When we use Google to explore the web, we're happy to accept the fact that we're venturing through a wild, open and sprawling information space. There’s always the possibility there’s a haystack out there we haven’t discovered yet, and in it, there might be some great needles that we want to find out about. Sometimes that's part of the fun.
Our expectations are very different when we search an individual website. For example, when we carry out a search on a classifieds or e-commerce site, we expect it to be able to show us every single item it has that matches our criteria. To use an analogy, imagine if you went into camera shop on the high street and asked to see all of their digital SLRs. Imagine if the sales assistant responded by showing you a few cameras, but then admitted that there might be more items in the back room, possibly better deals, and that you’d need to try re-articulating your requirements in various different ways to find out. A crazy idea. You'd think they were incompetent and walk out.
A shop assistant should know what stock they have, and they should be able to match it against your requirements. The same goes for websites.
Keyword search has been made famous by Google, but there are other ways to elicit users' requirements. Faceted navigation (aka "faceted search" or "guided navigation") is one of them. It is, in essence, a very simple concept. Instead of giving users text fields for them to fill in, you give them a list of links to click on. These links consist of lists of attributes, separated by type, each with a numerical count adjacent, indicating the number of results that will be displayed if selected. This may sound quite unremarkable, but it completely transforms the act of searching an unfamiliar database from an intimidating procedure into an easy task. The user just reads the options and clicks those that are most relevant to their needs. After two or three clicks, they have narrowed the results down to a manageable list of relevant items.
Although faceted navigation is fashionable right now, it’s not just a fad. Its true value is grounded in human psychology - let’s compare the thought process a user has to go through when carrying out a keyword search against the thought process involved in faceted browsing. We’re using job advertising in this example.
<tr > Keyword Search Faceted Browse
|The user looks at the blank search box and wonders "What should I type?"||The user looks at the sector list. He is reassured to see there are 971 Healthcare jobs available.|
|Since he is looking for a nursing role, he types "Licenced Practical Nurse".||The user looks at the sector list. He is reassured to see there are 971 Healthcare jobs available.|
|The results page shows zero results. He wonders if this is because the site doesn't have any matching jobs, or whether he's done something wrong.||He clicks “Healthcare” and is immediately taken to a search results page with 971 jobs.|
|He then tries the keyword "nurse" which returns a lot of results, but some are quite irrelevant.||He sees that for his desired salary range ($40-49k) there are 18 jobs - a manageable number of results, so he clicks the link and proceeds to the next page.|
|So, he tries "LPN" (an industry acronym) and gets the kind of results he is seeking.||Two clicks later, he finds himself looking at 18 relevant jobs.|
|He adds a filter to only view jobs in his desired salary range (£40-59k).|
|After a fair amount of thinking, typing, clicking, and re-trying, he ends up looking at a page of relevant results.|
So, to summarise, keyword search can sometimes be quite "hit or miss" and involve a fair amount of thought from the end-user, whereas faceted navigation helps guide users to their destination. To break the benefits down into themes, faceted navigation helps by:
- Deconstructing search into a series of easy 'bite sized' subtasks: rather than declaring a complex set of criteria 'up front', the user can declare one criterion at a time, gradually drilling in.
- Appealing to the psychology of recognition rather than recall: it's much easier for humans to recognise items displayed in a menu ("Which of these options should I click?") than recalling them without any cues ("What should I type into this empty box?").
- Setting expectations accurately: adjacent to each facet label, a number is displayed, showing how many results will be displayed if the item is selected. This means users are never taken by surprise, and it enables them to make informed decisions.
- Orientation and framing: When arriving on the site, the user can see at a glance the distribution of items within the various facets, which orients them as to what is on offer. If the site uses appropriate labelling, this can be very reassuring for the user as they will recognise the site 'speaks their language' and has the kind of content they are looking for. (This is known as "information scent" and "information foraging").
- Preventing users from reaching dead ends: links that have zero results associated with them are never shown. This means users cannot accidentally reach blank results pages - which we all know from personal experience is extremely frustrating.
By this point, you might be thinking we're advocating faceted navigation as a cure for all ills. It isn't - it has a mix of strengths and weaknesses, and this mix is quite different from keyword search. In fact, since each approach has strengths in different areas, it makes sense to provide both types of functionality and allow the user to decide which is best suited to their needs.
One of the problems with faceted navigation is that it can be implemented badly, making it hard to use. If you are considering developing your own faceted navigation system, here are some pitfalls you may want to avoid:
- Unfamiliar terminology: the language you use inside your organisation may differ considerably from your users. Be careful to validate your chosen terminology with real users, to ensure you do not alienate them.
- Overly long attribute lists: lists in user interfaces are known to work well at around the 7 item mark, but become taxing for users when they become very long. If you absolutely have to use a very long list, you should try to prioritize the most important items to the top, and add a 'show / hide' control which keeps the bulk of the list hidden away unless needed.
- Overlapping attributes and blurred boundaries: if users perceive three attributes of a facet to be almost the same, they are presented with a conundrum - which one do they click? To be on the safe side, they have to view each one in turn. (This is known as 'pogo sticking' and can be extremely frustrating). Some systems avoid pogo sticking by allowing users to select multiple attributes from each facet - however, this can create usability problems.
- Poor editorial control: if your content creators assign attributes incorrectly, this will make the entire system ineffective. For example, on job boards, some recruiters mistakenly believe that by assigning their job ad to all the categories they can, their job ad will get more valuable exposure (in fact it's a very poor tactic and reflects badly on the recruiter). The best way for a website owner to avoid this kind of behaviour is to provide a content creation interface that gives contextual tips and prevents users from making such errors where possible.
To sum up, faceted navigation isn't a magic bullet, but it can complement keyword search extremely well. Whichever method you choose for making your content findable, make sure it's easy to use. After all if your users can't find it, it may as well not exist.