When you buy a bulky item like a sofa on the internet, the last thing you need is for the product to arrive and look nothing like what you expected – this kind of mistake is costly and frustrating not only for the person buying the sofa, but also for the company selling the item.
So what if there was a way to use technology to make the more enjoyable shopping experience and to reduce the risk of unpleasant surprises on the day of delivery?
That’s exactly what Wayfair is doing with the help of machine learning. The e-commerce company, which sells furniture and household items online, uses a purpose-built platform that was built with tech specialist Snorkel AI.
Tulia Plumettaz, director of machine learning at Wayfair, says the platform helps her company improve the quality of the online search experience it provides to consumers and ensure that the sofa you receive looks like the sofa you ordered.
“We have these bulky items that are difficult to transport,” she says. “We want you to be inspired and confident that what you’re going to get is what you’re buying. And we want that to happen without you even touching the product.”
Delivering this type of high-quality online experience is far from simple. Wayfair’s site includes thousands of products with a large number of potential variables, including size, color, and texture.
An additional complication is that the e-commerce company provides a platform for its vendors to sell goods to customers. Plumettaz says Wayfair sometimes receives a limited amount of product information from its suppliers, so it can be difficult to provide detailed descriptions to customers.
This is where Snorkel’s platform plays a key role in providing rich product information.
“We want suppliers to find it easy to work with us using our advanced technology. We want them to say, ‘I gave Wayfair a picture, some information, and without much effort, my item just started selling,’” Plumettaz says.
Plumettaz also claims that machine learning supports “fast labeling operations” through a bespoke solution that was developed through a design partnership.
Snorkel already had its key product called Snorkel Flow, which is a data-centric AI platform for automated data labeling, embedded model training, and analysis.
But while Snorkel Flow focuses on text, Wayfair needed a solution that would support programmatic tagging of images.
According to Plumettaz, the solution, which was developed over a twelve-month period by the two companies in combination, offers benefits to both companies: Wayfair can shape the technology it uses, and Snorkel opens a path to a new, rapidly emerging market.
“We engaged together, and the result is a new development that brings programmatic labeling to computer vision,” says Plumettaz.
Now, with bespoke technology in place, the Wayfair team can label and re-label products quickly and efficiently.
Rather than having to rely on humans to manually tag up to 40 million products, automation takes care of much of the heavy lifting before business specialists, such as category managers, ensure the right images are delivered to online shoppers, Plumettaz says. “With programmatic image tagging, we can match catalog products to items customers are looking for as new trends emerge.”
Machine learning is also a productivity boost: with less time spent labeling images, employees can now focus on higher value activities. “It makes what we do a lot more interesting,” she says. “At Wayfair, our employees have no shortage of activities to do – consider maintaining such a rich catalog. This way we can now be more productive. It has made our lives easier and our work much more profitable.”
While Wayfair has chosen to work with Snorkel, Plumettaz acknowledges that there are other tech players who continue to develop their own machine learning solutions.
She says every company has its own stack, and in such a rapidly developing market, it’s hard to know where machine learning is going next. Plumettaz advises other professionals studying emerging technologies to make early breakthroughs and build strong partnerships.
“The field is changing so fast,” she says. “Five years ago, it was much harder to integrate with a machine learning vendor. Today, the hurdles to vendor approval are rapidly disappearing.”
While machine learning can bring a big boon to customer and employee experiences, Plumettaz says professionals shouldn’t let emerging technologies operate in isolation.
Left to its own devices, an automated system could start mislabeling products, leading to unhappy customers and what she calls “huge consequences.”
“You can have an amazing model, but the noise that can occur through a 1% error rate – like when a large item is delivered to your doorstep and it’s wrong – is huge.”
The lesson for all business leaders is to ensure that the human stays on top of what remains a nascent area of development.
“It’s a journey with a lot of these apps,” she says. “Let’s automate, but still keep a layer that checks that the automation is working.”
Plumettaz provides more details on how this process works at Wayfair. “When we’re not confident, we put the outputs in front of a human and get feedback,” she says. “It’s what I call the path to automation. It’s like a toddler; it’s not yet an adult who can run. And that’s the framework we use for these kinds of applications.”
Another lesson for professionals considering getting into machine learning is to focus on cross-organizational integration and processes, especially in terms of technology implementation, use, and operation.
Plumettaz says takeout will be familiar to professionals introducing new systems or systems: don’t just implement technology for the sake of it. “Building very close partnerships with business owners and product owners is key,” she says. “I think the blocking is less about the technology and more about thinking about machine learning as a driver of business value from the get-go.”