“Obama’s was such a brilliant strategy because they didn’t talk to everybody, but they knew exactly what to do to drive the scale benefit they needed,” she told AdExchanger. “If the Obama administration could figure it out, we had to figure it out.”
She asked Bob Rupczynski, the company's VP of media, data and CRM, to teach her what she needed to know. This was the beginning of what ultimately became a companywide infrastructure, centered around Kraft customer data, that allowed the company to quickly adapt to the changing ways consumers interact with the world.
These efforts include the Looking Glass, a sophisticated listening room that tracks trends, manages reputations, launches campaigns and watches Kraft’s competitors. It also includes Parallux, a proprietary data-housing system on top of which sits a data-management platform that builds custom segments and a demand-side platform that purchases ad inventory.
“We have over 800 different custom segment targets, which we action against,” Elsner said during a presentation at AdExchanger’s Industry Preview event. "We used advanced analytics to create a learning engine. Marketers shouldn’t think about how to do this, they should act on the information."
Elsner and Rupczynski spoke with AdExchanger about how that infrastructure keeps Kraft nimble as new developments – like addressable TV – threaten to further disrupt the way brands communicate with consumers.
AdExchanger: How did you prove out this vision to the CEO and the board of directors? How did you choose which proof points to show?
DEANIE ELSNER: The proof points in a typical marketer’s journey are very linear. In this journey, our objective was to prove a better, more effective way to spend our advertising investment.
It enabled us to approach this in a non-linear way and prove concepts along the way and help scale them across. We built from a targeting standpoint. Then we dove into tailored messaging, serving dynamic messaging, retargeting, matching consumer profiles outside of our dataset and still getting the same effectiveness. I don’t have to prove every test works, I just need to prove I’m making progress.
How do you relate to the tech people on your team?
BOB RUPCZYNSKI: In too many organizations, there’s a digital guy running around and nobody understands what he or she is doing.
DE: Letting this exist as a black box is the biggest mistake a manufacturer can make because then you’re at the whim of every social, data or media partner on the marketplace to tell you what’s happening with your consumer. And you live or die by the knowledge of your consumers.
We’ve spoken before about Kraft’s prioritization of transparency. Was that instilled from the beginning?
DE: From the beginning. But we learned a few things along the way. We thought transparency in our data also existed in external partner data. Our data is far superior to anything we can purchase. When we unwind other data sets, they’re not nearly as deep or accurate as ours are.
We’ve learned the standards we have to put into place and how to qualify third-party data before we bring it in, because we don’t want to kill the integrity of our own data.
How do you qualify third-party data?
BR: We won’t comment on the process we go through, but all third-party data is vetted at a very deep level. We have a sliding scale we rate it on, and all of our planners on the agency side know our ratings.
DE: The transparency with our agencies, external partners and internally around how we’re touching things is imperative to getting to the biggest bang possible. We’re not the biggest advertiser in the world. Today, we’re not global and we know we come in at a disadvantage. But we have an infrastructure and capability nobody else has, and we’re looking for partners to help us get better.
How do you onboard each brand?
DE: We moved the brands on by groups, and every test has been at the brand level, but from the start, it’s been a portfolio play. That’s why it has to be a top-down initiative. This is a reinvention of marketing.
Every one of our brands buys programmatically. This year, you’ll see a ramp up in how they action [Kraft’s data infrastructure]. We’re in the early days of the rollout, but some brands – like Oscar Mayer Bacon – have taken advantage of this system for a few years. Now it’s institutionalized.
What will this enable Kraft to do that it couldn’t do before?
DE: It’ll manifest in speed-to-market and agility. We had 40 middle to senior marketers out at CES two weeks ago, in meetings from 7 in the morning to 11 at night. Within days of getting back to the office, they’d jumped on partnership opportunities that will hit the market in the next three weeks.
Because we have the infrastructure, because of all the work we did up front, we have the capability to turn marketing decisions on a dime and take advantage of partners in the environment. You’ll see an acceleration in our speed-to-market and improved relevancy in our content, which I believe will drive top-line growth.
As a brand, numerous technology companies – both startups and established enterprises – are interested in working with you directly. How do you deal with all of that?
DE: We approach this very differently than other manufacturers. Data is the heart of everything we’re doing. We will own our data, and we will build a model and long-term vision on our terms. To the extent you want to play with us, our dollars will follow what we can validate and prove in market.
We’re not walking into these partnerships, throwing money and seeing what sticks. We have to define a plan. We know where the different social partners fit, how they can help us and who we can benefit from. Our dollars will follow their ability to align within our infrastructure.
With the explosion of the ad-tech world, marketers are disregarding the basics of the marketing relationship. You have to know your consumer, you have to have a product that’s relevant to that consumer and talk to them in a consumer-friendly way.
During your Industry Preview presentation, you said Kraft might be one of the first consumer packaged goods companies testing delivering TV messages at household level. Can you discuss what you’ve learned so far in addressable TV?
BR: Marketing automation needs to happen in the TV space. Much of what takes place today is manual. Automation needs to catch up.
When do you think addressable TV will be available for marketers?
BR: Let me clarify your question: You’re talking just addressable and scalable, not programmatic?
Is addressable TV a subset of programmatic?
BR: Programmatic implies real-time and auction-based. Programmatic TV is real-time bidding, whereas addressable is preset. I can buy an audience months in advance.
What we’d like to see is scalable, automated, addressable buying, so it’s not a manual process like it is today.
So when will we see programmatic and when will we see addressable?
DE: You’re coming into a debate we’ve had so many times. Programmatic works when there’s an abundance of inventory. Will TV ever be in a position where there’s an abundance of premium inventory?
What I’m salivating over is whether I can get my message appropriately delivered to the right consumer at the right moment. Can I influence her through that messaging?
BR: Programmatic is a long way off. Tell me why you’d even want it.
How about programmatic for whatever is on TV at 3 or 4 in the morning?
DE: The pipes aren’t ready. It comes down to infrastructure. How would this exist? TV is the most antiquated server in the whole industry. It is still highly manual and a lot would have to happen to make that possible.
BR: TV is still a very powerful medium. We’ll never walk away from television. But the infrastructure, not the medium, is dated.
DE: What I want to know is if I can get to addressable. That’s the win for us.
And you mentioned that addressable is closer, though you’ve got different points of view on how much closer.
BR: It’s much closer. You see AT&T and DirecTV, Comcast with NBC, all of these companies starting to merge. There will be more over the next year. This will accelerate the industry. And ad tech and content will come together in ways we didn’t envision. And that will lead to the digital delivery of linear television. From a scaled, addressable TV perspective, I’d say it’s within three years.
DE: Bob’s more bullish. I’m more in the three-to-five years range. But I believe it’s going to come.