Last week, James McCormick, Principal Analyst at Forrester Research, was a guest presenter for a webinar with our CEO, Karl Wirth, entitled “Digital Intelligence: The Key to Delivering Personalized Customer Experiences.” In the webinar, they explored a key challenge facing marketers across industries: the experience delivery gap, originally identified by Bain and Company back in 2005. Back then, Bain discovered that while 80% of the companies they surveyed believed they were delivering a superior proposition to their customers, only 8% of those companies’ customers actually agreed with them.

While a lot has changed since then, the experience delivery gap has not yet been closed. The data has shown that consumers expect personalization — in part due to the personalized experiences they are receiving from some of the bigger names on the Internet such as Amazon, Netflix and Spotify. In our recent research, we found that while over 70% of companies are using personalization in some form, most are using it for email personalization. Usage drops off when you explore other channels: 57% are using personalization for websites, 28% are using it for mobile web, and only 18% are using it for mobile apps. So marketers are not providing personalization at the level that consumers expect, resulting in a continued experience delivery gap.

In the webinar, James McCormick suggested that in order to effectively deliver personalization, you must have a good understanding of what “personalization” means. He presented the Forrester Research definition of personalization:

“An experience that uses customer data and understanding to frame, guide, extend, and enhance interactions based on that person’s history, preferences, context, and intent.”

This is an excellent and comprehensive definition. In this blog post, let’s unpack that definition — as each piece is critical.

Customer Data and Understanding

The definition begins with “customer data and understanding.” Why? Because you cannot provide a personalized experience for each individual unless you fully understand that individual. That understanding comes from the data.

But what specific data types are useful for personalization? Any data that gives a marketer insight into an individual can be used. All data needs to be available at the individual level, as opposed to the aggregate level, in order to be useful for personalization. For example, it is only relevant to know that one particular person is located in San Francisco right now rather than to know that you have had 50 visitors from San Francisco on your site over some period of time.

Typically, two types of data are used: attribute data and behavioral data.

Attributes describe a characteristic of a visitor. Those characteristics can either be detected automatically from the web (such as a person’s geolocation, referral source, company name, industry, device type, etc.), or they can be pulled in from any database (such as a CRM, marketing automation platform, data warehouse, etc.).

Behaviors describe any action a person has taken on any of your channels (website, web or mobile app, emails, etc.). This can include past behaviors a person as taken, as well as behaviors that they take in their current session. For more information on data types for personalization, check out this blog post.

Frame, Guide, Extend, and Enhance Interactions 

When all of this individualized information is stored in a central place and available in real time, it can be used to “frame, guide, extend and enhance interactions” you have with each individual. The two main ways you can personalize an experience for an individual are through rules and machine-learning algorithms.

Rule-based personalization allows marketers to deliver experiences to specific groups or segments of people based on the manual creation and manipulation of business rules. The simplest way to think about it is in the form of if/then statements. IF a person fits into a specific segment, THEN show her the appropriate corresponding experience.

For example, Dyn has used rule-based personalization to modify its navigation menu in real time when a small business visitor (one of its key segments) lands on the site. These visitors see a slimmed down version of the menu, with small business/consumer products prominently displayed. With this simple personalized experience, Dyn was able to help the small business/consumer segment quickly identify that Dyn could meet its needs — driving a 7% lift in conversions.

definition of personalization

Machine-learning personalization utilizes algorithms and predictive analytics to dynamically
present the most relevant content or experience for each and every visitor. It provides a more scalable way to achieve unique experiences for individuals, rather than segments of people.

For example, Zumiez provides true one-to-one experiences throughout its site – from the homepage to checkout. Zumiez responds to each shopper’s preferences to guide which brands they see, the products they discover, and the content they are served. These one-to-one experiences are driven by machine-learning algorithms that take a shopper’s preferred brands and styles into account. For Zumiez, shoppers who engage with these personalized experiences convert to a purchase 2.7 times more often than those who do not and also spend 2% more per order.

History, Preferences, Context, and Intent

To deliver these types of true one-to-one experiences, you need more than just the data. You need the ability to analyze and interpret that data to turn it into intelligence that you can act on.

Tracking what a person does on a site in real time is important, including the ability to combine current session behavior with any historical behavior. But context is the key to transforming data into insight. If you just have the data itself, you may only be able to understand that someone visited seven pages on a website. But when you have context, you can understand which categories, tags, brands, colors, keywords, content types, etc. those pages represent — which provides insight into that person’s true preferences and intent.

For example, a technology provider could analyze a prospect’s engagement with those seven pages — how much time he spent with them, what his scrolling and mouse movement looked like, etc. — and turn it into insights about that prospect’s preferred use cases, content types, keywords, etc. Those preferences can be used to provide relevant content recommendations across the site.

And finally, after you know a person’s preferences, you need to identify his intent in-the-moment so that you can understand what he is browsing for right now. If that prospect is browsing case studies, after spending his last several visits reading blog posts, he may be further into his journey and close to being ready to speak with a salesperson. This type of real-time insight can help you deliver those enhanced interactions as well.

Final Thoughts

During the webinar, James said that “to personalize at scale we must link our systems of insight with our system of engagement.” This means that in order to effectively achieve personalization as defined in this blog post, you have to combine data with action — the ability to collect data and the ability to act on that data in the moment.

To discover more personalization insights from James McCormick from Forrester Research and Karl Wirth from Evergage, watch the webinar replay. You can also download the recent Forrester Wave report to get a complete, unbiased view into the leading digital intelligence platforms and a helpful guide for evaluating Evergage and other vendor solutions.