Exclusive Q&A: Wayfair applies AI to supply chain disruption
Wayfair Inc. is solving issues such as demand spikes, storage, and on-time fulfillment with artificial intelligence (AI) solutions.
Chain Store Age recently spoke with Nitin Kapoor, VP of supply chain technology for Wayfair, about how the direct-to-consumer (DTC) home furnishings retailer utilizes leading-edge technology to increase supply chain efficiency and accuracy while reducing costs, in response to the disruption which has occurred as a result of the COVID-19 pandemic.
What issues has Wayfair been having with the supply chain?
The pandemic-induced supply chain challenges affected all e-commerce businesses with no exceptions. Back in 2020, Wayfair faced struggles with our third-party partners who help make our end-to-end supply chain possible—this includes ocean and trucking companies, last-mile assembly and delivery companies, and more.
Like many, these businesses faced labor and inventory shortages, which resulted in unpredictable performance variabilities. At the time, online shopping was soaring, but order shipments were often unpredictable—some were coming in late while others were arriving early. The ramifications of delayed shipments are fairly obvious, especially with customers who are counting on their items to arrive in the time frame promised.
But early deliveries also presented a unique set of challenges for Wayfair in areas such as our last-mile delivery stations, which are not designed to take on unexpected short or long-term storage needs. We also struggled with our demand forecasting.
No one expected demand to spike to the degree it did in 2020 and then soften so quickly in 2022. This caught the entire e-commerce industry by surprise, and for Wayfair, it showed that we needed a wider set of supply chain capabilities. This includes better technology optimization algorithms and a technology-enabled operations management system that can take into account the wider performance characteristics across all segments of the supply chain.
How did you decide to apply AI technology to your supply chain issues?
The infusion of artificial intelligence (AI) into the supply chain began with our demand forecasting models. We needed the ability to predict exactly what assortment would be needed, the quantity requirement along with a breakdown into the various geographic areas where we do business.
Most retailers, including the more traditional businesses, utilize demand forecasting to purchase inventory. But forecasting is a little more complex for Wayfair since we don’t purchase most of our inventory from suppliers. The majority of the inventory in our supply chain is owned by our supplier partners.
Therefore, our demand forecasting models require more ‘explainable’ AI - i.e., logic, that makes it easy for suppliers to understand why we make certain recommendations on how much inventory they should commit to Wayfair and where it should be placed.
This transparency is a big part of how we earn their trust, and it’s an order more complex than black box AI typically used in the retail industry. AI brings many benefits to our customers and partners, which for us includes suppliers.
What benefits does Wayfair’s use of AI provide your customers and supply chain partners?
For consumers that purchase from us, the big three are faster speed, lower costs, and a greater, more relevant assortment. We also apply AI-based models to make and deliver on customer promises.
For example, when a consumer finds an item on our site, they can see exactly when it will be delivered. That delivery date is our promise to them, and AI plays a vital role in ensuring that we make it accurately and meet it every time.
[Read more: J.B. Hunt, Waymo to expand Wayfair autonomous delivery pilot]
As for suppliers, AI enables more inventory turns, which increases their working capital and margin expansion. Lowering costs for us and our suppliers is also important here. Our north star is to remove costs from the ecosystem whenever possible by making the supply chain more efficient. When we succeed in reducing or eliminating costs, we lower prices for consumers while increasing margins for suppliers. Both of these are key to maintaining loyalty.
Are there any future AI plans you can discuss?
As a company built on innovation, Wayfair always looks for ways to expand our AI efforts. Right now, we are working on an AI-based overlay that will sit on top of the statistical/operations research models. There are a number of areas where these technologies have complementary strengths that enhance outcomes. I’m excited about some initial tests we’re running, which when scaled, will help us make even more accurate predictions for when products will get to the customer and anticipate any exceptions/delays before they occur.
We are also working on new geographic sorting (geosort) capabilities that leverage machine learning to identify and boost relevant products that are closer to customers. This has everyone excited, and here’s why.
Today, a typical product travels approximately 1,000 miles to a customer. With geosort technology, our team believes we can reduce that figure to less than 250 miles without impacting product relevance. This means customers will receive their products in less time with the likelihood of any item being damaged dropping dramatically, meaning fewer returns. The shorter travel distance will also reduce our shipping costs, allowing us to cut prices.
[Read more: Wayfair shifts data center to Google Cloud]