Building a successful search experience.
What do you do when you can’t find something? We get on our hands and knees, dump out every vessel-like object in sight, search the same empty car door pocket four times. We try to ignore the irrational urge to hit command+F to find that misplaced sandwich we really want to finish or pull out our phones to call our other sock. We blame innocent bystanders for putting it somewhere it doesn’t belong and throw up our hands.
Luckily, when we need to find something in the digital space, we can summon the hero we’ve been missing IRL – we have logic. But we’re still relying on designers and builders to make the search system smart enough to buffer our human forgetfulness or lack of intuition. A search box is a promise. It tells users that the system is capable of making its best guess.
It doesn’t say “I’ll take a quick glance at your query but only if you tell me verbatim what you’re looking for.” From a user experience perspective, having no search is better than having bad search. If the logic isn’t strong or the interface poorly designed, the all-too-familiar chaos of overturned couch cushions ensues in the form of irrelevant results or false empty states. You arrive at an empty results screen suspiciously quickly, stare at the cheeky placeholder icon shrugging his shoulders, and call him a liar. He’s not looking hard enough.
As designers, it’s important to remember that users might not immediately know how to search for what they can’t find. They don’t have an innate schema for how to craft the best query and tick the right facets in order to uncover their perfect search result. This is because WE put the thing users want somewhere in the system, not them. But search is ultimately a DIY experience. We have to anticipate those unknowns and focus on providing the right tools, not just the right end results.
In order to build the quickest path from query to answer, we have to consider the rest of the journey.
Empowering users through supportive search tools doesn’t mean handing them the kitchen sink. If your friend told you she lost her favorite pen in your living room, you wouldn’t say “Cool. Here’s a checklist of possible places you might find it in my living room and also the bathroom, kitchen, bedroom, and garage. Oh, and here are some qualities traditionally used to describe staplers, notepads, and other related office supplies to help jog your memory.” (Unless, of course, you’re a really bad friend). In a search experience, relevancy should always be the priority (and sometimes the only) filtering and sorting logic.
The inclusion of additional configurable options such as scope selection, facets, filters, sorting, and related suggestions should be purposeful, receiving just as much curatorial attention as the results themselves. This search framework should be tailored to the user based on a number of contextual factors including search history, the query, and the search origination location. In order to build the quickest path from query to answer, we have to consider the rest of the journey – where are users coming from before they initiate a search? What else might they be looking for? Where do they want to go next?
Designers need to embrace search as a living, breathing system [...] so that user intent becomes the feedback loop for search optimization.
Finally, notice that search is not just a way to trace lines to matching content. The concept of global search is expanding to fit more aptly into its namesake. Users want to use search in place of the main menu for navigation, or to ask questions expecting a human-like response. Designers need to embrace search as a living, breathing system. A dynamic combo of content tagging and analytics can be used to track user behavior, so that user intent becomes the feedback loop for search optimization. Features such as machine learning, natural language processing, and voice assistance are already driving the next generation of search, and users will soon look for these in the fabric of every search experience.
In summary, here are a few things to keep in mind when designing a search experience:
Let the engine do the work
Trust relevancy, don’t dilute it with too much user control. A good search user experience takes a short, uncluttered path to the best results, without hinting at the heavy lifting it took to get there. In other words, the best systems are those that are invisible to the user. The system should also automatically understand and respond to the user’s journey, including factoring in data such as account status, profile information, search origination location, and recent activity.
Help users help themselves
Offer building blocks for a better query, encourage early scope selection, and provide category navigation tools. When there’s still no luck, make sure empty states provide a pivot, not a dead-end. The system should provide guidance for the way there, and the way back, enabling users to back out of a search experience as confidently as they entered it. Furthermore, don’t ignore the framework – inject relevancy intelligence into the facets and listing styles, not just the results.
Don’t add or hide search features simply because of screen real estate. Design decisions should be based on the user and the content, not screen size. Do consider expected interaction patterns that vary from platform to platform, and look for “enhancement” opportunities that don’t disrupt the larger architecture (for example, thumbnail hover states on desktop).