infoneed - Meeting users’ information needs

On his Hitwise blog, Robin Goad notes an increase in searches related to redundancy.

“The top site visited by people searching for the term is a government site for businesses, the Department for Business, Enterprise and Regulatory Reform (BERR), which received 44.2% of traffic; it was followed by My Business (11.1%), a small business advice site. On the other hand, the majority of people searching for the basic term ‘redundancy’ (54.6%) went to the more consumer / public / voter focussed government site, Directgov. Of the people searching for ‘redundancy payments’, 53.4% went to BERR and 20.9% to Directgov.”

He seems to infer that searchers for ‘redundancy’ are employees and searchers for ‘redundancy calculator’ are employers looking to pay people off - presumably on the basis of which site they chose (Directgov for citizens and BERR for businesses). Although he does make the caveat that BERR and My Business are well optimised for ‘redundancy payment’

I attended an interesting presentation on ‘The next step for search’, from a search marketing perspective at internetworld.

Andrew Girdwood and others talked about brand and trade names, social graph driven search and more, but what interested me most was the challenges and opportunities with universal/blended search, where the major search engines are increasingly displaying other content formats - maps, images and especially video within the main search engine results.

So there is a challenge and, of course, an opportunity to get relevant videos etc. displayed in the main SERPs in addition to the regular text link. Perceived wisdom was that its relatively easy to rank well with video for relevant terms and, of course, Google loves YouTube.

On Wednesday, we noticed Google was starting to surface a search within box for some results for navigational searches:

It turns out it’s a wider roll-out by Google as the Official Google blog reports.

“However, one of the trends we noticed while studying teleporting was that there were lots of searchers who would type the name of a specific website as if they wanted to teleport, but would then immediately issue another more a refined search within this site.

Through experimentation, we found that presenting users with a search box as part of the result increases their likelihood of finding the exact page they are looking for. So over the past few days we have been testing, and today we have fully rolled out, a search box that appears within some of the search results themselves. This feature will now occur when we detect a high probability that a user wants more refined search results within a specific site. Like the rest of our snippets, the sites that display the site search box are chosen algorithmically based on metrics that measure how useful the search box is to users.”

The last sentence is interesting. We’ve certainly noticed a wide implementation across UK government websites.

I did a web search for ‘RSS feeds for search results’ and found a Live Search blog entry dating from 1995 that talks about the alpha feature of providing RSS feeds from search results: just add ‘&format=rss’ to the URL of the results and copy the full URL into your RSS Reader.

This still works on the main Live search site and is made easier by just clicking on the orange RSS icon in the URL. Here’s the ‘RSSed’ URL for Flooding information from the UK government: http://search.msn.com/results.aspx?q=flooding+site%3agov.uk&format=rss

Very nifty, but interestingly, Live don’t seem to draw users’ attention to it or even provide information in Help.

Yahoo! offers the similar features to Live, with different syntax, but the RSS icon.

In contrast, Google only offers RSS feeds (RSS and Atom) for its news results but does publicise it clearly. Similarly, the BBC offers RSS of its news and sports results.

Other web search engines such as Ask, Exalead and Clusty do not appear to offer RSS at all.

“Yet government websites rarely fit the way that people actually use the internet. A Google search for ‘UK‘ ‘government’ ‘childhood’ ‘obesity’ and ‘“help’ brings up a site that links to some mildly interesting statistics, but the excellent ‘children and healthy weight’ page on directgov does not come up among the first 100 results.” says Edward Lucas in an Economist special report “Look it up on the web“

I’d comment on two fronts:

• If you search for his keywords on Google (searching the whole web) there are relevant Directgov results at #7 and #11 - although not the article he refers to, they are relevant results. A search on Google restricted to UK gives the same results at #5 and #6. From these pages you can get to the article he refers to.
• He seems, like many commentators, not to understand Directgov’s proposition to write content in the general public’s language. If you change the search to ‘UK government childhood weight’ or ‘children healthy weight’, the “Children and healthy weight” article comes top in whole web and UK. So the article is optimised for general language, rather than more official language.

Nevertheless, it is vital for government to recognise the importance of web search when citizens are looking for information and to start taking visibility in search engines seriously.

John Ferrara writes an excellent piece at Boxes and Arrows on designing for different Search behaviour patterns. He argues that the simple model of search; enter search term/get matching results/read and select ‘best’ result’ overlooks complex human behaviours and differing information needs according to many factors.

He lists six factors, among others, that influence search behavior:

1. Domain expertise
2. Search experience
3. Cognitive style (global thinkers <—-> analytical thinkers)
4. Goal type: navigational searches (to reach a particular location), informational searches (to seek out any documents relating to a topic), Transactional searches (to achieve something online)
5. Mode of searching - browsing and ‘berrypicking’
6. Situational idiosyncracies - search behavior may vary due to mood, workin context, time pressures etc.

