Vendors Take Note! 10 Key Features Missing from Nearly Every Personalization Platform

March 14, 2016 | John Berndt

question mark surrounded by toolsWe are at an unprecedented time for the design of digital user experiences—one where it has finally become operationally possible to plan experience not just based on the user’s last click but on a much richer panorama of information known about the user and also to do so without specialized programming skills. How far this approach will be adopted is largely a function of a digital team’s ability to effectively have a vision for these new experiences and to efficiently manage them in an operational context. But there is a third important limiting factor, which is the subject of this post.

While we have long been fans of the features of personalization platforms (everything from Sitecore, to Acquia Lift for Drupal, to 3rd party optimizers like Monetate and Thunderhead) much of this software seems to be very long on raw features and capabilities and very short on refinements and features that allow for frequent usage and long-term usability. 

A Personalization Platform Walks into a Bar… 

In fact, around TBG and with our clients, it has become a standard joke to say “you know, none of these systems were actually designed to be used”—an overstatement that captures the fact that most systems seem to be built for occasional use, either in the context of SEM landing pages or eCommerce optimization but without much thought to broader operational issues. With vanishingly rare exceptions, the systems today simply lack most of the features that would make them more viable for real integration with the mainstream of digital activity. One assumes this is for two reasons—firstly, that these systems are relatively new and haven’t been “beaten up” by users, and secondly, because they tend to be used in very limited ways—a problem that may be the circular result of usability decisions by vendors. 

Vendors Take Note! 

So, vendors take note! Here are TBG’s top 10 missing features from personalization systems—the features that would make these systems truly exceptional and give them the best chance of being adopted in more meaningful ways. 

1. Easy Personalization Proofing 

Ability to easily simulate user states within the same UX as they are personalized and in the context of the site. The personalization manager should be able to pick from exactly the same variables that are referenced within rules and edit various states to see the results at the page level. Currently, a few systems have separate utilities that allow simulation of a subset of user variables, but none of them is truly complete, easy to use, or integrated enough. This feature is crucial to be able to make sure personalization is working and also to be able to catch incongruous effects of personalization within the UX. 

2. Personalization-Centric Views

Ability to review an automated summary of all personalized experiences within a site, in an organized way, and with their page-level collisions highlighted. Current systems, almost without exception, attach personalization rules to the user experience components of pages (portlets on the page), which makes sense for designing experience but misses a crucial use case for managers—the need to simultaneously see all the places in a digital property currently running personalization from a top down perspective. A system should be able to derive that report with the ability to both link directly to the page in question and make changes in the context of the UX. Without this ability, the effort to keep track of larger amounts of personalization in a site can only be the province of hard-to-read and hard-to-use spreadsheets—a potentially huge operational barrier.  

3. An Independent “File System”

Ability to self-organize and categorize personalized experiences outside of site structure, as well as the permissions and workflows that characterize group collaboration environments. Building on the previous item, a best case system should permit users to create their own custom organization of personalization with a utility similar to a modern file system—the ability to organize a personal “drawer” of folders and sub-folders, easing browsing and accessibility. With the addition of an ad hoc classification (categorization), individual users’ permissions, and a sophisticated search application, the resulting system would be versatile and scalable.

4. Group Tracking

Ability to “group track” a set of personalized experiences as a composite, named “personalization program” (or campaign, if you like). For the purpose of tracking outcomes, most systems allow the user to easily create a report between a particular element of personalization and conversions (or other metrics). While this is the base case, what about situations where a more complex “campaign” of multi-part personalization has been deployed? Some systems address this, but the ability to be able to flexibly group personalization elements together and track their effectiveness as a group has broad application.

5. Tracking Rules That Are Triggered

Tracking outcomes in relation to what personalized experiences have been triggered—not just “potential” personalization. Almost all systems assume that what managers want to see is “since we implemented personalization X, showing to some users and not to others, how did overall conversions change?” In fact, though that is useful, it is more useful to understand “of the people who saw personalization X in action, how many of them converted and what was the yield, that is, the difference between those who converted and for whom it didn’t fire.” These are far better statistics to understand the impact of personalization.

