I “attended” the webinar on Findability: Designing the Search Experience. It was given by Tyler Tate, who jointly runs a company which produces a toolkit for Solr / Lucene search engine development.
The talk essentially expanded on an earlier article by Tyler (“The Scent of Search” – see recommended reading list at the end of this section), and I took a bunch of notes, written up here.
To view the on-demand webinar (including audio), you need to register your details.
You can also download Tyler’s slides.
As you might guess, Tyler discussed ideas for designing for findability. This is something I’ve mentioned before, and I often like using Peter Morville‘s “UX honeycomb” to illustrate the facets of UX design. We often talk about usability, but this is only one facet that we should consider; another important one is how findable your content / data is.
How do people search?
We should consider both search expertise (how many people you boolean operators in their searches?) and domain expertise (e.g. search engines for domain-specific information sources).
Designers need to find the right balance of interface complexity when they consider these factors. Commonly, search interfaces are designed for the lowest common denominator (think of Bing and Google) because very few assumptions can be made about the users’ expertise.
Specialist search facilities are often more complex, taking into account the domain expertise of the users.
[FR] – Sometimes though, we have to remember that there is a range of domain expertise, and although a user may be have a lot of domain knowledge, they might still be a relatively inexpert search interface user.
Specific vs open-ended tasks
The examples used here were Kayak vs Yahoo. Kayak has a quite a complex interface, but this is appropriate for the specific task of finding and booking travel details. Yahoo is suitably simple because the tasks are quite open-ended, and the results page is not the end-point for the user.
Graphing the relationships between expertise and specificity:
[FR] – I think that in many cases, the “scientific research” point might belong in other places… the user might have a highly specific task (I tend to think about “goals”); the user might also have low search expertise, as mentioned above. That complicates things a bit!
Recall vs Exploration
Trying to recall facts, as opposed to using search as a way to explore an information space. The example used was Channel 4′s Mapumental.
Immediate vs. Iterative
An example of an immediate search plus goal achievement might be asking the Wolfram Alpha search engine what is the population of Brazil.
Iterative searches are where the user drills down to what they want, using hints from the search facility itself. The example Tyler used was looking for a guitar to buy on eBay. You start with “guitar”, but then you can filter results by various categories (acoustic / electric; brand; etc).
[FR] – I guess this also touches on “pearl growing” in search. A later point also covers this…
10 tips for Designing for Findability
- make the search box obvious
- provide as-you-type suggestions (e.g. auto-completion) [FR] – top tip!
- indicate the number of results
- use descriptive titles, highlight search terms, differentiate visited links
- use “top hits” to show more info on key results
- never let “zero” results happen
- use breadcrumbs [FR] – again, very helpful for pearl growing
- make metadata clickable (i.e. apply metadata to search query)
- indicate the number of results for each filter (facet)
- only use meaningful / useful facet visualizations [FR] – I’ve suggested result bar charts for the upcoming Enzyme Portal results pages… could work really well
Download the slides
Download the PDF of Tyler’s webinar slides (12MB). Note that there are a couple of errors in the links to recommended reading. This list should be:
- Read Peter Morville’s book “Search Patterns” – http://searchpatterns.org
- Read Tyler’s article “The Scent of Search” – http://johnnyholland.org/2010/07/05/the-scent-of-search/
- Read Marti Hearst’s book “Search User Interfaces” – http://searchuserinterfaces.com/book/
- Follow TwigKit’s blog, “The Dao of Search” – http://blog.twigkit.com