How AI Improves Personalization for Condé Nast and the NBA
By Mike O'Brien | April 25, 2019
Googling “hottest technology of 2019” brings up 214 million search results, most of which concur: AI. However, while marketers all generally agree that AI will radically transform the industry — McKinsey reports that retailers that have wholly embraced AI generate profit margins ten percentage points higher than their competitors who haven’t — not everyone is sure exactly how.
Though the National Basketball Association is a large organization, its marketing department is fairly lean. AI is an invaluable asset to the company’s email marketing program, enabling the team to A/B test subject lines and send times far more efficiently, for one.
“How do we automate that in a smart effective way that makes our lives easier but also delivers relevant content to each of our fans?” asks Kara Loyal, the NBA’s Director of Email Marketing Solutions, who spoke at Sailthru’s recent AI Roadshow event in New York City. “With AI, we can automate and personalize messages, pushing customers down the funnel.”
“AI is the meeting point of data science and automated decision making. It’s about making your life easier,” adds Kara’s fellow panelist, Andrew Perell, Director of Email Strategy & Operations at Condé Nast.
What Automation Means for Personalization
No matter your company or your category, your customers almost certainly prefer personalized communications. Surveying more than 2,000 people around the world, Marketo found that 78.6% of consumers are far more likely to engage with a brand when offers reflect previous behaviors.
Kara thinks of her own experiences as a customer. A working mom who’s short on spare time, she doesn’t always have time to browse and appreciates a personalized email. But at the same time, something that’s too personalized can be a little off-putting.
“The convenience is great, but I just checked out beds for three seconds and I don’t need 75 emails,” she says. “We have a fine line as marketers; automation isn’t the entire answer. You don’t want to feel like someone is following you.”
This blend also applies to content. For the NBA, that could mean making sure subscribers’ daily newsletters reflect their team loyalties. For Condé Nast, that could mean anything. Andrew’s job involves generating traffic and revenue from email, while increasing the channel’s footprint for 18 different brands.
Each publication sends at least one marketing email per day, totaling between 4 and 5 billion per year. And the titles — as diverse as Vogue, GQ and Golf World — have wildly different audiences. Andrew points out that at its best, AI complements human editors. An algorithm may put a Self article about weight loss next to another about positive body image; a human would see why that may not be the best placement.
“For Vanity Fair, that means featuring content that’s no longer relevant, like publishing something about the Mueller report six months from now because it’s popular and has a lot of page views,” he says.
AI Enables Smart Segmentation and Thoughtful Triggers
Though many marketers conflate the two strategies, personalization and segmentation couldn’t be more different. Still, AI is equally beneficial to both. Segmentation involves grouping customers together by identifiable characteristics. Determining segments involves weeding through data, which AI can accomplish in seconds.
Those identifiable characteristics can be demographics, such as age, geography or gender. They can also be based on interest engagement levels, which is a big factor in Condé Nast’s triggered messaging strategy.
Responsible for the vast majority of email revenue, triggered messages should be a crucial part of every email marketing program. While these emails often responses to consumers’ onsite behaviors, they’re also far more complex than that.
“If we have permission to email someone, how do we make smart decisions around what kind of content we can send that they didn’t necessarily ask for? We have to think about frequency caps and reservation systems in a way that doesn’t require a lot of manual work and account for every conceivable scenario,” explains Andrew. “A GQ sneaker drop is safe because they’re not happening that often. But if an Epicurious reader has a demonstrated interest in chicken, we can’t constantly send them chicken content.”
A Marketer's Guide to Evaluating Personalization Technology
Personalization has never been more popular - or more peddled. With thousands of marketing technology vendors to choose from, marketers may be tempted to simply “add on” personalization to their ESP. But doing so could result in siloed data, half-baked personalization, and wasted money.
Recommendations to Revenue: 8 Algorithms to Drive Sales and Personalize the Customer Experience
By Admin Adminer