6. Rule Modularity

Easy reuse of personalization rules. In a sophisticated personalization operation, experimentation with rules and their interaction with different UX areas is at a premium. The ability to easily and quickly redeploy existing rule sets into new areas of the UX (optimally by dragging and dropping), or to mix and match existing rules, would enable modularity. Most systems don’t treat complex rules as “objects” and require that they be recreated from scratch in each instance.

7. Global Personalized Elements

Provision for personalizing elements across multiple pages in a site or sites. Most systems hang personalization off of individual UX components on a specific page, but in many cases, personalization is most effective when it appears globally across a site, even across different kinds of templates. All systems allow this to be implemented in a very manual way, but the ability to recognize a global element and personalize it with a single click would be highly helpful.

8. Personalization Frequency

Allowing personalization rules to run inconsistently for a user, so that the frequency of personalized experience can be globally adjusted—optimally at either the audience segment level or the individually personalized experience. In another blog post, I wrote about the dangers of over-personalization, where too much obvious personalization creates a claustrophobic experience, one that can be a creepy turn-off for the user. This can occur for a variety of reasons and always needs to be factored into personalization practice. A powerful feature to combat this would be the ability to control how often personalization triggers—allowing a fair amount of personalization to be active within a site but without overwhelming the experience for the user. Of course, this creates problems for tracking effectiveness, but it has the potential to allow personalization of various kinds to co-exist with non-personalized versions in a more balanced experience.

9. An Audit Trail

Accessibility to the history of changing personalization configurations over time with rollback. For large amounts of personalization that changes over time, the ability to see the history of experiments and changes can be highly important, not the least of which is the correlation of changes in rules to changes in analytics. It would be great if personalization systems would treat rules (and instances of rules) in the same way that content management systems treat content, which would provide a lucid, usable audit trail of changes.

10. Redeployment

Ability to export (serialize) personalization rules (and audience definitions) as XML or a similar format, so they can be captured by concurrent versioning. Following on the audit trail above, a simple serialization of personalization rules into files that could be captured in concurrent versions systems (CVS, such as MS Visual Studio Online or Git) and that could be selectively deployed or rolled back to different instances of a site would be beneficial. This would, for instance, enable the roll-back of just the personalization rules of site during troubleshooting. Or allow for selective deployment of copies of rules among servers, en masse. This feature is less meaningful in the context of SAAS platforms that operate as cloud-based services.

Giving User-focused Features the Love They Deserve

We hope this list of features is helpful for vendors and those who are trying to understand the systems they are using and where they might go.

The absence of these features is evident in all the current platforms (varied, and generally powerful as they are), which is unfortunate because they are the capabilities that would bring use of these platforms more into the mainstream, outside of their obvious niches in conversion optimization and eCommerce. We love most of the advanced features that are the main development focus of platform vendors—everything from advanced profiling of audience segments, hybridization with multivariate testing, automated audience discovery, and purchase lifecycle tracking to integrations with CRM and predictive modeling. But for the sake of those who use the systems (as opposed to those who sell them!), we hope that the more user-focused features get the love they deserve.

We’ve had a flood of features and a bit of a feature war, most of which go far beyond what Web teams can utilize. A lot of skill building should happen, but we’d love to see vendors concentrate more on these usability basics and try to apply the same sort of usability thinking that arrived a few years ago among content management systems, which finally began to reflect on how they were actually used. Today, the personalization vendor market has the opportunity to quickly embrace these sorts of features and usability concerns and transform personalization into a much more indispensable layer in the digital tool kit. Failure to do so will likely lock the systems in their current niche—as optimizers largely used by agencies and specialists—far from the mainstream practices of Web managers, content strategists, and campaign managers. Personalization needs to take its rightful place as a new part of user experience construction.

About the Author

JB-avatar
John Berndt

I'm CEO of TBG and I've been thinking about the Web in creative ways since the year it began.

Leave A Reply

comments powered by Disqus