How Universal McCann Does Data

By
  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Any media agency will tell you it's innovating hard around data in campaign planning and reporting. But what does that mean exactly?

For Universal McCann, the global media arm of McCann Worldgroup, today's challenge in analytics isn't methodological. It's all in the follow through – creating standardized measurement approaches and then getting insights into the hands of media planners in a predictable and repeatable way.

To meet that challenge, UM set up a unit called Decision Sciences, led by Managing Partner Michael Horn. Horn, whose background straddles agency (R/GA, OMD) and media (NPR), describes Decision Sciences as "incremental to the agency's research and advanced analytics functions."

He recently spoke with AdExchanger…

How do you operationalize analytics in the Decision Sciences team?

MICHAEL HORN: Every team optimizes, reports, and projects performance and benchmarks. The question is, how well and how efficiently do they do it? We found there was a tremendous lack of efficiency and training skills management in the junior levels.

If you look at the teams from media director to associate director on down, there's a lack of important technical skills and the translation of methodology to actual planning and optimization.

So part of that initiative and part of UM's transformation overall is following a financial services model - this idea that your ratio of portfolio managers to portfolio analysts needed to shift and we needed more analytically focused people in the trenches on the accounts.

These people are dedicated to their account. On Wells Fargo, for example, it's four analysts sitting alongside about 12 media planners.

Can you quantify the overall balance of analysts versus planners?

It varies account by account. We're about a year into the North American rollout now. We've got analysts in place on Wells Fargo and Charles Schwab, Johnson and Johnson, Verizon, and Chrysler.

Where we have them deployed depends a bit on the scope. The more digital and the more DR oriented, the more aggressive the optimization philosophy, the more likely you're going to find analysts there.

Wells Fargo is pretty high at a three to one ratio, but there are lots of places where it's higher. On Johnson and Johnson you have a team of four analysts who are just focusing brand by brand where the impact will be felt greatest.

Talk about talent. Do you interview lots of people?

My discipline is very different from advanced analytics or from research in that regard. There you want people with a background in stats or who are econometricians, have advanced degrees, and whatnot. The people on my team are media strategists at heart and they just pick up additional technical skills. We can train analysts out of a lot of different backgrounds. We look at the ability to tell stories with numbers as being the one thing you can't live without and virtually everything else can be trained.

Who are the most important internal constituents to tell the story to?

The ultimate audience has to be the client decision maker. While we're doing optimization on our end, that relationship doesn't work unless the client fully buys into the strategy and why the strategy is supported by the data.

We do a lot of work with dashboarding, high-impact visuals, and ways of telling a story using all this data.

How is your attention divided between creative and media?

Most of our focus is on being good partners for creative. We typically own the data set and then we're providing that data set to creative. It used to be very behaviorally focused. Creative optimization was click-through rate by creative concept or by offer. But now we're able to talk about visibility and audiences - which audiences respond to which messages.

Are vendor companies and their services arms an extension of the agency services layer?

I think that agencies are rightfully the trailblazers on customizing solutions for our clients. In some ways you've got some of the big technology providers drafting on us because we've already found out the optimal ways of customizing a vertical category or business challenge. The question is, can you build technology and automation around what those methodological best practices are?

There's definitely some give and take where we do client service really well. We can provide customized contextual solutions and insights while, at the same time, the technology players are doing technology well and trying to see how much they can edge into the insights piece.

Do you think marketers want Google's unified stack?

There are a whole bunch of tools and platforms - it's not just Google. There are lots of other software as a service or platform as a service companies out there who are tremendously empowering to small to mid-sized advertisers.

If you have a marketing budget in the six figures, you have many more options available to you now and much more robust methodologies and targeting. It's a golden time to be a smaller marketer.

The stack has less applicability for the Fortune 500 scale. On the one hand, it raises the lowest common denominator substantially because suddenly you've created a new floor where certain types of measurement and targeting become commodity services, which is great in terms of upleveling the conversation as a whole.

Some of the Fortune 500s are making independent decisions about their platforms, and agencies are obliged to use that platform. Is that a problem?

No. For my team, our job is to get actionable information in front of the client so that we can all make decisions together. The agency team can make decisions; we make the decision with the client.

There is a technology agnostic aspect to that. We'll use the best data and the best tools available. Sometimes clients have strong preferences and that's fine. It takes a long time to implement a good analytics platform. If you're talking about people that are looking at Coremetrics versus Omniture versus Webtrends, those are multi‑year investments.

We totally understand, whether it's an ad server or one of those big platforms, the need to commit and support our clients’ choices.

It gets more complicated when choosing where you're going to spend your media dollars. Clients can create relationships that are based on personal associations or some impressive-sounding technology, but it's got to perform.

Can brand building be done in programmatic media?

Of course it can be done. What I find amusing is when people say, "Well, in the programmatic space the units are so limiting." I'm like, "Well, there’s a 30-second spot. No one says you can't brand in a 30-second spot. It's all the same modular structure." It's all about how you target it.

If you can target based on brand-relevant attributes - for example, if you're targeting on peoples' values and their brand equity and lifestyle considerations - then you can do branding media. It's more about the targeting than it is about the units.

If there's no opposition between custom and datadriven ad buying, what's to prevent a creative agency from functioning as a media agency?

Nothing.

So what's the future for that?

It obviously puts pressure on the agency landscape. The way those roles are defined, if it's all going to be about delivering the best outcomes for clients, there will sometimes be blurred lines, just as in the offline world.

Now every agency's a digital agency. The creative/media distinction is ripe for some disruption, and the reality is, it's been disrupted all along. I think it comes down to clients and the way the clients want to work.

Sometimes clients want to have a single agency, sometimes they like having best in class across six or seven different agencies, and that creative and media distinction comes down to the customer's experience with the media. That is defined by what the creative is, by where you consume it and when you consume it. I don't think that consumers, in the end, distinguish between these things.

What else?

One thing I'm really excited about is how the IPG Media Lab has been doing testing with biometrics in a production-ready way. In terms of conversations we'll be having a year from now, I love the fact that things like eye tracking or even facial recognition are becoming available on a more standardized, deployable basis. I love the fact that we're putting quantitative metrics behind what had previously been pretty subjective judgments.

It's interesting to see The New York Times and The Journal do these big stories on biometric analytics.

It’s finally here. You can tell a really interesting story with it, compared to when you'd have analysts who were just struggling to find a storyline. That doesn't happen if you've done a good job selling a measurement plan now.

 

  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Email This Post Email This Post

By on at

Leave a Reply