicators of value, but poorly adopted– that’s engagement. Others are reasonably hard to set up, very strong indicators of value, but are non-standard measures that require translation to other parts of ecosystem to act on. Marketing ROI requires client-specific set up, works really well, but when you want media partner 466 to do something different, you need to convert it into CPM. As to where the world goes, we see custom KPIs per brand assembled from a library of trackable metrics, where most brands start with mostly default settings and then get more sophisticated as they test. The brand’s DSP then translates these metrics into action across the supply chain. In other words, more Facebook sponsored news to these segments, less branded search to those, etc. It remains a long process.  But, inside of five years you’ll see more of the most sophisticated marketers with this system in place.