Deckers Brands is a global leader in footwear, apparel and accessories. Founded in 1973, the company has five world-renowned brands in its portfolio, including UGG, HOKA, and Teva.
"Each of our brands was started by an innovator who was passionate about something in their lives, whether it's surfing, or ultra-marathons, or whether it’s river rafting," says Richard Russell, Vice President of Omni Marketing at Deckers Brands. “And that passion, understanding what consumers needed, still lives with Deckers today.”
For much of its company history, Deckers Brands’ marketing strategy was channel-oriented, where individual teams worked in silos on singular objectives. This led to a fragmented strategy, and members of the Deckers Brands marketing team realized they were not adequately engaging with consumers throughout the customer journey. The team also continued to hear the same question from other parts of the organization: How do we know our marketing is working?
To address these concerns, Deckers Brands chose to reorient its entire marketing team to be more customer-centric. It also needed to invest in a first-party data strategy to connect customer data across multiple touchpoints. For this reason, Deckers Brands partnered with its digital marketing partners at Jellyfish and adopted Google Marketing Platform as its ads and analytics solution.
Spotting consumer trends with Google Marketing Platform
While Deckers Brands always prided itself on being agile, the global pandemic in 2020 was a challenge beyond anything it had previously faced. Fortunately, since the company had consolidated its advertising and analytics efforts with Google Marketing Platform several years before, it could better adapt to the rapidly changing marketplace.
Deckers Brands used several Google products to gain a more holistic view of its first-party and customer data. It used Google Analytics 360, part of Google Marketing Platform, to visualize how consumers engaged with its websites, and then it ingested its customer data from Analytics 360 into BigQuery, part of Google Cloud, to better understand customer buying behavior. Finally with Data Import, the team was able to bring those insights back into Analytics 360 for review. As a result, Deckers Brands was able to quickly spot and react to customer trends throughout the pandemic in 2020.
When many countries around the world instituted stay-at-home orders, Deckers Brands noticed a trend of consumers buying shoes for below their work-from-home desks: UGG slippers. Later, when people began to leave their homes, orders of HOKA running shoes increased. And when people finally started venturing into the outdoors and exploring nature, Teva shoes were top-of-mind for consumers.
When the Deckers Brands' marketing team noticed each of these trends, it could surface them throughout the company – influencing everything from media investments to supply chain management. These quick adaptations to its business operations helped Deckers Brands better serve its customers throughout the pandemic.
Optimizing ad creative with machine learning
Once the Deckers Brands’ marketing team knew what customers were looking for, it also sought out ways to deliver the best creative to them across media channels. Previously, it had been difficult for Deckers Brands to scale and optimize its ad creative due to extensive and costly manual checks. To solve this challenge, Deckers Brands worked closely with Jellyfish to move towards a data-driven framework for improving ads for one of its brands, UGG.
Prior to the global pandemic of 2020, Deckers Brands had worked with Jellyfish to automate creative insights at scale by combining Google Cloud products like Cloud Run, DataFlow, App Engine, and BigQuery. This unlocked the ability to scan a variety of creative elements to test against performance signals, including color composition, logo, faces, objects, and text.
The data science team at Jellyfish could then build an interactive dashboard using Data Studio that allowed the marketing team at Deckers Brands to easily access and interpret those creative insights. At a glance, the team could now answer questions like, “Do ads featuring happy faces lead to higher engagement?” or, “Does the number or the type of products shown impact performance?”
Modeling consumer behavior with BigQuery
The Deckers Brands team wanted to ensure that its now optimized creatives for UGG products were reaching the people who were most likely to make a purchase on its websites. In partnership with Jellyfish and Deckers Brand’s account team at Google, Deckers Brands set out to develop a propensity model, which would apply machine learning to the company’s first-party data to predict first-time purchasers.
The Jellyfish and Google teams took UGG's historical purchase data and used BigQuery Export to export its Analytics 360 data to BigQuery. The Google team then used this data to build a propensity model with BigQuery ML to predict the likelihood of a website visitor making their first purchase.
To assess the model’s proficiency, the Google team tested it with a sample of UGGs’ first-party data that was not previously used to train the model. After this series of tests, the model proved to accurately predict purchasers.
Using the Data Import feature once again, the propensity model insights were pushed from Big Query to Analytics 360. They were then used to make Analytics 360 audiences, which were finally pushed into Display & Video 360 and Google Ads for activation across its media campaigns. The audiences from the propensity model drove a higher conversion rate than standard site visitors in UGG’s display advertising.
Preparing for the future
With Google Marketing Platform and Google Cloud, Deckers Brands was able to effectively use its first-party data to grow its business, communicate more meaningfully with its customers, and prepare for an uncertain marketplace.
The company plans to continue to lean into new projects with Google Marketing Platform, including the scaling of its propensity model to other brands and marketing channels, as well as building a lifetime value model to better predict high-value customers.