Personalization long ago reached buzzword status, yet customer demands for hyper-personalized interactions have yet to become a reality. And that’s a problem. After all, if you are collecting customers’ personal and behavioral information as they interact with your brand, it’s fair that they expect you to return the favor with a personalized, highly relevant experience.

Businesses turn to vendors from a wide variety of categories to deliver on this demand, including (but not limited to): customer data platforms, CMSs, personalization engines, digital experience platforms, ecommerce systems and more.

Scott Brinker's martech landscape supergraphic breaks down the 8000 plus marketing technology applications listed into six high-level categories. The two primary categories relevant to the personalization conversation are "Content and Experience" and "Data." Data gives us the understanding of the customer, and the Content and Experience category covers the ability to create, manage and deliver content to your audience.

personalized content

Personalization happens at the intersection of Data and Content. It is in the careful orchestration of customer data and content that truly meaningful experiences are born. Personalization, Optimization and Testing is the martech category that aims to bridge this gap, however personalization engines rarely succeed due to the wide distance between data and content and its complete separation from the data side of the equation.

Vendor Categories Are a Manifestation of Organizational Boundaries

When executives are researching vendors, they look for software providers that fit into their personal domain and world view. So you have Technology (lead by a CTO), Marketing (lead by a CMO) and Sales (lead by the Head of Commerce). When thinking about the ownership of Data vs. Content and Experience, in a traditional organization, IT would own data as well as relationships with technology vendors, while Marketing would own the process of creating and delivering content and experience.

This can reduce software selection to a domain or territorial question rather than being rooted in the desired outcome — to provide a personalized and relevant customer experience. This kind of territorialism makes it difficult for teams to work towards the final goal of providing the best possible experience to the customer.

With this in mind, let’s dig deeper on the current state of each side so we can arrive at a way to bring them together.

Related Article: Why Personalization Efforts Fail

The Current State of ‘Data’

Customer data comes from the variety of sources and channels a customer interacts with over the course of their journey with a brand.  Organizations collect this data and ideally use it to better understand the customer with the end goal of providing meaningful experiences to them, such as personalized offers, messaging, products and content. CDPs have recently ascended in popularity as they profess the ability to create a unified 360 view of your customer across all channels of engagement, allowing data analysts and marketers to create complex audience segments to power highly targeted and personalized cross-channel marketing campaigns. 

The integration between a CDP and the remainder of your marketing stack is typically limited to sharing of audience segments. This is a result of the territorialism we noted above: IT/Data teams are in charge of stitching the raw data and marketing only gets involved when the unified customer profile is ready for segmentation. 

The Obvious Gap Between Understanding and Execution

While audience segments are useful and do a good job capturing the various types of customers you want to target your marketing efforts towards, they fall short of 1:1 personalization or hyper-personalization. It also leaves it to each of the marketing platforms to work it out on their own what content should be delivered to who, making it difficult to create a unified and consistent experience through all channels. The result is you may be sending customers something different through email than what they see on your website.

At the end of the day, creating a unified customer profile is only a means to an end. The end goal again is to provide better experiences to your customer to improve your bottom line, boost conversions and increase loyalty and engagement — you know, all the impressive outcomes you hear every MarTech vendor tout on their marketing materials.

The question is, how do you go from collecting, unifying, segmenting and analyzing tons of customer data to delivering a more meaningful experience? How do you activate this data to offer relevant and personalized content to your customer? In order to answer these questions, let’s pay a short visit to the world of content.

Related Article: Curiouser and Curiouser: Drawing the Line Between DXP and CDP

Treat Your Content Like Your Data

A lot of emphasis has been placed on customer data and ways to extract insights from it. Every CDP and personalization vendor talks about unifying, enriching and segmenting customer data. However, content is often treated as a second-class citizen, left for individual channel owners to figure out. 

Just as we talk about creating a unified 360 view of the customer by combining data from multiple sources, content too needs to be pulled together from the multiple siloed CMSs, DAMs, PIMs, ecommerce systems, portals and other systems it resides in. Currently it remains locked in each channel and only works for narrow use cases. It’s not an option to migrate all of your content into a single CMS or repository because you don’t own that content and it would be highly disruptive to throw away existing workflows, not to mention extremely expensive and time consuming. So, what should we do? Well, if you can create a CDP to unify customer data, there is no reason why you wouldn’t have a similar approach to content. This would be your content 360.

Similar to customer attributes and traits, you have content metadata that describes the nature and characteristics of content. If you pay attention to it, it will give you signals about who would be interested in this content as well as when and where it should be served.  As an example, unless a movie record is tagged with the director, genre, year, actors, etc, you can’t determine what kind of customers would enjoy such a movie.  Sometimes, content comes with rich metadata and other times, work is required to make it so.

If you are tracking your customer, you should be tracking what content they are viewing, clicking and converting on. Just as there is a need for marketing analytics, there is a place for content analytics. You should be able to enrich your content with insights about how each content or catalog item is being perceived by your customer. For example, content analytics would show what segments of your customer are interested in this content, has the conversion on this content gone up or down in a given time period, in which geographies is the uptake of this content item higher or lower, etc.

