No more segments, no more personas. Just reach out and pitch to the individual. This is the holy grail of personalization.
Many of the technological building blocks have been used for years in an attempt to reach individuals. Is personalization finally achieved when you bundle all these marketing apps into one platform? Are we there yet?
Yes. No. Maybe.
“We think it is,” said Karl Wirth, CEO of Evergage, a firm specializing in real-time personalization. The concept of personalization has been around for a while, but could not be realized due to the limitations of technology. What makes Wirth’s answer a “yes” is the combination of three things: big data, machine learning and the cloud.
“No, we’re not there yet,” said Kevin Lindsay, director of product marketing for Adobe Target, which offers testing and personalization. “The bar keeps getting higher, the opportunity keeps getting higher.”
“Personalization is an evolution, not a revolution,” added Maribeth Ross, senior vice president of marketing at Monetate, which offers a personalization platform. Getting to the destination will not be like “flipping a switch”.
It depends on what company you look at, noted Mark Abraham, principal of the Boston Consulting Group. Only 15% of all companies are doing personalization today, while another cohort of established companies is still learning how to use the technology, he explained. (A full explanation from BCG can be found here.)
Machine rules
A personalization solution depends on a combination of machine learning, artificial intelligence, and a business person or marketer who can determine how the solution is applied. How people interact with these systems, and what results they obtain, will vary somewhat from solution to solution. And sometimes those approaches are dictated by outlook, not technology.
For Adobe’s Lindsay, the difference is “contextualization”, which should take personalization to the next level. What enables this is the mobile device. As a users get around, a company can serve up data that is relevant to their location. Still, while automation will make up 80% of a solution, the remaining 20% is where the human resides. Just because a system can do something does not mean it should be done.
Lindsay offered one example. “Selling shoes. It is not a big risk if you show a pair of women’s shoes to a man.” That mistake probably won’t offend, yet an automated system can make this error if the data adds up a certain way. “We [at Adobe] make it easy to understand what is coming out of machine learning,” he said. “We give the marketer the ability to understand and take action.”
Evergage relies on a plain-language interface so that a non-expert can analyze data and craft rules that will translate into machine-operated algorithms, Wirth explained. “A human can jump into the system any time they want.” The company refers to this product feature as “Recipe” — a collection of rules and recommendations that should produce a personalized pitch, based on the customer data a company possesses.
Personalization “must be unique and right” for the individual, Monetate’s Ross said. But that experience does not have to be 100 percent unique to them. There is a subtle distinction here. “Artificial intelligence can help design what is served and where…It can choose from a library of stuff you already have. No human can do this in real time.” So while a group of individuals can share an interest, how you reach each person and with what can still be unique. The outreach is custom-made, while the components of that outreach are catalogued and tapped as needed.
Automation does away with the need for hiring creatives to reach every segment, group and sub-group with tailor-made messages, Ross noted. But that does not mean a marketer is cut out of the loop. “The black box is the marketer’s biggest fear,” she said. “We provide visibility back to the marketer why the machine makes the decisions, and that understanding can provide insights,” she added. “I never want to be replaced by a machine.”
Implementing personalization requires a different approach. Companies are “usually set up in a product-fashion first,” BCG’s Abraham said. Each product line becomes its own silo in the organization, which is geared to pushing out more stuff. “They are not thinking across the company in a customer-first fashion.” Personalization requires the firm to cut across silos, not just product lines, but channel silos as well—store, web site, e-mail. “The customer journey cuts across all channels.” He said. “Teams are not used to working together.”
“This is not a problem of ideas…the real issue is around execution” Abraham continued. To lick this problem, a company can pull together a small team made up of people from IT, analytics and marketing, and focus on delivering minimal steps towards a solution, he said. If you can make a one-percent improvement in two or three increments every week, then some KPIs will see significant improvement after a year, he explained.
The personal touch
How marketers handle data for personalization will also have some impact — negative if they get it wrong, positive if they get it right.
Opting into personalization, and having a sales associate know you when you walk into store, is good, noted BCG’s Abraham. But: “If it happens without consent, it can be creepy.
A different approach is to look at a product or service as a customer experience, Abraham continued. One general principle is to look for “pain points” in the transaction, then change the process to produce a convenience. Abraham offered some examples. Starbucks allows customers to pre-order from their smart phones while waiting on line, so the coffee is ready by the time they get to the register. Or Sephora, which uses augmented reality to allow users to “try” various shades of lipstick, using their picture or smartphone selfie. Yet these two solutions required both companies to see their services as creating a customer experience.
Ross offered Office Depot as another example. There, Monetate’s AI and machine learning was used to alter web pages delivered to customers, depending on where they were in the purchase cycle. Again this paid off, with a $7 million sales gain in four months.
For many companies, personalization is a when, not an if. While established brands may not be as far along as “digital natives”, they must find a way to adapt the technology to suit their customers’s needs.
Or they can do nothing and wait for Amazon to eat their lunch.
