Can a site design itself?
With this post, I’d like to touch upon two separate trends for the future: personalisation and artificially intelligent websites.
Personalisation has been around for what feels like an age, from Google first tailoring search results over a decade ago in 2005 (I’m really wishing I hadn’t looked that up!) via Netflix and Spotify making suggestions based on previous choices, to today when an individual site visit can be tailored based on how a user has interacted with it so far. But these are just the beginning and it feels like there’s a long way to go on the web generally. The vast majority of sites are still designed in a generic manner with the same content for everyone, forcing the user to sift for what’s relevant to them rather than presenting tailored information based on things like what they’ve already seen, their location and other relevant information.
I think all organisations understand the importance of better personalising web experiences for the user but actually implementing and managing it often raises difficulties for them. For example, the extra work often involved in producing tailored content and managing the necessary rules for everything to work smoothly.
But over time, and with systems like Sitecore, we’ll get to the point where the majority of sites are adapting based on all the relevant data that’s available rather than just one or two pieces of information.
Understandably, this usually raises concerns regarding privacy and permission, like the oft-cited example of Target, whose use of purchase history meant they were aware a teen girl was pregnant before her father was and sent her a baby tailored mailer in the post that he was not best pleased about.
As companies start to end up with a Single Customer View (SCV), it allows them to infer a lot but also cross boundaries. Generally speaking, this is most commonly discussed from an organisation’s point of view but what if we pass control of everything to the user?
This essentially comes down to implicit and explicit personalisation. Implicit personalisation takes a user’s past behaviour and characteristics to provide content a webmaster thinks will be relevant to them whereas explicit personalisation puts that choice in the user’s hands. There are pros and cons to each approach but generally speaking, it makes sense to let the user control this as much as they choose to.
By doing that, we are helping users to understand how and why they are seeing certain things and also allowing them to correct what we think we know of their characteristics and preferences, especially when people are using shared devices.
Google’s MyActivity is now part of the account settings, allowing users to review everything that Google has tracked from search, YouTube, Chrome, Android and the rest then edit or delete it should they want to.
This is an example of a potential move towards giving users further control but what if we make sure users can have complete control?
For example, I have all my main information held in once place, control my own ID and my browser has access to it. This could be a universal profile including: my name, my date of birth, my interests, the industry I work in, what types of things I buy and on and on. Whatever I’m happy to share.
I’m then in a position to grant access to certain brands/sites and they can tailor what they show me based on this information or I can chose to withhold it from them.
When a site knows who’s visiting it and what their preferences and interests are, then it can show me only the most relevant content, thereby improving my experience. Since I have control over sharing that information, there’s no longer a need for the developers to try and walk the tightrope between personal versus intrusive, which is different for every user anyway.
Anyway, what about that second area I said I was going to touch upon - artificially intelligent sites?
Well, with AI and machine-learning becoming more common in other areas, what does a website with the capability to imitate intelligent human behaviour look like?
Let’s be clear, some people refer to AI sites but they mean things like thegrid.io, which is pretty nifty but is essentially just automating how some design decisions are made (if I stop and listen carefully, I think I can hear a designer crying somewhere in the distance at this very concept).
So, let’s take it a stage further. What happens if a site could change itself?
Imagine a site that’s doing multivariate testing every single time someone visits it. That tailors itself based on how it’s being used and adjusts content, navigation, layout, format etc. to better achieve defined organisational goals.
In this case, there’s no right or wrong “solution” that everyone sees. Instead, we’d all see something different based on the information that we provide: what we originally searched/asked for, what our universal profile says about us and how similar people to us used the site.
If 1000 people are visiting the site simultaneously then they could all be viewing something entirely different and personal to them.
In human interactions, we are often asked exactly the same question (e.g. “how are you?”) by our spouse, boss or old school friend and we give an entirely different response based on our knowledge of the asker and what information will be of interest. Why should a website be any different?
Obviously, like all AI, we can expect some unintended consequences. Or as one of my colleagues put it, the joke probably goes a little something like this:
“Before we created this new site, we sold plumbing equipment but now we build bespoke business intelligence systems as that’s what our website decided to start selling to users…”
In conclusion, users could have more control of their data on a browser level and also be given more direct influence over what data they share, thereby getting a better level of personalisation.
Secondly, once AI begins to take hold, we can probably expect sites that look completely different for everyone who visits them, constantly test themselves to improve their efficiency and of course, the usual unpredictable consequences.
As always, it would be great to know your thoughts on this so feel free to share, comment below or ping me a message on Twitter (@ianlcrawford).
By Ian Crawford, Digital Strategist