Even if you are using machine learning to predict what content would be relevant to what customer at what point, you can’t do that without having a clear map of content metadata.  Machine learning algorithms are not magical — they identify user behavior patterns by determining what features (or attributes) result in an outcome such as interest in a category, or genre, or actor, etc. The algorithm is essentially finding patterns in the user’s behavior and predicting content metadata that best describes the kind of content that the customer would find relevant based on their individual attributes and traits. This process gives life to the content recommendations we're all familiar with, such as "People who liked this movie also liked," "Similar articles read by people with your taste" and more. 

Learning Opportunities

Related Article: 5 Drivers of Personalized Experiences: A Walk Through the AI Food Chain

Existing Personalization Engines Fall Short of the Promise

Over 200 vendors can be found in the martech supergraphic's "Personalization, Optimization and Testing" category and both Gartner and Forrestor have introduced their own categories to encompass this idea of delivering meaningful experiences for the customer. Some are hybrid CDPs, some started off as DXPs, others arose from ecommerce technologies. Many vendors claim to be "personalization engines," but never built the foundation needed to truly personalize the customer’s experience — in other words, to understand the customer, create a holistic, semantic view of the siloed content sources and then connect the dots between data and content repositories with a vendor-agnostic orchestration engine. 

Related Article: Decisioning vs. Orchestration: What's the Difference?

A Fragmented Technology Stack Demands Orchestration

Vendor-agnostic orchestration is key because organizations own applications and systems from multiple vendors and need to power experiences in a myriad of channels. You can expect to see one or more CMS, ERP, PIM, CDP, DMP, ecommerse platform to support the multiple digital applications (website, mobile application, chatbots, IoT, etc) your customers will interact with. For a unified experience across all digital channels to happen, the orchestration engine needs to be able to connect all these disparate systems without requiring a migration to a single vendor suite.

A real-time customer interaction (such as a purchase) in one channel is relevant to that customer's experience in another channel (self-help through a chatbot). You can’t wait for a batch process within your CDP that slots customers into audience segments, then ask your IT team to share the customer list with you so you can update the logic your chatbot uses to respond to the customer about their recent purchase.  You need to act in the moment when it matters.

The fragmentation of technologies is one of the biggest reasons almost none of the current vendors truly deliver on their promise of personalization.

When selecting a vendor for personalization, organizations should pay close attention to the vendor’s approach. Ask "How exactly do you do this and can you support my current technology stack?" Any claim that "we use AI algorithms to determine the next best action or recommend relevant content" isn't enough. 

The following capabilities are what I believe are the essential components of an experience orchestration stack:

Data: Understand your Customer

  • Stitch together customer data sources to create a unified, deduplicated customer profile across all channels.
  • Allow marketers to create and apply audience segmentation in real-time using a combination of rule-based and AI-powered tools.
  • Ensure that the customer profiles are available in real-time through secure APIs to experience delivery applications such as ESPs, DXPs, CMSs, ecommerce, chatbots, etc.

Content: Build a Content 360

  • Unify content from PIMs, CMSs, ecommerce systems for product catalogs, marketing content, knowledge articles, offers, promotions, into a centralized content hub that marketing teams have full visibility into. Note that this does not mean migrating your content into a new CMS. It means creating a semantic view of your content from all sources. Most existing DXP vendors do not provide this capability.
  • Allow content owners and marketers to classify, standardize and enrich content through rule-based and AI-powered techniques.
  • Get a clear view of content analytics or campaign performance broken down by customer segments, geography, time ranges, etc.
  • Headless APIs for real-time access to the content hub so all your applications get their content from a single source. Warning: This is not the same as a headless CMS because we are not trying to migrate content to a single source of record.

Personalization: Deliver the Right Content to the Right Customer at the Right Time

  • An API-first central intelligence layer, built on modern architectural principles, that guides the rest of your digital marketing stack.
  • Ability to determine the next best action and trigger these action on other systems through API invocation based on user activity or system events.
  • An AI engine that creates a dynamic relationship graph between content and customer profiles that is continuously evolving based on ongoing customer interactions with the brand.
  • Tightly connected with your unified content hub allowing digital and marketing teams to assemble template driven content blocks that can be consumed by any application or digital property in real-time. 
  • A rules engine capable of taking inputs such as customer’s demographic, psychographic and behavioral information, customer’s real-time context such as their recent activity, time of day, weather, and dynamically return content based on its metadata and classification.

Break the Habit of Separating Technologies by Organizational Boundaries

Most of the vendors listed in the DXP, CDP or personalization engine categories don’t provide even 70% of these capabilities. The capabilities also can’t be neatly separated into marketing, technology, data or digital. An ongoing collaboration between subject matter experts and owners in each of these domains is the only way this will move forward.

If we continue to separate technologies in ways that parallel organizational structures, we’ll never truly achieve a comprehensive and holistic approach to customer experience. It is time to close the chasm between business and technology. It is time to build teams centered around the customer experience and then reorganize and refactor our technology categories accordingly.

fa-solid fa-hand-paper Learn how you can join our contributor community.