The complex task of designing search for data heavy products
Google makes designing search seem like the easiest thing in the world.
Most of us turn to it multiple times a day, relying on it to correctly interpret our queries and point us in the right direction. It’s pretty much the simplest task we perform on the web. If Google can make finding just what we’re looking for feel so effortless, then anything less from the digital products we use, will naturally seem frustratingly inferior.
Unfortunately, many product teams run up against major headaches when designing their platform’s search experience.
This is especially true of the data heavy, complex products we typically partner with, where ease of surfacing relevant content is key.
We’ll drive into the intricacies in this article, uncovering what makes designing search one of the toughest but most vital UX challenges for products of this sort.
“Search is a defining element of the user experience… Unfortunately, it’s also the source of endless frustration. Search is the worst usability problem on the Web.”
-Peter Moville and Jeffrey Callendar “Search Patterns: Design for Discovery”
#1 Different users search differently
With any data heavy tool, it’s really common to have several very distinct user personas. We might divide them into ‘casual users’, who only really scratch the surface of its capabilities, and ‘power users’ who dig much deeper.
They represent two very different use cases for how deep people want to go into search, and how complex their queries are likely to be.
The casual user is a run of the mill ‘googler.’ They’re often fairly infrequent visitors who aren’t likely to dig deep, and tend to use simple queries to view a broader range of data at a higher level.
A power user is likely to be confident with Boolean operators and multiple parameters to build much more targeted datasets. They’ll want to get down to the granular level on a more regular basis, answering highly specific questions efficiently.
There’s a natural tension in interface design between emphasising simplicity for the infrequent, straightforward searches carried out by the basic user and making immediately clear the powerful tools available to the advanced user. Add too much complicated ‘clutter’ for the basic user and they’ll be overwhelmed by choices they don’t need. At the same time, make the advanced user jump through too many hoops to access their higher level functionality and they’ll end up frustrated.
Getting the balance right is no mean feat!
#2 Designing results takes a lot of thought
Another potential minefield for product teams working on data heavy platforms is designing how the content generated should be organised.
There are plenty of decisions to be made in order to create something that’s maximally engaging and digestible. A results page that isn’t scannable is going to switch all but the most devoted user right off. That’s because humans are notoriously bad at taking in anything but a modest quantity of data at the best of times. Psychologist George Miller famously discovered the limits of information processing in 1965, finding that the average number of items we can hold in our short-term memory is “7, plus or minus 2.” Presented with too high a number of items all at once, we begin to lose accuracy and ultimately our interest wanes entirely.
A good UX designer applies ‘Miller’s law’ across loads of elements of their work. We’d certainly recommend keeping it in mind when designing forms, for example.
Bringing it into play in search results is likely to involve ‘chunking’ and/or ‘blending.’ (incidentally, also key for a perfect soup 🥣 – yum!)
Blending
Blending results, that is to say mixing different visual elements in the display design, is something Google is unsurprisingly very good at. Think of all the different items displayed on a Google results page (aka a ‘SERP’) these days.
There’s the results themselves, a knowledge graph, a featured snippet, an answer box, images, shopping results and more, all sitting.
Depending on the type of query and the data Google finds, different elements appear in the blend in order to ultimately give us relevant information in a much more manageable form.
Chunking
Chunking results follows a not-dissimilar principle of grouping results into – you guessed it – chunks of meaning to make information easier to scan and understand.
A designer is likely to use Gestalt grouping principles (such as decisions around proximity, use of spacing, borders and similarity), assisted by colour and iconography to visually represent this.
It’s all about making it easier to see at a glance the results that belong together. Showing an article is different to an event, for example, or that a result come from an external source rather than being generated by the platform itself.
Heads up – we say ‘assisted by’ because just showing a state or difference with colour or iconography wouldn’t be accessible. Any modern design team should be designing for everyone by default.
#3 Fitting the wider digital landscape
Regular readers of our articles (or just fans of good UX practice in general) will know that user research is a non-negotiable part of any design project.
Walking in your users’ shoes and gaining a solid understanding of how they interact with the digital world and the other platforms they encounter is always a smart idea.
Humans are – as we keep saying – creatures of habit by default, and meeting user expectation by delivering an experience that feels familiar is an important part of designing search.
When it comes to data heavy platforms, it’s worth bearing in mind that the ‘competitor’ that users are highly familiar with searching probably isn’t a shiny SaaS platform.
It’s just as likely to be Excel, or a database with massive technical capabilities but a UI that was designed by engineers in the early 2000s.
Obviously we aren’t saying you should design your product’s search capability to look and feel like an old school Windows product (please don’t), but that there could be certain comfortable searching habits that have become ingrained in users from their experience with such applications. They may be surprisingly loath to give them up.
It’s all about balance.
Good search UX is often a tightrope walk between familiar functionality – even though the platforms used were clunky and on the ugly side – and enough newness to spark delight. A rigorous and well-executed research phase will ensure you don’t lose your tread as you navigate along this tightrope.
In conclusion…
Like almost every UX decision, the search design choices made by product teams hinge on understanding what works for their own real users.
We can say with some certainty though that all users of data heavy products need a search experience that is intuitive no matter how deep they intend to dig.
From the casual browser to the multiple-times-daily superfan, the platform they rely on must surface relevant content and present it in a digestible, scannable manner that fits comfortably within their wider digital landscape.
Need help designing an effective search? Get in touch and we’d love to help.
