Article

Discover the key to creating a 1-to-1 customer experience

Attaining the holy grail in customer service is extremely challenging but doable. One viable approach is using a recommender system.

September 17, 2015

By Chris Bryson

The 1-to-1 customer experience is the marketing holy grail. It's physically very challenging, but a very wonderful idea. Just imagine, to offer a thousand customers a 1-to-1 customer experience, a thousand marketers sit at computers, monitoring, tracking and engaging one unique customer.

Okay, now wipe the sweat off your brow. Here’s the good news: a 1-to-1 customer experience is possible and we’re going to show you how. One approach involves the recommender system.

Why segmentation does not suffice

As an experienced marketer, you know a first step toward a 1-to-1 customer experience has been taken and that step lies in segmentation. Segmentation divides a retailer’s customer set into smaller groups based on demographics, buying behavior and personal preference.

The problem with segmentation is it relies on dividing a retailer’s customers into groups so it can be imprecise. Plus, it can be stagnant. After all, customers evolve, and unless you have the man power to update the segmenting rules daily or more frequently, your system won’t be fulfilling the need of your customer.

Enter the recommender system

While marketers have been getting good at using segmentation, technology has been evolving. Enter the recommender system. Cloud computing combined with machine-learning result in recommender systems, which historically were so advanced they were only available to companies who could afford to build it themselves. For example, Amazon and Netflix have had a “we think you would like” recommender systems for years.

But now machine-learning recommender systems are available in an easy and affordable way through select software companies. Recommender systems can be used for a huge variety of marketing efforts including 1-to-1 emails, personalized web and mobile experiences, eCirculars and even list planning.

Why you need a recommender system over all other options

Alright, you’re an advanced marketer, you get it. But why do you need a recommender system?

The most basic reason is reliability. Recommender systems are built on the foundations of machine learning. They are intentionally designed to yield predictive results based on actual trends in real customer data. This is the opposite of segmentation, which relies on arbitrary, human-generated ideas.

Let’s look at a literal example. Take Mary and Sue, both women who recently purchased a baby product. Leveraging segmentation, a big box retailer offers both Mary and Sue a discount on Pampers diapers.

But just because Mary bought a baby product doesn’t mean Mary wants to buy Pampers. Machine learning understands subtlety, and when used by the big-box retailer, offers Mary an offer on her preferred brand of diapers, Huggies, and Sue an offer on her preferred brand of diapers, Pampers. Two happy customers.

Another benefit to machine learning is that predictive results evolve as data changes – like when a new purchase is made and when a new product is offered. So retailers can evolve as their customers evolve, instead of being reactive. Talk about advanced.

With a technology update, you can offer a 1-to-1 customer experience across all of your channels – without hiring a thousand marketers.

Chris Bryson is the CEO and founder of engagement solutions provider Unata. To learn more about developing a 1-to-1 customer experience download Unata’swhitepaper.