Roughly three years ago, dairy and food manufacturer Land O’Lakes tested a customer data platform (CDP) to help with site optimization.
People who visited pages about horse management, for example, or how to set up a home chicken coop, were targeted with relevant ads. Land O’Lakes sells farm feed and equipment, not just products you see in a supermarket.
“Unsurprisingly, site conversions jumped,” said Steve Rude, manager of personalization and analytics at Land O’Lakes, and the company decided to make a more thorough review of the CDP market.
Rude, whose job is focused on business intelligence rather than advertising analytics, happened to join Land O’Lakes on the very same day it settled on Lytics as its CDP. “It was not a coincidence,” he said.
Now, with a few success stories under its belt, the brand is consolidating more data from across its business sectors into the personalization program. For instance, Rude’s team recently merged with the email team, further aligning its corporate function with marketing.
The move should help Land O’Lakes do more with its data.
Many executives at companies with access to first-party data don’t fully appreciate the utility of the data they’re sitting on, Rude said. One, because they don’t have marketer instincts, but also because the data is only useful when combined.
Land O’Data
There’s a challenge, though: Food and beverage brands don’t have what the ad industry would generally define as first-party data. Meaning, they don’t operate well-trafficked sites or apps and they don’t collect purchase data. But many still have valuable data sets.
The liquor company Beam Suntory, for example, which owns Maker’s Mark, Jim Beam and numerous other alcohol brands, started incorporating distillery visitor data into its digital media data – and turns out that doing so creates a very valuable potential retargeting and lookalike seed audience, said Jessica Spence, Beam Suntory’s president of brands, speaking at an AdExchanger conference earlier this year.
The berry brand Driscoll’s, meanwhile, incorporated data from its “cold chain” into its online advertising with good results. “Cold chain” is the perishable food industry term for what happens once a product like a raspberry is plucked … and the clock is ticking on freshness. There’s no point in advertising expensive berries if they’re going to be delivered warm and gross.
For Land O’Lakes, the distributor of Purina feed for horses, cattle, poultry and other farm animals, first-party data collection starts with its own sites and sign-ups for newsletters about owning farm animals. (Nestle operates the Purina dog and cat food brands you’ve seen in stores, but Land O’Lakes has the feed business).
During the pandemic, there was a spike in people setting up home chicken coops as part of what you might call “hobby farms,” Rude said.
“We started building some audiences around the visitors coming in: other behavioral traits they had, what they were doing on the website,” he said. “Many people were new to us and it was a great opportunity to expose more of the brand than just our flock content.”
Wait, FLoC? Hold your horses.
“A very different kind of flock,” Rude said.
One way Land O’Lakes was able to bake more data into its newsletter strategy was to consider timing. After 15 to 18 weeks, baby chicks transition to adult feed. Home farms with dairy cows are also on set timelines for product and milk replacement.
With a consolidated first-party data set woven into its personalization program, Land O’Lakes has been able to keep more “hobby farm” customers and increase the overall number of animals they manage, he said, partly because of the improved cadence of the educational resources for those who subscribe to their newsletters or log in to the site to buy animal feed.
Another advantage is knowing when not to sell. “One of the huge benefits to me of knowing things about our visitors is that we can isolate them and stop asking them to convert,” Rude said. Rather than continually pushing the sales message, the marketing team can step back to consider its next step.
Land O’Lakes is now testing many of its assumptions about seasonal sales cycles, Rude said. The company had seen sales spikes in farm animal electrolyte products that it expected to be tied to drought conditions. But that turns out not to be the case. Farmers just give the animals more animal Gatorade at times when they themselves are thirstier.
“So now we’re going through a process to align distribution centers with potential spikes in demand,” Rude said.
The first-party stack
Land O’Lakes has two different types of customer data within its first-party data set: aggregated and personal.
Aggregated audiences include reporting data from walled garden platforms. FLoCs (the ad tech FLoC, this time) can’t be attached to a CRM profile or allow for customers to be retargeted online. Retailers that carry Land O’Lakes food brands (human food brands) provide similar data that is useful for analytics, but can’t activated in a programmatic or social media campaign, Rude said.
Land O’Lakes works with Salesforce as its CRM cloud, with Adobe Analytics for ad measurement and also partners with Snowflake.
The CDP Lytics serves as a personalization engine for the other set of Land O’Lakes data: the kind that can be attached to a person (hence the focus on email sign-ups).
When Land O’Lakes has first-party data to anchor a profile, the information can be expanded beyond the CRM function, which primarily tracks purchases over time, said James McDermott, CEO and co-founder of Lytics. CDP data is more behavioral and matches to encrypted audiences elsewhere, so it can bridge to anonymized audiences, he said.
Land O’Lakes, for instance, appends data to customers based on affinities, such as whether they’ve bought maple-flavored or pumpkin-flavored products in the past or whether they’ve searched for ham recipes. This isn’t CRM data, McDermott said, but the expanded profile information could be useful for ads personalization, customer service or to alert people about sales on particular products.
The point is to help brands do more with the data they’ve got.
“It often starts as a process where the business is turning profiles into smarter audiences that can be used across a variety of purposes,” McDermott said.
