personalization technology for 2019

Believe it or not, 2019 is almost here! As you think through your marketing and customer communication strategies for the coming year, personalization is undoubtedly top-of-mind for you (after all, we found in our annual survey that 88% of marketers believe their prospects and customers expect a personalized experience). But a personalization strategy is only as good as the technology powering it. So whether you’re a personalization pro or you’re just beginning to explore its potential in 2019, you need the right technology and underlying platform to execute an effective personalization program.

In this blog post, I’ll describe a few of the capabilities you should look for in personalization technology for 2019. There are, of course, many features you’ll need to meet your company’s specific requirements, but this list should cover the main areas of concern. Let’s dive in!

Your personalization platform should address the following five key requirements:

1. Bring together customer and prospect data from multiple sources into single, unified profiles

The key to delivering a truly personalized experience is a deep understanding of each person you interact with. And how do you understand someone really well? The answer, of course, is with data. A good personalization solution should be able to collect plenty of data on its own without having to bring in lots of additional data from outside sources (I’ll get into more on this later), but I’m sure you can think of a few data sources from other systems across your organization that you’ll want to use to inform the personalized experiences you deliver.

As a result, your personalization solution can’t exist in isolation and needs to be a customer data platform (CDP) at its core. It should allow you to store all relevant customer data in a single place, with a single unified profile for each person (and account for B2B companies). That profile will contain all the data the personalization solution will use to determine the most relevant experience for each person.

As you evaluate personalization technology, pay attention to how the solution connects to other systems and how it stores the data. It will save you a lot of headaches in the future if you think about data aggregation and integrations upfront.

2. Collect and analyze deep behavioral data

As I alluded to in the previous section, a good personalization solution can collect data on its own — it shouldn’t just rely on data from other sources. This allows it to deliver a personalized experience to anonymous and/or first-time visitors who don’t exist in any of your other systems yet. And it can act on any data it collects about someone, including anonymous and known visitors/users, the moment it learns it.

Make sure you look into what data is collected by the personalization technology you’re considering. It should be able to collect data to help you understand what a person’s true interests and intent are based on past and in-the-moment activity. In other words, it should collect in-depth behavioral data. And I don’t just mean which pages someone has clicked on — I also mean what behaviors were taken or not taken on those pages. That includes time spent on a page, mouse movement, scrolling, hovering, inactivity, etc.

And, of course, all of this data means nothing if it can’t be analyzed and automatically interpreted to create a clear picture of the individual. A good personalization solution should be able to combine this in-depth behavioral data with in-page contextual data (content or product categories, tags, brands, keywords, etc.) and apply machine learning-driven analysis to provide an accurate indication of someone’s affinities, interests and intent. This type of information is critical to maximally relevant 1-to-1 personalization.

3. Personalize across channels from a single platform

While the importance of data can’t be overstated (personalization based on bad information isn’t really personalization at all), activation is the ultimate goal. You need a personalization solution that can actually use all of the information collected and aggregated to deliver relevant 1-to-1 experiences.

During the planning phase, it’s important to identify which channels you want to personalize. Think broadly: the personalization “umbrella” covers exit and cart abandonment messages, segment-based communications, product and content recommendations, account- and industry-specific experiences, web and mobile app messages, push notifications, triggered emails, bulk emails, digital advertising, search, and more. And while individual solutions exist to personalize specific aspects of specific channels, you may want to consider a platform that can deliver personalization more broadly across many channels.

4. Use machine learning to identify the best experience for each individual

A personalized experience can be delivered via rules and/or algorithms. With rules, you manually define which group (or segment) of people will see a specific experience. With algorithms, you let machine learning decide which experience to show each individual person.

There are many occasions when you want to manually set rules (and your personalization solution should allow you to create and prioritize as many rules as you wish) — but they can become labor-intensive and cannot be used for 1-to-1 personalization. If you’d like to deliver relevant product, content, category or brand recommendations; ensure lists and search results are sorted in a relevant way for each person; or pick the most relevant promotion that has the highest potential value to your company — you need to use machine-learning algorithms.

Make sure that you understand how the algorithms in your personalization solution work. Are they customizable or do they operate in a black box? This is an important factor in successful 1-to-1 personalization, so you want to be able to control them.

5. Combine A/B testing and personalization

Just because you’re delivering a personalized experience to segments or individuals, you don’t throw your principles of A/B and multivariate testing out the window. You just need to find a solution that allows you to combine your testing and personalization approaches. It no longer makes sense to test to find the single experience that works for everyone when you can find the experience that works for each group or even each individual (read more about this concept in this blog post).

Find a personalization solution that lets you combine testing and personalization. Your personalization technology should let you test multiple experiences to find which one works best for each segment. It should let you test and tweak different algorithms to find the one that performs best — it should even let you test different algorithms for different segments. And it should let you to set up algorithms that continually test and learn in order to select the ideal experience for each individual and for the company.

With effective testing, you’ll be able to continue to iterate and deliver better customer experiences in 2019 and beyond.

Final Thoughts

As with any piece of technology, there is a lot to consider. With something as important as your customer experience on the line, it’s worth the time to pick the right solution.

As you evaluate your personalization technology needs for 2019, keep this list in mind. You should look for a solution that offers a single unified platform, in-depth behavioral data, omnichannel capabilities, customizable machine-learning algorithms, and integrated and comprehensive testing — these are all critical components of a successful personalization platform.