Alaska Airlines has a straightforward approach to measurement and attribution. The company uses Google Analytics for multi-touch and data-driven attribution, paired with media mix modeling (MMM) analysis, alongside incrementality and lift studies.
Wait, did I say straightforward?
I meant extremely complicated.
But there’s no way around using an assortment of methodologies for marketing measurement, said Owen Bickford, the paid performance media director at Alaska Airlines.
“I don’t think they can work in a vacuum,” he said. “It’s just not feasible to rely on last-touch or multi-touch attribution anymore, so you have to run the gamut of all these other ones as well.”
Why MMM? Why not
Even with its already crowded plate of measurement methodologies, Alaska Airlines is one of the first enterprise brand testers of Meridian, Google’s open-source MMM product that launched last year.
The brand is working with Adswerve, a Google platform services provider, to develop its MMM and customer analytics. Previously, Alaska Airlines used an in-house MMM tool from Adswerve.
The first test of Meridian showed promising results, Bickford said. The brand will begin running quarterly MMM tests using the open-source model, he said, which should provide a rolling benchmark that helps allocate budgets to the right media channel.
For example, Meridian helped shine a light on some media channels, including CTV and out-of-home media, which simply aren’t highlighted as well using data-driven online attribution methods. Those impressions don’t plug into the data-driven system that relies mainly on user-level info like cookies.
This is particularly important for Alaska Airlines right now, Bickford said. The company is expanding to new cities in Asia and Europe and is incorporating new routes with Hawaiian Airlines, which it owns. In other words, Alaska Airlines is pushing its brand to new places and trying to reach new people as a global airline. Thus, there is a new imperative, he said, to prove the worth of its upper-funnel brand marketing as an engine of new bookings.
MMM will play a key role, he said, in demonstrating ROI “that was never going to be captured by a multi-touch or data-driven attribution model.”
Crossing the Meridian
Why turn to Meridian, rather than Adswerve’s off-the-shelf MMM or another MMM software service?
Many brands have expensive MMM and measurement service providers, and some are eyeing Google’s open-source version that might prove more effective without the large subscription costs, Luka Cempre, Adswerve’s head of data modernization and cloud strategy, told AdExchanger.
There are trade-offs, as always. Meridian is especially useful because Google has added features that incorporate Google first-party data. It is also not a simple drag-and-drop solution. Meridian still requires data science expertise and a knowledge of Bayesian causal inferences just to set up.
So there is still a need for service providers who can run MMM programs for big brands, Cempre said.
But the big draw for Meridian is the Google data.
One particularly useful feature, according to Cempre, is Google’s incorporation of search trend data into the model. The big benefit of this new data set in the measurement is that it maps results for the brand with the airline industry writ large. If Alaska Airlines sees a strong jump in travelers from a particular region or during a period of time, was that because Alaska’s marketing paid off or was it simply in line with the overall market trend?
For marketers, he said, there’s a challenge in assigning ROI to a media or creative strategy, when perhaps the results just reflect normal seasonal patterns.
And new features that advertisers would like to see in Meridian are also based on access to Google’s first-party data.
For example, Bickford said one request he’d have for Meridian would be for Google to tie it more closely to Google Analytics (GA4) data. Using GA4, he said, it’s possible for marketers to connect users within a chain of events, like being targeted by a search ad, then reached by email before converting.
Meridian wouldn’t deterministically attribute upper-funnel ad impressions to bottom-of-the-funnel performance campaigns and conversions, because MMM relies on aggregated data, not user-level connections. But, he added, Google could do that modeling itself based on its first-party data and report back results in the system, more like how a brand uses GA4.
Although even getting a handle on GA4, which underwent a major overhaul a few years ago, is a journey. Luckily, Alaska Airlines is now lapping the new measurement system used by GA4, Bickford said, so year-over-year data provides useful comparisons.
“It’s been a ride,” he said of the changes to Google Analytics and to ad measurement overall in the past few years. “That’s why we’ve needed quite a bit of adjustments to understand what’s happening now versus what we were doing before.”
