Carhartt’s signature workwear is near ubiquitous, and its continuing presence on factory floors and at skate parks alike is fueled in part thanks to an ongoing digital transformation that is advancing the 133-year-old Midwest company’s operations to make the most of advanced digital technologies, including the cloud, data analytics, and AI.
The company, which operates four factories in Kentucky and Tennessee and designs all its products at its Dearborn, Mich., headquarters, began its digital transformation roughly four years ago. Today, more than 90% of its applications run in the cloud, with most of its data is housed and analyzed in a homegrown enterprise data warehouse.
Katrina Agusti, a 19-year veteran of the company who was named CIO six months ago, has played a pivotal role retooling the workwear retailer for the modern era, under previous CIO John Hill.
Now Agusti, who began her Carhartt tenure as a senior programmer analyst, is charged with leading the company’s transformation into its next phase, one that is accelerating daily with the barrage of complex technologies changing the global supply chain and business practices, Agusti says.
As part of that transformation, Agusti has plans to integrate a data lake into the company’s data architecture and expects two AI proofs of concept (POCs) to be ready to move into production within the quarter. Like all manufacturers in the information age, Carhartt is also increasing relying on automation and robotics at its service and fulfillment centers as it faces challenges in finding talent on the technology side and in the labor force to meet growing demand.
And demand certainly is on the rise for the workwear manufacturer, which is currently experiencing double-digit growth in all three of its lines of its business — direct to consumer, direct to business, and wholesale.
Tuning a transformation to make the most of data
Carhartt launched its Cloud Express initiative as part of a foundational transformation to shift the company’s 220 applications to Microsoft Azure. Two legacy applications, its warehouse management solution and its payroll and benefits solutions, still run on premises but those applications may soon be replaced in favor of cloud-native solutions, Agusti says.
Moving to the cloud — even amidst the pandemic — was a major win for Carhartt. Aside from the obvious speed to market and scalability gains, the vast improvements in stability, performance, uptime, maintenance, failover monitoring, and alerting has automated many of the costly, time-consuming IT tasks, thereby freeing up the IT team to tackle advanced data analytics and to experiment with other new technologies.
Agusti says Carhartt will likely embrace a multicloud architecture in the long run, but for now she and her team are ramping up their cloud expertise in part through conversations with other CIOs about best practices.
“We’re still learning and building the muscle internally to properly run in the cloud and how to manage in the cloud, and not just the management of systems but how to size them,” she says, adding that she is also homing in on data architecture and retention strategies. “It’s a different beast to manage workloads in the cloud versus workloads on premise. We’re still in that journey.”
Like many CIOs, Carhartt’s top digital leader is aware that data is the key to making advanced technologies work. Carhartt opted to build its own enterprise data warehouse even as it built a data lake with Microsoft and Databricks to ensure that its handful of data scientists have both engines with which to manipulate structured and unstructured data sets.
“Today, we backflush our data lake through our data warehouse. Architecturally, what we’d like to do is bring the data in first into the data lake, whether it is structured or unstructured, and then feed it into our data warehouse,” Agusti says, adding that they continue to design a data architecture that is ideal for different data sets.
She does not currently have plans to retire the homegrown data warehouse in favor of the data lake because the team has customized many types of certified data sets for it.
“The data lake will be more in service to our data science team and consumer-facing teams that are building out journeys using unstructured data to inform those personalization,” Agusti says, noting Carhartt’s six data scientists have built several machine learning models that are currently in test mode.
Two such projects are nearing production, the first of which supports Carhartt’s replication of inventory for its five distribution centers and three different businesses.
“We’re trying to use it for decision support and to plan all of that inventory into different distribution centers based on service levels,” she says, noting that the model can optimize Carhartt’s distribution network by taking into account capacities as well as supply and demand and inventory levels.
The second POC is aimed at helping data scientists collect consumer data that can be leveraged to “personalize the consumer journey,” including demographics information and data from consumer surveys, Agusti says.
The power of tech
Like many CIOs, Agusti’s biggest challenge is change management — especially when it comes to persuading employees that the company’s AI models really work.
“Teams are skeptical that technology can provide the decision support and automation that they do today,” the CIO says. “We have a lot of use cases and we’re running them in POC mode because we need to prove to our end users and business community that these models can make those decisions for you.”
Agusti expects many companies are in this transition mode. “There are different functions along the maturity curve,” she says of the AI efforts under way, “but I think there are so many potential applications that can leverage technology especially in data analytical spaces.”
To pique her resolve about the power of technology, all the CIO has to do is think about how, without investments in technology and talent, the pandemic might have derailed the company’s business.
At first, during the pandemic, many essential workers needed to be equipped with Carhartt work gear for extra protection. As a result, the company’s revenue stream grew in the double digits, even when certain business segments were curtailed due to widespread work stoppages.
Once work stoppages started taking hold, Carhartt gained a rare glimpse into its supply chain, enabling its data analysts to view the steps of the supply chain in exquisite detail, like the individual frames in a film.
“What the pandemic did was create the need for that visibility and proactive exception management,” Agusti says. “Every leg of that journey becomes important when you’re having disruption. It was the catalyst for us to get more granular in the visibility and exception management of every single step in the supply chain.”
Thanks to that visibility — and IT’s push to keep Carhartt’s businesses humming — the company is in a better place with its supply chain. It’s still not at the “predictable” level that it was pre-pandemic, Agusti says, but “we’re starting to see logistical lead times level out and improvements of lead times for goods creation getting better.”
Analytics, Artificial Intelligence, Data Management