The 3 Stages of Dysfunction for “Programmatic” Advertising Organizations

Does your company have an Internet expert? Let’s hope not. There was a time, not too long ago, when every company that “got it” had an Internet expert. If this sounds ridiculous now, ask yourself this: Is there someone in your digital advertising organization exclusively responsible for programmatic?

Yes? Uh oh. You are in stage two of the three stages of dysfunction for data-dependent organizations. Here is each stage in all its gory detail:

Stage 1. Someone has the great idea that: “We should really collect all our data. There has to be valuable stuff in there.”

Action: Call IT. Hopefully, you are not reading this on your abacus, and you have passed this stage.

Stage 2. “Wow, we sure collect a ton of data. But what does it all mean? We need someone to figure that out.”

Action: Build a data science team. Maybe call it your “programmatic team.” Hiring someone (or even a team) to handle “programmatic” is not a sign that your digital advertising business is going to succeed. In fact, it means you are about to go to stage three.

Stage 3. “Hmm. My email inbox is filling up with reports from my data science folks. If I wasn’t so busy trying to help my clients with their real-world problems, I would totally read these. What do I do now?”

Action: This is when it gets interesting. There are two wrong moves you can take – and one right answer. We will start with the wrong answers because schadenfreude.

Idea A: “Let’s turn that programmatic team into account managers. This way they can apply all that data to help each client.” This is a great idea until your data science team is unhappy because they did not sign up to deal with clients. Also, clients will not love losing their favorite account manager. (You know, the media expert.)

Idea B: “Let’s train our account managers to be data scientists.” This is a great idea until most of your account management team is unhappy because they signed up to help clients, not iterate on algorithms.

So what do you do? There are clearly insights contained within all this data you have collected. But those insights aren’t impacting client results very often.

The Answer: Embedded Analyst

Keep the data science team. Keep the account management team. But create a new position: a data analyst that is embedded into your client team. This person is well-trained in data science, but is also experienced in media and account management. Perhaps it is one of the few people who wasn’t unhappy when you tried Ideas A and B.

The embedded analyst uses her knowledge of specific client campaigns to give feedback to the data scientists so they can improve their algorithm. She can also review their reports and understand where there are gaps and ambiguities in the data because she knows the client. Account knowledge is crucial to avoid dead-end analysis and wheel-spinning.

You can’t silo your programmatic team away from your client-facing team. But your data science team does not need to know the idiosyncrasies of each client’s account either. Someone needs to know when something doesn’t look right with the client data or the algorithm output. This person is your embedded analyst, who will soon become the prototype of your most important future hires.

We Must Kill Programmatic to Save It

Programmatic is just a fancy term for data-driven decision-making at scale. Soon, the phrase “programmatic” will seem as banal as the phrase “information superhighway.”

We no longer hire Internet experts because everyone in digital advertising is trained on a baseline level of fluency with the technology. Time to do the same with programmatic.

As our business is becoming increasingly data-driven, our teams will be staffed with folks who have basic data science skills. Technical experts will always be needed, but some of them must be fully integrated into the media side. We need to demystify programmatic and stop pretending it’s a special calling for a small group of high priests.

It starts by embedding programmatic experts that can sit within your client team and connect data to real-world client needs. Take this step, and many the insights contained in your data will actually impact client campaign performance. Just like you always knew it would.

Related reading

prog
hillary-clinton-text-message-signup
screen-shot-2016-09-13-at-10-20-04
mike-andrews
<