Skip to Content
IT

On the Front Lines of Retail: Why We Used AI To Drive Stronger Customer Engagement

It's time to make the right connections with customers.

Tim Delles

This article is contributed by Salesforce customer Orvis, an outdoor retailer who believes the most meaningful experiences are created by sharing the love of nature and being inspired by its endless possibilities. 

Marketers have traditionally lived by a steady philosophy: get the right message to the right person at the right time. Today, it’s about having the right conversation — the right connection — with the right person at the right time, and it has never been more important than during the current COVID climate. 

It’s a shift from one-way messaging to building authentic conversations. Meaningful relationships with our customers has always been a priority for Orvis, and we are always looking for new ways to deepen them. While we used artificial intelligence (AI) to find the best time to deliver better, personalized content, we didn’t know how AI could help drive engagement during a time when making that deeper connection mattered most. 

What customers want: Experiences

Over the last three years, we’ve shifted our focus to the experiences that inspire a love of the outdoors. During the holiday season, for example, we ran campaigns with “Moment of Chill” videos to help our customers unwind from the stress that can come with the holiday season. Think paddling on a serene lake, puppies running through a field, or the rhythmic hum of a grouse drumming its wings in the spring (a fan favorite); these short nature videos help people replenish and restore.

When COVID-19 hit, we quickly recognized that people needed these short nature-inspired reprieves more than ever. They wanted to connect with something helpful, something that would make their lives better in the moment. That lit a fire under us to figure out how to engage customers with Moments of Chill better and faster.

We already had a strategy of sharing content and telling stories; how could we put it on hyperdrive? Using AI, we focused on finding that sweet spot: where connection builds demand by delivering the right pieces of content in a given moment to generate the best customer response. 

Here are five takeaways from the last six months as our digital team worked to connect with customers in new ways. 

1. Focusing on inspirational experiences leads to stronger brand engagement

We started by integrating “Moment of Chill” into our email campaigns, and then expanded that with curated, brand-related video stories inspired by adventure and wonder in the natural world.  Using Einstein AI, we were able to optimize when to send different content to get the best reaction.

Many Orvis customers own dogs, and (no surprise) they love our Moment of Chill puppy videos (enjoy!) To personalize the experience for each customer, we load all eight puppy videos into Einstein’s Content Selector feature. The AI learns individual preferences about what and when content should be delivered, and then applies those learnings in the moment of engagement.

For example, the AI may learn that a customer engages very heavily with certain pieces of content and not others, so it shifts the conversation to what the customer is interacting with. 

By inspiring customers to reconnect with nature, we’re also inviting them to connect with our brand in ways that are meaningful for them, and it pays off. From April through July 2020, our most engaged customer segment grew by 20%, and that goes hand-in-hand with sales growth. And since we shifted to focus on the right customer conversation, traffic has gone up more than 30%, and click-through rate has increased 22%. 

2. Using AI uncovers unexpected trends 

We often go into a new season thinking certain products will be winners, but that can change. 

For example, we were surprised when the AI recently recommended we feature a shirt that we didn’t consider a high-revenue product at the start of the season. But when we sat down with our web merchants who do merchandising and digital positioning, they confirmed that customers were in fact consistently viewing and purchasing the product. The AI helps us understand what products are starting to outperform what we planned and what’s trending from an overall sales perspective.

Using the insights from AI based on our own marketing efforts builds an efficient feedback loop. We can respond to trends customers are interested in 6x faster, and can now quickly serve up personalized emails in 30 minutes rather than three hours. 

3. Testing is your best friend

Every marketer should be testing. But I confess: before COVID-19, we did very little testing. I’m happy to report that’s changed. We’ve since embraced an always-on testing philosophy. Within a given day, we test upwards of 10 versions of an email based on AI selection criteria like types of images used, as well as where/when/how an email is presented. One email may start with brand messaging and try to sell product second; maybe another email flips that. 

We measure success based on key metrics, including email open rates, click-through rates, and customer value, and continuously refine our AI strategy to optimize each one. Rather than just running with what AI has calculated as the “best” prediction, we’ve found that transparency into how AI arrives at specific insights is critical for hitting these metrics. 

Testing has helped us prove the value of our campaigns, and it’s also enabled us to get buy-in across our organization. Our email team took our early results to management and found an executive champion who immediately asked, “How do we make this bigger?” — a question every marketer loves to hear!

4. Decision making needs human input

There are times when AI may make predictions that at the outset seem like a bad idea, like suggesting we focus on selling our $20 dog collars. Why focus on inexpensive dog collars when we could be pushing our high-quality dog beds at a $200 price point?

That’s where the human component comes in. AI doesn’t always understand context, especially when it comes to the big picture marketing strategy, such as: Where within the customer journey are customers buying those? Are customers more likely to start with a dog collar and then graduate to a dog bed? Whether we’re looking at segmentation or merchandising, the human touch is still needed.

5. AI isn’t an all-or-nothing proposition

You have as much control over your AI strategy as you want to have. Many people think using AI in their email strategy means that all of their emails need to include AI. That’s not the case.

There are specific ways that we onboard customers and we’ve found that it’s more effective to not leverage AI. Instead, we’re focusing on helping new customers learn all about the brand. In other areas, like nurturing someone from an initial lead to a buyer, we use a combination of AI-driven content and static content.

When we take a step back and think about the strategies we have and the experiences we want to create, if AI falls into that experience, that’s where we use it. It’s about finding the balance that works for you and your organization. 

The ongoing adoption and growth of AI will continue to help develop highly personalized conversations with our consumers. The foundation we have built with Einstein AI, and the growing capabilities, will ensure we can meet our customer expectations not just today, but in the future.

If you want to learn more from other customers turning to AI to help their business navigate and recover from the crisis, watch our recent Learn AI event and get to know Einstein for Marketing Cloud.

Related Stories

Astro

Get the latest stories from The 360 Blog, every week.

Get the latest stories from The 360 Blog, every week.

Enter a valid e-mail address
Select your Country
Select a state
Please read and agree to the Master Subscription Agreement

Yes, I would like to receive the Salesforce Weekly Brief as well as marketing communications regarding Salesforce products, services, and events. I can unsubscribe at any time.

Salesforce values your privacy. To learn more, visit our Privacy Statement.