Author: Gabriella Vas
Personalization has become such an integral part of our lives that sometimes we’re not even aware that our experience has been personalized; we just take it for granted.
Remember that movie you streamed the other day based on a Netflix recommendation? Or the pair of sneakers that caught your eye just as you were about to proceed to checkout, and you ended up adding it to your cart? The book, the album, the holiday home, the Tinder date you “stumbled across”?
What exactly is at play? How does the magic happen? Let’s take a detailed look at personalization and the artificial intelligence behind it, also known as personalization engines.
There’s more to personalization than newsletters that address you by your first name. In order to fully understand what it is, let’s rewind back to the beginning.
Recommendations have been a thing ever since the first shoemakers set up shop in Florence. But there’s a major difference between general recommendations, basically pointing to bestsellers (“These ones are selling like hot cakes right now”) and personalized ones (“Wait, let me show you something I know you’ll love”). For one thing, these latter are more efficient. Furthermore, they ensure a more even turnaround of stock, as not everyone is after the same popular items.
That’s why, a few centuries later, computerized recommendation systems were invented: to come up with suggestions that a certain person will find helpful in making a decision, such as what to buy, where to dine, and so on. They do this by modelling the person’s taste and trying to predict her responses. And because recommendation engines are able to process reams of data in the blink of an eye, they can evaluate way more factors than any human being, resulting in more nuanced recommendations.
One level up, personalization is an even more sophisticated process that aims to enhance all or many aspects of an interaction to make it more relevant to the user. Though recommendations remain at the heart of this function, it extends to include interfaces – what this person will see on her screen, how it’s ordered, ranked, filtered; the look and feel of it all – and reactions to her moves, like reminders to continue shopping if she has abandoned her cart.
All this is achieved through insights built on data specific to the user, or derived from the choices of people like her, in order to create an experience that meets her precise preferences.
It’s no surprise, then, that personalization at scale has a massive potential to create new value – McKinsey Global Institute puts it at 1.7 to 3 trillion dollars and that’s not just profit for brands offering personalized services, but also savings and added value for consumers.
Categorized as Software-as-a-Service, a personalization engine is a digital solution that defines a person’s best possible experience in a given context and acts accordingly. It changes the appearance of an online interface, or triggers an automatic response. Alternatively, it provides analysis to those in charge of running the given service (say, an e-retailer or a media streaming platform) to help them decide on the best way to serve this person.
To put it very simply, a personalization engine does three things. First, it collects data:
Next, the personalization engine applies an algorithm (or several) to the data. This is where it gets complicated – for the purpose of this article, let’s just say an algorithm is a formula, a set of rules to be followed in a calculation in order to estimate an unknown outcome. Each personalization engine provider has its unique portfolio of algorithms – this is what sets them apart in terms of performance.
Finally, having put the data through the algorithm grinder, the personalization engine produces an output that will inform the user’s decision about what products to buy or what content to consume. All this happens in a matter of milliseconds, so people get the impression that their options are being served up in real time.
It’s easy to see that personalization can be put to good use in a wide range of industries, from e-retail through banking and insurance to telco, travel, healthcare, and media platforms. In fact, 64 percent of respondents in Gartner’s 2019 Marketing Technology Survey said they’re using a personalization system.
Only a few decades back, you had a problem and you wondered if there was a solution; you needed a product and you weren’t sure it was available. Today, the abundance of options at our disposal has become our greatest hurdle.
The more choice we have, the harder it is to choose: we’re aware that settling for one option means deselecting all the other possibilities, and the resulting anxiety often prevents us from making any decision at all. It’s called choice overload: research has shown that the human brain struggles to choose from more than a dozen items.
Search engines have come to the rescue. But, as anyone who has tried Googling something more complex knows, it’s often difficult to formulate our questions to ourselves, let alone to the computer. And, even in case of a simpler search, the sheer amount of results is overwhelming.
Recommendation engines are often better at identifying users’ needs than the users themselves. As a consequence, the alternatives they serve up are more likely to be relevant than in the case of basic search.
Personalization engines take this one notch higher by anticipating users’ needs, and by going beyond recommendations to enhancing every aspect of their experience, every step of their journey.
From a consumer’s point of view, the main benefits of personalization are saving precious time and effort. Personalization shaves off those mind-boggling minutes of confusion when you first encounter the seemingly endless array of products or content on a site. It cuts back on time spent browsing aimlessly by pointing you to potentially interesting stuff. Most importantly, it reduces the number of your options to a manageable amount.
What’s more, good personalization finds the right balance between spot-on and variety; it doesn’t enclose you in a so-called filter bubble.
No wonder 91 percent of consumers are more likely to shop with brands who recognize and remember them and so provide relevant offers and recommendations, according to Accenture’s Pulse Check 2018. In fact, people have come to expect a decent level of personalization at webstores, content portals, and the like. A man shopping for shoes online would probably be nonplussed if he got recommendations for ballerina flats.
For companies implementing personalization, its main advantage is being able to sell more products. It’s a no-brainer, really: people are more likely to buy or consume stuff that’s relevant to them. As mentioned earlier, personalization also makes it possible to sell a more diverse set of products, as it matches up even the most obscure, back-of-the-shelf items with someone who wants just those.
All this adds up to increased customer satisfaction and, in turn, loyalty. As a result, personalization can cut acquisition costs by up to 50 percent, grow revenues by up to 15 percent, and boost the return on marketing investment by as much as 30 percent.
Case in point: as a traditional brick-and-mortar retailer, Cora Romania badly needed to transition to digital in order to keep up with the competition. The personalization solutions engineered by Yusp transformed Cora’ website, mobile app and loyalty program reaching consumers both online and offline, and ended up making more money than what it cost to implement them. Personalizing the website alone resulted in a 10 percent increase in conversion (that is, 10 percent more folks decided to buy, not just browse) and a 7 percent growth in total revenue.
One of the crucial challenges of personalization is being able to serve users one-on-one and yet shape the experience of millions at any given moment. India’s largest premium video streaming platform, Hotstar boasts up to 100 million daily active users. The media giant decided to test the personalization system put in place by Yusp against their existing provider.
Yusp outperformed its competitor in all metrics, attaining a 50 percent increase in click-through and 30 percent more watch-time. Even more importantly, Yusp scaled up flexibly to serve Hotstar’s massive user base, crafting one billion personalized recommendations every day.
There you have it. Now you know how personalization evolved from simple recommendations. You’ve caught a glimpse of the science behind the process. You’ve discovered how the mechanism of streamlining available choices affects the lives of consumers and the livelihood of brands worldwide. You’ve seen examples of the power of personalization put in place by Yusp. Read on – we have more case studies here.