As someone who has been in consumer-based merchandising and retailing for her entire career, Lynsey Munn is bucking the status quo.
Munn, Global Online Trading Manager for Jack Wills, believes that customers, not the retailer itself, should dictate what products are right for them. That’s why one of the first decisions she made when she joined the company in early 2017 was to broadly implement AI-based personalization with Salesforce Einstein.
“When merchandising is done manually, you’re telling the customer what’s right for them, but that doesn’t necessarily mean it is right for them,” she says. “That’s just your opinion of that customer. But with Einstein, you get insights that you haven’t been able to see before, and you get a better understanding of what the customer wants to see. That means you can put the customer at the center of the experience, and that’s a game changer.”
One of the first decisions she made was to replace its existing product recommendation engine with Salesforce Einstein.
“I had introduced Einstein into several websites, including UGG and Teva, across Europe when I was with Deckers Brands, so I was already aware of the functionality and how good it was,” says Munn.
At Jack Wills, a global UK-based clothing brand with more than 100 stores and a thriving ecommerce business, Munn is leveraging Einstein not only for product recommendations, but also for site search, with Einstein Search Dictionaries, and in-depth reporting, with Einstein Commerce Insights.
Munn and her team are laser-focused on delivering the most relevant products and content possible to its shoppers. That means, among other things, that they no longer guess which products a customer want to buy, or guess which products to group together on the site.
One example: using Commerce Insights, Jack Wills determined that 11% of shoppers buying a particular dress also purchased another particular dress along with it. Armed with that information, it categorized the dresses together on the site, driving incremental revenue.
“It’s not perfect, but we can see what people are basketing and buying together, and we can make edits to the site based on what we know our customers love.”
According to the Salesforce Shopping Index, 13% of all site visits used the on-site search function, and those visits contributed a whopping 28% of total revenue. It’s clear that shoppers who use on-site search are looking to buy.
Until recently, 11.3% of site searches at JackWills.com resulted in the dreaded “no results found,” even though the company likely did have what the customer was looking for. The problem is common, and stems from misspellings, typos, or variations in product descriptions. For example, if a customer searches for “trousers,” will the site return products covering “pants,” “bottoms” and trousers? Or how about colors? Will a search for “peach” also return products in “coral” “blush” or “pink?” How about regional differences? A “jumper” in the UK is a sweater in the U.S., and on and on.
The variants are endless, but must be accounted for. In January 2018 Jack Wills overhauled its site search and implemented Einstein Search Dictionaries. The results, nearly overnight, are stunning.
Its “no results found” rate has been nearly halved, to 6.1% of visits using site search, but that’s only part of the story. “Only about 8% of our customers use on-site search, but they’re converting at a rate 147% higher than those who don’t. They spend more time on the site, and they have more items in their order. It’s absolutely fantastic,” says Munn.
Shoppers who use site search, in general, are more productive buyers. They know what they are looking for, and when used in conjunction with other Einstein capabilities that quickly display relevant products and content, it’s an unbeatable combination.
The goal for Jack Wills: reduce its “no results found” rate to between one and two percent.
Simple, Integrated Product Recommendations
Product recommendation engines are not a new concept in ecommerce – we’re all familiar with “you may also like” on product detail pages – but they don’t all provide insight that empowers retailers with information they need to make tweaks that drive revenue.
In April 2018, Jack Wills replaced its previous personalization engine with Einstein Product Recommendations partly because it did not provide the depth of reporting they needed.
“We couldn’t tell how much conversion was being driven by personalization. Was it even moving the needle? There was no granularity in customer data. Right away, we identified that as an real opportunity. We were already paying for an ecommerce platform (Commerce Cloud Digital), and this is built right in.”
Munn says Jack Wills saw strong conversion improvement within the first week of testing recommendations, which are now running on four sites.
“We already see product affinity driving incremental revenue gains. We are now tracking and measuring engagement with Einstein recommendations, and conversion rate for customers who click on a recommendation is already much, much higher than our previous solution, higher numbers than we anticipated. When you’re able to see great results and reach that confidence level within a short period of time, you know it’s good. ”
The company saw such a quick improvement partly because of the ease with with product recommendations can be rolled out.
“It was so easy and fast,” Munn says. “We set up one recommendation rule and used the same custom attributes, and rolled it out across four of our five sites. The speed is just incredible – before, we were only able to manage recommendations for one site.”
The intelligent, automated nature of AI-based recommendations has made the lives of the merchandising team “so much easier” since they’re relieved from manual administrative work. One example: its previous recommendation engine relied on the team manually specifying the products or categories to include in recommendations, and the products they appeared on, which meant that we would sometimes recommend out-of-stock products – particularly at the end of the season. Now, we can set up a few simple rules using custom attributes, and can be confident in their success.”
Its use of Einstein has been such a success that the Commerce Insights dashboard is built into its weekly activities, with the hope of sharing the data with other departments to identify new opportunities.
Next up, Einstein Predictive Sort, which personalizes search and category pages for each shopper, displaying the products (in order) that the shopper will be most interested in buying.
“I’m really excited about it. We spend a lot of time looking at sorting rules. There are so many options, but you never really know if you’re serving what the customer wants. But if you can get to what they want, it can be massive.”
And that’s where Lynsey Munn bucks the status quo. She’s a tireless advocate for intelligent, AI-based automation across the Jack Wills customer journey, to deliver what the customer wants, not what Jack Wills thinks they want.
Check out our report, Personalization in Shopping, which analyzes the impact of AI on revenue. A key finding: personalized product recommendations drive just 7% of retail site visits, but an astounding 26% of revenue.