There’s a disconnect between ad buyers in the trenches of online advertising and their own corporate leaders up the C-suite, who are accustomed to determining marketing group budgets based on ROI and other basic advertising KPIs that, frankly, are going haywire right now.
One way to deal with that disconnect is to add someone who speaks both languages, as it were. To bridge finance and marketing for the children’s clothing and shoes brand Carter’s, Jeffrey Coleman, a former finance analyst and agency analytics director, made the jump a year ago to the brand side, starting a role as leader of digital marketing science.
Coleman is serving as a finance-trained go-between for the marketing and finance teams, along with senior leadership, as the brand works through its budgeting and analytics process, Coleman said.
“Every day is a new challenge and it feels like the sand under our feet is shifting on a day to day basis,” Coleman said. “Last year wasn’t like this year; and even since then there have been major changes in Apple iOS, Google and Facebook. We’re all learning on the fly as we go.”
AdExchanger caught up with Coleman, as Carter’s negotiates a tricky marketing measurement overhaul that actually assigns credit (and budgets) to the media that deserves it. He also shares how the retailer must strengthen its first-party data set.
AdExchanger: What is Carter’s attribution and analytics setup right now?
JEFFREY COLEMAN: We think about getting as granular as possible, however it makes sense to do that type of attribution.
Right now, we have a media mix model that allows us to look at it from that lens as well. We had a multi-touch attribution model (MTA), but we’re in the process of finding another vendor to go with MTA at this point in time.
Are you ditching the MTA model, rethinking it or just adding a different vendor?
We rely heavily on media mix modeling. The reason is that, at least on the social media side, it’s more and more of a black box. But with media mix modeling we can look at [social media] from the lens of incrementality.
There’s always a factor of error in any attribution model. Media mix modeling allows us to ask, “What are the things that our model says are driving incremental gains?” Whether those things are traffic, or sales or new customer acquisitions.”
From there, we can portfolio manage spend accordingly. If you look at online video, for instance, somebody may watch a video and not necessarily click. How do you measure that? How does that come out in an attribution model?
Then we start to bring in the world of digital audio – all of these none-clickable or non-interactive channels. But digital marketing measurement is built around interactivity and events that can be tracked by user engagements. How do we assess how impactful those channels are on outcomes?
Media mix modeling allows us to do that when other models, whether that’s last click or multi-touch, are going to have holes.
How does the media-buying team navigate attribution changes that could mean big swings in their budgets or reported performance?
The emphasis put on privacy has changed the landscape of what digital marketers have access to from platforms. Google and Apple of course are huge players in this ecosystem. And some of the things they’ve introduced to safeguard user privacy have played a big role in what we can and cannot get a hold of.
Now it’s on you, the marketer, to develop first-party data. Media mix modeling affords us a way to still see and report to our leadership an indication of the value of marketing efforts where the data doesn’t allow that transparency like it once did.
But when you get down to the channel level, and you’ve got your paid search or paid social channel manager looking into the weeds on measurement reporting, it does become an effort of art and science.
Art and science?
For instance, campaign reporting suggests at first blush that perhaps you should not invest in digital audio, or to pull back there and spend more in our direct channel and search keywords because that’s performing. Well, that’s a false positive.
What’s happening is people hear the audio ad and go type “Carters.com”. That’s where it’s an exercise in art and science, to determine what’s working and what’s just reporting.
Online video has one of the highest incremental ROIs for us. The media mix model also says that digital audio ads have a high incremental ROI. But you would never know that from an attribution perspective because you can’t click an audio stream.
There’s a lot of art that goes along with interpreting what you see from self-reporting platforms, along with what your media mix model tells you is happening.
Have you added (or subtracted) vendors while you deal with these advertising and measurement issues?
When I got here there was a lot of manual reporting. We deal with a ton of disparate platforms, as I’m sure you know about digital marketing. So the first thing that I needed to do was to be able to pull in data with a sufficient enough velocity. I added a vendor called Adverity, which brings that data together so we can start to do the types of more advanced analysis.
Do the changes and updates made by Google Analytics affect your work closely?
We aren’t a Google Analytics shop, per se. We use Adobe analytics. But we do have to pay attention and have consideration for those changes because we use many other Google products, whether that’s GA360 (Google’s enterprise analytics service), Google’s display ad network and campaign manager.
The reason why GA has an impact on this is because IP addresses have been used, or even have been the main way to understand geo-location analytics. So when they pull away those data points, it just further crystallizes in our minds why we need to have as robust a first-party data set as possible. We have to depend less and less on the third-party data or platform-reported data and get more and more focused on building out our first-party data warehouse.
Some retailers have launched ad businesses or commercialized first-party data for outside advertising. Is that in the plan for Carter’s?
Where Carter’s sits, for instance, in kids apparel, you can see interesting intersections and commerce opportunities with other companies that serve the children’s space or young families. At some point in time we may be a good partner for that. Frankly, our first-party data is in the early stages of this maturation.
We have an app and a loyalty program that allows us to know our customer and their buying habits. And once we beef up our first-party data set we may be able to partner with non-threatening, non-competitive businesses that need data and insights around children and the parents of children who shop with Carter’s.
What are your first-party data sources and potential ways to “beef up” that data set?
One thing that we want to do is build a way to track our own own customers’ buying habits and behaviors outside of what Facebook and Google can give us.
We will also be launching our updated CRM this summer. That is going to allow us to really scale up the functionality in a way that we just weren’t able to in our old CRM. It’ll be a cloud-based solution built on top of the Snowflake environment. That CRM is going to allow us to start to scale our ability to collect the data and to track our customer IDs across different digital channels and activities.
This interview has been condensed and edited.