Despite the wide variety of variables, Ferrara has observed six behaviour patterns that should be accommodated in design:

1. Alternating between search and browse - have robust cross-linking and hierarchical clues such as crumb trails
2. Minimizing the results set - let users filter search results by categories, include a numeric count of the total number of results and the total number for each category , use “and” as the default operator rather than “or,” so the number of results narrows instead of growing as the user adds more terms
3. Surveying quickly - ensure that result titles are comprehensible at a glance, including application files like PDFs, highlight the terms that match the words originally submitted, let users change the number of results shown per page
4. Making immediate judgments - optimise results for common queries and use best bets
5. Agonizing over the query - provide suggestions as user types, show ‘top’ searches, store user’s previous searches
6. Pogosticking - where user’s bounce around - clear titles and descriptions, show visited links on the results page and consider preview windows of the landing page

This is a really useful reference piece and I’d agree with nearly all of it, except “Highlighting the terms that match the words originally submitted to help people scan the titles and descriptions more easily.”

Shari Thurow has recently blogged on Understanding search usability 1 and 2. I was expecting something on designing search interfaces for usability. Rather, Thorow focuses on the message that optimising the search experience is much more than traditional SEO and should take account of users’ search behaviour.

Thurow argues that search usability is misunderstood by a range of new media disciplines and aims to dispel misconceptions around the use of usability studies, focus groups and web analytics. She argues that key to understanding SEO and web site usability are the human factors. “Why do people do what they do before and after they arrive on your web site? By objectively observing target audience members and carefully analyzing their search behavior, web site owners can improve their web sites.”

Thurow outlines Marcia Bates’ concept of berrypicking - searching is not a linear behavior, rather it comprises a wide variety of behaviors including, but not limited, to: querying, refining, expanding, browsing/surfing, pogo-sticking, foraging, scanning and reading. So users’ berrypicking behaviour needs to be recognised in designing more effective search interfaces.

“The term “search usability” addresses all search behaviors on a single web site, not only querying behavior, and not only browsing behavior. A user-friendly, search-friendly web site accommodates berrypicking behavior and delivers searchers to the information they desire as quickly and easily as possible.”

Search usability needs to incorporate not only the usability of the query interface (search box and results pages etc.) but also ensuring sites are search-friendly (rather than purely search-engine friendly).

“The term “search usability” addresses all search behaviors on a single web site, not only querying behavior, and not only browsing behavior. A user-friendly, search-friendly web site accommodates berrypicking behavior and delivers searchers to the information they desire as quickly and easily as possible.”

So it is important to:

• Identify types of search behaviour
• Understand how these search behaviours are all related
• Design a web site that addresses all or most of these search behaviors, by understanding the searcher’s experience as an objective observer.

Thurow advises remembering three things:

1. You are NOT the user
2. Even if a user fits a profile, persona, or role, a user is not objective or accurate about evaluating his own behavior
3. Users are not always right

So Thurow concludes that “Search usability is a complex subject. There are many types of search behaviors, and plenty of elements on a web page that need to be formatted in such a way to accommodate these behaviors.” I particularly liked her point that “You are here” cues are very important. When users click on a link from web search, they are likely to land in the middle of the site. So good label to help searchers orientate and form a mental model of the site are important in instilling confidence.

Strange to see lots of Japanese language Adsense placements in my Gmail. The content of the email had no obvious Japanese context.

Writing recently at Boxes and Arrows, Stephen Turbek advocates Advancing Advanced Search. I agree with his articulation that “The essential problem of search — too many irrelevant results — has not gone away.” and that most implementations of Advanced Search make little contribution to fixing the problem.

Little contribution because:

• The link to advanced search is usually small and hard to find
• Advanced search pages are usually confusing
• Functionality is often little more than with simple search and it’s hard to revise your query
• You have to make a choice between doing a search and clicking a link to do a search. “In other words, do you want it now, or want to go somewhere else to look?”

After a brief review of alternative approaches - defined parameters, drilling down on tags, and facets, Turbek suggests an alternative approach of progressively disclosing further options.

So after the initial search, a user is presented with a very visible options box above the results. Once the user has invoked this, further options could be built up.

Of course, this is best suited to an enterprise or e-commerce context, where there is likely to be more metadata and users may be more accustomed to filtering their results.

Nevertheless it’s an interesting way of making more advanced search tools available to basic searchers and helps expose relationships the user may not have thought of.

The April 2007 edition of the Journal of Computer-Mediated Communication ran a special theme on the Social, Political, Economic, and Cultural Dimensions of Search Engines.

Despite search engines’ central role in how people access information, little social science study has focused on the non-technical dimensions, the search companies or users’ the practices.

As the search industry has consolidated, decisions made by just a few players can have considerable repercussions for what material is realistically within the reach of users. So it is important to look at what factors determine inclusion and exclusion criteria in search results and how users approach them; so as to gain a better understanding of how users’ access to content is being mediated by a handful of commercial services.The first two articles examine people’s search engine uses at the level of responding to results pages.

• Heuristic and Systematic Use of Search Engines
• In Google We Trust: Users’ Decisions on Rank, Position, and Relevance

The next cluster of articles considers the larger social context of searching.

• Searching for Culture—High and Low
• Learning to Search and Searching to Learn: Income, Education, and Experience Online

The last three articles focus on material covered by search engines, including a look at what decisions influence the content and presentation on these services, a comparison of search engines in different countries, and the possible manipulation of one such tool.

• Is Relevance Relevant? Market, Science, and War: Discourses of Search Engine Quality
• Equal Representation by Search Engines? A Comparison of Websites across Countries and Domains
• Google Bombing from a Time Perspective