The Importance of Bringing Content and Context Together
by Dave Ingram
Building a platform and strategy that will take your digital experiences into the next 5 years means stepping into uncharted territories. No one knows the devices or channels that will be most popular in the future. With the pace of technology growth, predicting what you’ll need or what will be the preferred device in even a few years is incredibly difficult. While today you may be thinking about mobile applications and websites, a few years from now augmented and virtual reality could be your largest growth potential.
Fortunately, we’re not starting off blind. We know from early tech pioneers the kinds of problems that we'll face and the terrain we’ll be on. But the first step on any successful journey is to make sure you know exactly where you’re going and what you’ll need for the trip.
Preparing for the Destination
For personalization, the “destination” is when every customer, regardless of the device they’re using, receives the content that makes them most likely to buy, engage or come back for more at just the right time. Everybody in the personalization space talks about this, but no one is actually there yet. The pioneers who come the closest are the heavyweights in the tech industry like Google, Facebook, Netflix and Amazon. They have teams of data scientists and engineers working on this problem, yet they still have a long way to go.
One aspect of the destination we can be completely sure of is that it involves showing content to users. Content refers to anything a user will see on a website or any other channel; from products to technical documentation, from the footer to images. When you add personalization into the mix, now you’re potentially dealing with the need for individualized content that fuels tailored experiences for different segments and user profiles.
Managing dozens if not hundreds of siloed data sources, each one duplicating content for a different channel, will simply not work. Knowing the complexity that awaits us, and generally what it will look like, allows us to bring the right tools in the right configuration so that we can make it through this uncharted territory unscathed.
Organizing and Utilizing Customer Data
Over the past several years, organizations have become increasingly aware of the power of understanding and leveraging customer data. This is evidenced by the proliferation of data related marketing tools and the push by leading content management platforms into data services. Furthermore, the rapid improvement in tools related to very large data sets such as the Hadoop and Apache Spark ecosystems, as well as AWS and other cloud services, have helped to accelerate the adoption of powerful data tools and also increased awareness of the possibilities to the technology industry at large.
One obvious trend that stems from the adoption of these tools is that in order to reap the full benefits of a data related strategy, data from around the organization must be centralized and normalized. This means that data coming from every touch point and every moment that an individual interacts with an organization must be brought into the same data store so that it can be analyzed and acted upon. This data store must be sufficiently flexible to meet this need, and it must be able to scale rapidly in support of millions if not billions of interactions per day.
Two general approaches have emerged to this problem from an architectural standpoint. One is the creation of a technology stack often referred to as a “data lake”, which can take all data from disparate sources and make it accessible to analysts or data scientists in order to discover the insights needed by the organization. The second approach is to normalize data at the time of collection. Various solutions have emerged for this task which provide APIs, integrations and aggregations.
Either approach can be beneficial for analytics, but the latter is preferred when building a system that is meant to adapt to user needs in real time. This is because structured data can be fed into a personalization system as well as better understood by predictive algorithms. For this reason, our future ready personalization toolset includes a rich, structured and highly flexible data store, or Unified Customer Profile, which provides the context of users at any given time.
Sustainable Content Management
Many personalization and recommendation systems today fail to take into account the central importance of well structured content and the challenge of maintaining that content. As a result, only deeply technical individuals can make changes, or business users are severely limited in the scope of what they can accomplish. In either case, the end result can suffer significantly.
In the case of automated recommendation, a compromise is often made by simply using algorithms that attempt to extract the appropriate data from the content or that use other individuals’ signals such as views or shares to derive their recommendations. As examples, collaborative and content based recommendations can do a fair job, but they are always incomplete. For example, a search engine rarely has access to deep semantic knowledge of the content, and while the semantic web movement is making strides at fixing this problem, deep knowledge of the content is often still missing.
The other common approach by personalization vendors is to provide What You See Is What You Get (WYSIWYG) editors that overlay on top of a page. These tools are easy to get started with, but the results are unsustainable. Not only do you have two completely different editing systems (one you created the base page with, the other for creating variations), but it also becomes very difficult to understand what a page might look like for a given user.
These are all problems that can be addressed through deeper integration with content management systems such as Drupal.
A successful journey begins with planning. A great architect begins with a firm foundation. While the ideal of 1-to-1 personalization has yet to be realized, we can all see it getting closer. To prepare your organization for what’s to come, you need to plan carefully for how your content and your user’s context can be integrated together in a central place. From there, you can start incrementally building out a personalization team and program, knowing that you can scale into the future without the foundation collapsing. It’s an exciting future, and we should all be ready.