Marketing is no longer about right-brain versus left-brain. It’s about AI versus IQ.
Technology is getting smarter, and consumers are turning to machines to help them make better purchase decisions. As people turn to their devices to make product recommendations and automate their services, one has to wonder: Will brands start marketing to machines versus humans?
“Marketing to machines — that kind of scares me,” says Jeff Barnett, CEO of CRM giant Salesforce Commerce Cloud. “Who’s working for who now?”
The history of marketing to machines
While the concept may seem a bit dystopian, it’s not completely farfetched.
The truth is that companies have marketed to machines for decades. John McCarthy is credited with coining the term “artificial intelligence” while serving as an assistant professor of mathematics at Dartmouth College back in the mid 1950s. Andrew Frank, research VP for technology research firm Gartner, says even SEO can be considered a form of marketing to machines — although, he notes that the attempt to generate higher page rankings by attracting robots has always been done with the intent of reaching consumers.
While the power and capabilities of technology have evolved over time, its role in marketing has remained fairly constant. Jay Henderson, director of cognitive platform IBM Watson Marketing, sums it up as helping marketers better understand their customers, and then using those insights to create more relevant experiences.
Fast forward to today and there’s a new kid on the block that can help marketers do this even more: virtual personal assistants (VPAs). It’s still early days for these voice-activated devices, but they’re already seeing adoption. In December, Amazon reported that it sold “millions” of Alexa devices over the holidays and that its Amazon Echo sales increased nine times compared to last year.
What marketing to machines could look like
Henderson considers VPAs a “natural extension” of the way consumers search for items via their Web browsers today—allowing them to research and discover new products with a single voice command. But if marketers can influence how results show up on the Web, what’s to stop them from swaying the way VPAs or other connected devices filter results in the future?
How this could be done is yet to be determined, but marketers can hypothesize.
Frank suspects that VPA filtering will be influenced by patterns, such as common voice commands or repeat purchases. So when a consumer asks Alexa to order laundry detergent, he explains, Alexa will know which kind of laundry detergent to buy based on that consumer’s history, similar to how advertisers leverage retargeting now.
Over time, the VPA will develop algorithms for what it determines best represents the consumer’s preferences, Frank says, leaving it up to marketers to figure out how their products meet the machine’s standards of relevancy.
“In order to clear those algorithms or clear that filtering process,” he says, “the marketer is going to have to understand what those trigger points are [and] what those references are that are likely to register [an item] as relevant to the agent.”
These “trigger points” could come in a variety of forms. Perhaps a discount on a comparable laundry detergent would be enough to spur an alternative option, Frank notes, or maybe an unsatisfactory customer experience would cause a customer to tell the VPA that he is no longer interested in making repeat purchases.
In the end, brand preference could be the greatest barrier competitors have to break through, and Frank says building that level of loyalty will become essential.
“When it comes to buying something through a VPA, I think the brands that have an existing relationship will have an advantage,” he notes.
But impacting filtering options might not be marketers’ only way impact on VPAs. Similar to how social networks like Facebook and Twitter (and now Snapchat) had to tinker with their marketing model, Henderson says marketers will have to experiment with VPAs to figure out how to best leverage them as a channel, whether that’s as a content play or an advertising structure.
VPAs aren’t the only technologies marketers have to consider, either. Other connected devices, like wearables, may also start serving more targeted content or products to consumers. Take the Fitbit, for instance. A consumer may use the device to help achieve his weight loss goals, Henderson says. Based on the fitness data collected, he adds, the Fitbit might serve that customer recommendations for health coaches or meal plans.
How much say will machines have?
Still, Frank suspects that it will be a while before machines start swiping the credit card.
“Ultimately, I think we’re quite a long way from the idea of the machine being the ultimate decision maker for buying something,” he says.
But Henderson isn’t so sure. He predicts that marketers will see more automated purchases for everyday items, like tooth paste or dog food—similar to how people reorder items with the tap of an Amazon Dash button. He also suspects that machines will be able to self-diagnose themselves and purchase items when they’re in need of maintenance.
However, he forecasts that machines will ultimately support human decision-making rather than replace it.
“I think both marketers and consumers fear a little bit [of] a black box that [they] can’t see or understand,” Henderson says. “It’s human nature to not want to give up control over either a purchase decision or a marketing decision. Again, I think there’s this big opportunity for machines and artificial intelligence to complement the human decision making—to have it be man plus machine.”
Some brands are already venturing down this path. 1-800-Flowers in one example. Henderson said the flower delivery and gifting company is leveraging an IBM-powered platform to create a personal shopping assistant called GWYN. The technology collects data about customers’ gifting preferences—such as whether they need an immediate present or have a favorite color—and makes recommendations.
Staples is another example. At Retail’s Big Show in New York, the company’s head of growth and applied innovation Ryan Bartley discussed how the office supplies retailer is working with IBM Watson to develop a voice-activated ordering system that will make it easier for B2B customers to purchase products on-demand.
Forecasting the future
Whatever role machines end up playing, both marketers and consumers will certainly face their fair share of challenges. Setting boundaries in terms of data collection and use cases is a big one.
“Often times, when it’s early in a channel’s lifecycle, it’s sort of unclear what’s appropriate in terms of a brand interacting with a consumer through that device,” Henderson says.
Ethical issues is another concern. Frank points out that it might be difficult to determine whether a machine is serving up content or products based on the consumer’s best interest or based on brand motives. And while he says consumers can look at whether the VPA provider is getting paid by the user or the advertiser, he said this might not always be clear.
Mastering these new technologies won’t happen overnight, and Henderson advises marketers to establish a structured experimenting process to determine how to best advance new machines into their standard marketing mix. But as marketers move forward in this ever-evolving digital world, Barnett urges marketers to not lose sight of the most important business element: the human element.
“At the end of the day, behind every app, behind every device is a customer,” he says. “Marketers need to empower customers on the channels they live on, and many of them still want to speak with people.”