James Glover, co-founder and CEO, Coherent Path
Whenever we stream a song on Spotify, open up a Facebook page, or indulge in some retail therapy we’re bombarded with personalized recommendations.
Personalized community pages, playlists, and product recommendations are just a few examples of the individualized, one-to-one content produced when algorithms and predictive analytics work together. But what is traditional personalization really doing for customers? And where is it falling short?
As machine-assisted companies improve at collating data, they’re advancing at recommending products similar to the ones we already know and love. The downfall is the same categories and products are recommended over and over again
Take Spotify for example, the more we listen to music and the more our personal-preference data is fed to the algorithm, the more likely it’ll suggest songs and artists similar to the ones we like. Their traditional machine learning personalization doesn’t lead us to discovering new genres but rather stays near the artists they already know.
The same goes for clothing companies like H&M—search through the sweaters and its algorithms will populate your recommendations bar with nothing other than the latest sweater fashions.
Personalization based on reinforcement falls short
Even the most well-known companies like Amazon are guilty of using machine learning for safe, cyclical personalization, reinforcing a few basic truths but hampering new discoveries.
While its design is predictably complex, Amazon’s personalization algorithm is simple in its application. To establish a customer’s preferences and tailor products accordingly, it analyzes their past purchases, the contents of their virtual cart, products they’ve viewed or rated, and similar products other customers have bought.
We see the algorithm’s effects every time we visit Amazon’s homepage or receive an email—we are greeted with scores of recommended products similar to previously purchased items that have been “picked just for you.”
But Amazon’s algorithm has its flaws. After all, when you’ve just purchased a new microwave, it doesn’t do any good when Amazon reveals all the better-rated, better-value microwaves you could have chosen from instead.
What if companies could break free from the reinforcement cycle—and recommend categories and products that felt new and different, yet familiar at the same time? And what if these algorithms performed like a personal shopper who knows your individual unique tastes and what you’re in the mood for?
Personalization focused on discovery will improve brand loyalty
Traditional personalization algorithms can easily identify the products that are closest to past purchases. Only the machine that draws on deep wells of data and synthesizes insights from vast reams of product information can create magical discovery moments in the customer’s journey; the moment you present someone with a perfect product they never knew existed, exactly when they need it.
The rewards for any company that can do this are clear, especially concerning trust and brand loyalty. Successfully recommending a personalized product that delights the customer and takes them out of their comfort zone, puts a brand on par with a close friend or personal shopper.
Customer-focused personalization leads to increased life-time value
Let’s say a customer has been browsing tennis clothing on a sporting website. Machine learning could synthesize all the data about this customer—past searches, age, location, interests, and seasonality—and discover truths about them.
A premium machine learning algorithm might determine that the customer doesn’t want just tennis clothing to wear at their local court; they want an entirely new experience, like a week-long tennis vacation.
Maybe the customer isn’t interested in a tennis vacation. In this case, they reject the offer and the machine learns from their action. Or it was the vacation they’ve always wanted and the personalized recommendation came along at just the right moment.
Better yet, maybe by harnessing the power of machine learning, your brand just introduced the idea to the customer, providing that magical moment where it takes on the role of the genie or the wise counsellor in their individual buying journey.
Exceed customer’s expectations with advanced machine learning personalization
A recent Salesforce report revealed that 52 percent of consumers would abandon a brand that didn’t personalize communication for one that did, while 65 percent said that personalization influences their brand loyalty.
Today’s customers are willing to give their data to trusted brands in exchange for highly personalized experiences. Customers expect brands to understand what they’re in the mood for today, how their tastes change over time, and to be introduced to new products that speak directly to them at the right moment. Only advanced machine learning algorithms that can manage a series of interactions through time are able to provide the level of personalization customers now demand.
Aside from the financial rewards, the biggest long-term benefit in achieving this level of personalization is brand reputation and loyalty. By delivering customer experiences that continually delight, customer engagement increases and a lifelong brand relationship develops.
The virtuous cycle that happier customers create—where frequent engagement leads to better value which leads to higher conversion rates—will let companies maximize their lifetime customer value. Your customers may not know it, but within their data lies the secret of what they really want. Your job now is simply to show them.
Coherent Path provides predictive analytics software designed to surface products and categories that meet consumers’ evolving needs over time.
Favorite