Since our founding 130 years ago, those of us at Carhartt have always maintained a tight focus on what we see as our mission: providing hardworking people with durable clothing they can rely on. Along the way, we’ve also become an iconic brand with a broadening appeal and growing visibility in popular culture.
But the recipe for growth is complex. If our experience is any lesson, it’s that great products and a strong brand indeed create the potential for strong growth, but solid execution at the people and process level—from suppliers to our sales channels—is essential to achieving it. Execution means decisions and actions. And data provides the signaling to guide it.
Predicting with AI
Carhartt’s traditional model of selling through wholesalers has always been inherently challenging because we’re a step removed from the end customer. But it’s something we’ve managed. A kind of turning point came a few years ago, when growth was ahead of our expectations and it took us too long to serve our wholesale customers and that, of course, trickled down to our end customers. The experience exposed a replenishment problem and we were determined to solve it.
The most basic thing we needed to achieve was more visibility into our wholesale customers. But the real heart of our vision had to do with broader transformation, affecting all the downstream processes related to replenishment decisions. Let’s start with prediction. We’re using IBM Watson AI technology to build algorithms that take into account a huge range of factors—from economic indicators to weather, even down to micro factors like changing retail store footprints—and automatically generate demand forecasts at the SKU level.
Speed through automation
In our vision of replenishment optimization, the speed of turning insight into action is the real game-changer, and our way to get there is through the intersection of AI and robotic process automation (RPA). The crux of our plan is that all the replenishment orders coming through will happen automatically, without anybody touching anything. With this in place, we’ll be able to give our wholesale customers the same level of service as our direct-to-customer sales: out the door in two days.
Learn how to delight your retail customers
Say one of our wholesale customers had a great weekend with a particular cold weather product like a jacket or boots. Our demand prediction models will take that sales data in, along with weather data that might show a big snowstorm expected in the region. Combining AI and RPA, we’ll be able to ensure that those products will be in stock in time for that next weekend.
A new role for planners
As we’re putting the pieces of this new model-driven, automated replenishment approach in place, we’re also beginning to see an important evolution in the role of planners in the process. If there’s been anything “robotic” about the daily activities of our planners, it’s the fact that they’re constantly making calculations. Under our emerging framework, planners will have as big a role as ever—but their activities will be far more strategic, pragmatic and analytical.
And that means new priorities. Like figuring out what new factors to include in their models to get better predictive accuracy. Or adapting and applying insights from some accounts to others. Or down the road, working with our growing team of data scientists to look for ways to increase conversion rates for a particular customer.
It starts with data
That last point highlights how vitally important it is to put in place both a data strategy and resources to implement a vision like ours. In the last year, we’ve doubled the size of our data team and centralized it so they can work as a team with our engineers and architects.
It didn’t happen by accident. Our CEO and COO took the initiative and provided unwavering support for our vision. We’re not there yet, but that support is helping us get there faster.
Listen to John Hill discuss how Carhartt is engaging IBM Planning Analytics to deliver the highest possible service level to its wholesale customers: