Physical Meets Digital in the Future of Retail
Back Bay, Boston: location of eTail East, 2017
At this year's eTail East conference held in Boston's Back Bay, it was clear that data-driven retail is in the process of taking a huge leap beyond eCommerce. From all sides, the message was clear: The retail customer journey no longer traverses two distinct channels, online and in-store. It's a holistic journey, crossing many channels and devices, and retailers who can't keep up with today's agile customer are in jeopardy.
Indeed, a number of observers were quick to name — off the record — high profile retail chains facing imminent peril. And of course everyone went on the record in naming Amazon as the number one challenge all other retailers face. But what struck me was the consistency and consensus when it came to identifying the key themes in retail's immediate future.
Leveling the playing field
Whether a customer walks into a store or goes online, it's possible to personalize which products are presented, Ian Strain-Seymour told me. He's VP of retail strategy at UGC content marking platform Bazaarvoice. I met a quorum of Bazaarvoice representatives at their eTail booth, including product marketing manager Amelia Carry, and Doug Foote, a new recruit on the advertising business development side.
Bazaarvoice is best-known for encouraging and aggregating content from the user side — in particular, product reviews and ratings, but also Q&As, videos and images, and social content generally. But it's also in the advertising business, leveraging data on shoppers from across the Bazaarvoice network to deliver relevant messages when they're truly in-market (although Bazaarvoice has executed media buys itself — not least to validate its segments — most clients execute on the Bazaarvoice segments through agencies or in-house).
What's more, Foote told me, they've now partnered with Placed, the location marketing vendor, allowing them to know where their in-market customers are.
Being able to personalize an experience for shoppers online, in transit, or in-store is great, but isn't it limited if it only reaches shoppers who've shown up on the Bazaarvoice network? Not really: Bazaarvoice reaches around 214 million addressable shoppers in the U.S., out of an estimated total of 240 million. And for any give retail site, added Carry, an average of around 53% of visitors had previously visited other Bazaarvoice network sites (in a defined time period).
As a general proposition, that means Bazaarvoice knows one in two of your customers; and potentially has relevant information about their interests and intent based on behavior. Standard practice among retailers is to cut off shopper tracking after a certain time with no activity (30 days, say) on the assumption that the shopper isn't in market any more. That means built-in atrophying of the ability to recognize even a good customer who hasn't recently shown a presence. It's one thing Bazaarvoice is trying to solve for. Crucially, Carry pointed out, this helps retailers begin to level the playing field with Amazon, which recognizes its customers as a matter of course.
Asked about other themes of the week, Strain-Seymour mentioned AI, but with reservations. "Lots of people are talking about AI," he said. "The graph of the hype train is in the ascendancy." There remain a lot of challenges, "steps people will have to work through," he said. He mentioned in particular the unreliability of a lot of data, drawn from multiple sources, and likely including many lookalikes rather than actual known customers.
Nevertheless, Bazaarvoice has twinned natural language processing with machine learning to offer a new feature on the customer review side: highlighted excerpts from reviews. With so much UGC available, it was necessary to automate the process of searching reviews for clear expressions of sentiment — positive, neutral, and negative — which can be surfaced to save shoppers ploughing through reams of content to find the nuggets. The solution is driving "double digit improvements in conversion," Strain-Seymour said. It tests well when it comes to audience enthusiasm too, with 89% of shoppers finding summarized information from reviews helpful.
"I give you my data: reward me"
Emarsys is not your typical marketing cloud. Although Salesforce made a strong push into B2C with the acquisition of Demandware, Emarsys fully identifies as a B2C marketing cloud: "designed to help retailers retain customers," said Nicole Hutzul, RVP Sales Americas, as we huddled around a table on the exhibition floor.
AI is very much part of the Emarsys roadmap. "The requirement," said Hutzul, "is to get to 1:1 personalization." AI will take away the manual process of building a campaign, conducting A/B testing, etc. "We already have AI within product recommendations. Incentives are AI-driven." Once the marketer sets the KPIs, the Emarsys engine generates personalized offers, discounts, coupons — whatever is appropriate. "It's got a brain, so to speak."
If the future is using AI to build entire campaigns, Hutzul admits that many marketers are not yet quite ready for that; but "they are now more interested in hearing a lot more about it." Why is AI necessary? Without it, "human personalization doesn't scale," she said. The biggest hurdle to AI adoption, she believes, is prioritization at the C-suite level.
When it comes to both AI, and the challenge of making in-store data actionable, Hutzul views the bigger retailers as facing the greatest obstacles, not least due to the "antiquity of their systems." For the most part, the in-store experience hasn't evolved; it certainly doesn't meet the expectations of millennials.
So who does get it? "It varies," she said, "from pure-play, single product companies" — like clothing companies, who need to retain customers for periodic repeat purchases — "to some very large companies" — like the large culinary brand consolidating four separate platforms in the interest of creating one, unified customer experience.
Emarsys serves B2C at all levels, from a small gift-basket vendor primarily interested in using marketing automation to support its email campaigns, to an international theatrical events producer. "It's about continuing to be relevant," she said, "almost like being your own personal shopper." Customers are telling retailers a lot about themselves; they're also saying "reward me" for that.
Mail is physical too
One aspect of the offline customer journey which is increasingly data-driven and digitized is that old standby, pieces of paper delivered by post. Pioneering postal-programmatic is New York-based Pebblepost, founded in 2014 by entrepreneur (and "Chief Stamp Licker") Lewis Gersh. I spoke with VP sales Tom Barbaro about recent developments, including the appointment of a Chief Growth Officer (Geoff Dodge, ex-AOL) and a GM for European operations.
"At the moment, we live solely in the U.S.," he said, "but we're starting to kick tires overseas." The logical next steps are Canada and the U.K. he said. "English-speaking makes most sense out of the gate," but there's also been "a lot of interest from France." The company is also going through the process of seeking to trademark the label "programmatic direct mail."
Unusually for a marketing tech company, Pebblepost finds itself in an uncrowded space. "When we started, we really didn't see any competitors," said Barbaro. "There's now a handful, maybe five to eight, who have some components of what we do, but I'm not certain any has the full sweep and breadth." Of course, Barbaro points out, growth of competition is positive in that it tends to validate the concept.
Pebblepost itself has expanded services on its roadmap. Right now, it collects and analyses intent data to generate a personalized mailing within 12 to 24 hours. "Conceptually it mimics re-targeting," Barbaro explained, "but the output is physical mail." Which so far has meant a Pebblepost postcard or a programmatic eight page catalog. Variations being considered include both larger formats and smaller — "Some marketers might not need eight pages" — as well as sealed envelopes for financial services clients.
Location plays a role for Pebblepost too: "We're certainly thinking about how to optimize campaigns, and one of the variables is zip codes" — reflecting store openings, or areas where there's need to drive more traffic to stores.
There's an analogy here with the potential for digitizing and driving data from the in-store experience. Physical shopping remains a very important part of the customer experience, and it ceased to be a non-digital experience once customers started carrying mobile devices. Similarly, postal marketing still has advantages over digital, but even more so if it's driven by digital insights. Mailings can have "a longer-standing impact than an ephemeral banner ad," said Barbaro; plus they can be viewed by a number of members of a household.
"The ideal brand to work with us," he said, "has an order value higher than $50, and more of a considered purchase period.
It's a mobile, mobile, mobile retail world
The key to understanding these themes is the dominance of mobile. Of course, I'd expect to hear that message from Marty Berman, VP of re-marketing at Inmobi, a solution which uses data to drive personalized, contextual advertising to mobile devices. But of course, he's right.
"From a mobile perspective," he said, "the tipping point was last summer." Consumers are now "comfortable with the mobile shopping environment." Inmobi offers dynamic creative (based on behavioral data) for mobile, such as product carousels; deep linking takes shoppers direct from the creative to the purchase page, within the mobile screen; and the transaction process on mobile is becoming increasingly frictionless. Last summer, Berman saw impressions migrating from desktop to mobile for many retailers: "It's only accelerated since."
Mobile first shopping creates opportunities, but also problems. "Right now is a critical time for retailers," he said, "and it's not just Amazon's fault. They've been caught flat-footed; they've not invested as much as they should have [in the relevant technologies]." Many big retailers are struggling, he said, because they're still not digital first. "The Wayfairs of this world are digital first."
As for the in-store experience, "mobile means you're always connected to the consumer," he said. "No other media provides as intimate an experience." This means retailers can potentially know if the shopper is in, or near, the store; the frequency of the shopper's visits; the shopper's online behavioral signals. In which case, it should be possible to serve the right type of ad, or exploit dynamic pricing opportunities.
Based on some 22 billion data points each day, Inmobi can tell the kinds of creative customers respond to best (video or carousel, for example), and uses AI to "connect those dots," as Berman puts it. An in-house team of data scientists provides the models to drive dynamic audience building and dynamic creative, and the quantity of data available provides the "scale and reach to bid more effectively" on media.
And something big is coming down the road. Apple and Google are paving the way for a cookie-less online environment, which will truly disrupt many digital business models. "It's a little ways off," he said, "but people aren't yet truly aware of how much it will affect their businesses."
Who would have thought, just three or four years ago, that in 2017 we'd be contemplating the declining importance of desktop-based eCommerce, the possible extinction of cookie-driven customer-tracking, and even a resurgence of the importance of physical retail environments — but not your mom'n'pop store — all centered on the mobile revolution. Although devices will continue to evolve beyond the smartphone (think wearables), there's no reason to think they'll be less mobile in future; and, as ever, the scale of data points to the importance of AI.
Those were the universal talking points at eTail East. Oh, as well as what Amazon will do next.