Why the complexity of your users should be driving rather than hindering your web personalisation
Web personalisation is no longer a choice. Customers expect a tailored and engaging experience that’s optimised to their needs, and if you aren’t providing that, you’re going to stand out from the crowd – for all the wrong reasons.
But web personalisation is also not new. Almost 20 years ago, web-usability consultant Jakob Nielsen shared his concerns about trying to predict users’ needs when he wrote, “This is difficult to do and even more difficult when you consider that the same person may have different desires at different times.”
Since then, personalisation capabilities have improved dramatically – but the risks remain.
Although retail businesses have led the way in perfecting this personalised experience – Amazon has been making purchase recommendations for decades – other organisations continue to face challenges.
“Organisations such as universities and government agencies often have much more complex audiences and goals,” says Chris Grist, solutions architect at Squiz. “They have a diverse range of customers trying to use their site for a range of purposes.”
Traditionally, organisations approach these customers by segmenting and modelling them as personas. Personas make for great tools to better understand the different aspects of your audience.
As Nielsen pointed out in the ’90s, your users’ desires will constantly shift. Chris believes that modelling your users too simply creates even more complications. “Any one of your users is more nuanced than you can cater to in a single persona,” he explains. “They often fit into multiple personas at one time.”
When these users are visiting sites, such as government or university sites, their goals are much less predictable than they are on retail sites.
So, how do you approach these complex users and ensure that you can provide a personalised experience that caters to them – whatever their goals and desires might be?
Capture the nuances of your users’ behaviour to understand their needs
Digital personalisation aims to create dynamic content that’s relevant to a particular user. But if you’re shoehorning these users into single persona buckets when they fit into two or more of your defined personas, you’re potentially missing out on providing them with the content they need at particular times.
One way to approach this complexity is by using multivariate personas – a concept that considers that a user may concurrently fall under the umbrella of multiple traditional personas. “The idea is similar to the difference between A/B testing and multivariate testing, which you can use when testing design and content changes on your site,” says Chris.
Rather than focusing only on high-level segmentation, you should be capturing all the nuances of user behaviour that you can and targeting personalised content based on the different levels of detail of that user model.
Caption: When you employ multivariate personas, you may find that users meet more than one high-level persona (such as A and B shown here). When you then personalise content for these users based on these multivariate personas, you can deliver not only appropriate A and B content, but also niche content, such as Q and X, that you identify as relevant to this particular persona combination.
“When you design your personalisation based only on those high-level personas, your users receive one version of your content,” Chris says. “By using multivariate personas, you can generate a mixture of content based on any number of personas the user matches, emphasising those for which they have a closer affinity.”
By using a multivariate approach, you’ll be poised to create and deliver niche content that you might otherwise overlook when focusing on top-level personas.
User diversity is shifting how organisations approach personalisation
By tracking user behaviour, a university can identify whether a user is likely to fit into a prospective student persona or a current student persona and optimise the site accordingly.
But sometimes this collected information can contradict itself and break simpler persona-based models. For example, many current students may also be prospective students if they are currently studying but looking to change course or plan for postgraduate study. So, which experience should you optimise for?
With a multivariate approach, you can do both.
There are many top-level audiences for each organisation; in the above university example, for instance, a user could be a current student, a prospective student, a staff member and an alumnus – all at the same time.
“When you’re not using traditional, rigid user personas, you’re not locked into that A or B content model,” Chris explains. “You can be agile and responsive to the user data you’re collecting and provide more personalised and engaging web experiences that actually work.”
Deliver a personalised and engaging experience for all users
It’s rare that we can be certain about segmentation that we infer from user behaviour. “Many of your personas at large, complex organisations such as a university will overlap,” explains Chris. “They could fit any number of roles, and your content should work for all facets of that user independently.”
By targeting and emphasising content on different areas of your site based on what you’ve inferredabout your user – without broad assumptions – you’ll be able to optimise an experience for all your users. And you won’t be excluding content for unique users that you haven’t predicted in your persona modelling.
Take a harder look at your audience. Begin capturing nuanced user behaviours so you’ll be in a position to deliver the personalised, content-rich web experiences that your users expect.