The future of search
When we talk of search, the first things that usually spring to mind are powerhouses like Google or Bing, churning away billions of lines of code and giving you a set of results that you whittle through until you find what you were looking for — not too far removed from searching your home for something.
You know what you want is there, you’re just not sure where it is.
Now I’m not saying the modern-day search engine is bad; it has come on leaps and bounds from the days of Archie. But it is rather antiquated and needs an upgrade.
When the World Wide Web was launched on August 6th 1991, the personas of those people who were trying to access data were used to that style of input/output method.
Fast forward 26 years, and we still search pretty much in the same way, albeit now instead of entering ‘white paint’ users are more inclined to use a phrase like ‘I need white paint tomorrow’.
If we bring this search model across into a website, things become even more of a muddle. Yes, there are some great sites out there that have great search experiences, but they’re not engaging and there is little to no emotional connection. But emotion sells.
With the advent of technology we now have the opportunity to implement engaging interactive searches that actually work.
Let’s look at combining two relatively new technologies.
First, let’s go back to the drawing board using a metaphor I have overused in client meetings for far too long: “If you walked into your local store and couldn’t find what you were looking for, would you walk out?” That’s the same for websites, except in a store, there is usually someone to ask and get directions to what you want.
When looking at providing that same level of experience that you get in a store, there was one obvious answer we came up with. A chat function. But human-to-human interaction would be too expensive and time-consuming! So instead, we looked at introducing ‘bots’ into the fold.
The challenge with bots is that they’re built for structured and repetitive jobs; even though they can be fun to interact with they’re not the most intelligent. The input of information is still requiring you to define search parameters, in much the same way as you were doing already, for very little improvement on the results being returned.
So the next step was to look at how we could turn a mundane task into something that was engaging to use.
We turned our eyes to machine learning (or machine not-learning, but more on that another time) and examined how we could couple the two together. The technology is very much in its infancy, but it’s gaining headway and the possibilities are exciting.
The problem with bots, even with a machine learning upgrade, is that they’re still pretty stupid. Having used Shopify in the past — they have a bot called ‘kit’. It’s pretty basic, and in the majority of uses it would have been quicker for me to have just logged into Facebook to place an advert.
What machine learning could provide is a higher level of engagement, it could detect the dialect in which I converse and mimic that based on pre-defined rules. It could remember my purchase history, my sizes, my colour preferences my understandings.
The more I ‘teach’ it by coming back again and again, the more it understands me and tailors my online experience accordingly.
Websites will be completely customised experiences that don’t contain the traditional ugly mega menu, making you drill through tens of refine options. It will be a conversation “I need some jeans for a party next week” “ I want to go on holiday somewhere in Europe next march, I only want to spend £800 for the week all in”. I’ll let your imagination fill in the replies.
To you and me we know it’s not technically learning per se, but to the end user they’re having a fully customised experience that they can engage with and create emotional attachment to.