Thursday, March 25, 2010

Intranet Search is Overrated

You should not rely on your intranet's search capabilities to ensure your users find the tools, content and collaboration that they need. Many intranets appear to cede the battle of providing proper and useful organisation and navigation under the premise that their search capability can quickly find the content their users can or should find value in. I believe that intranet search has value but is inherently limited, both functionally and technologically. Search can be improved, but is only one of many solutions for connecting your users with content, tools and collaboration opportunities that form the basis of intranet business value.

From what I have seen, intranet search is overrated.

I had the idea to write about this topic from both reading some of the discussions on the Linked In group Intranet Professionals and my recent experience at a large organisation. Intranet search can be an excellent tool for connecting users with the content they want to access, relatively quickly and without reliance on knowledge of the intranet structure and organisation. I do not want to devalue that capability, but rather point out the important limitations that we should be aware of. Some of those limitations are functional, and intrinsic to search. Others are technological, meaning that the right search engine may alleviate the limitation, but I do not think the solution is widely available.

Search is a poor tool for discovery. This is a functional limitation. Users may not know that specific content exists or could be useful, and search can not reveal this to them. The concept of enabling discovery is key to the success of many public social web applications, and for good reasons: As good as a search such as Google or Bing becomes, search does not surface interesting content before you know that you need it. Social networking applications such as Facebook and Twitter, or social bookmarking applications such as Digg and Delicious are valuable because expose information we would not have guessed that we cared about. In the intranet environment discovery it is not hard to find the value of discovery. It may mean that a user is made aware of a new procedure, or of an existing initiative they could add value to, or of a document summarizing highly relevant and costly won experience.

Search can also founder on providing context and authority for the items returned as results.

The context of a result in this instance refers to its place within the website hierarchy that can expose important contextual information. For example, imagine a set of guidelines on the use of social networking websites by employees. That set of guidelines may be interpreted significantly differently if the user understands the placement of the document within the marketing department's section of the intranet, and not as a part of the more general policy documents applicable to all users. Another example could be an exciting project proposal which, unbeknown to the user, was actually written years past and was since moved to the  'shelved projects' area. A good intranet search facility can highlight the place of a document within the intranet within the search results, but I would be willing to bet that everyday comprehension of this context by users is lost.

Authority is a related concept to context. Users commonly need to judge the authority with which a piece of content is written. Is the content old? Was it written by your manager, as opposed to by the over-zealous contractor? In other words, especially as the intranet become more encompassing of all business content, the user needs to know whether the search result is current and should acted on, or whether it should be treated with skepticism. Again, a good search engine can expose the information and signals around a piece of content that influence the user's judgement of validity. However, user ability to interpret or even notice this information may not exist.

The above are functional issues with search, but there are a number of practical issues as well. Users can have difficulty formulating search queries that will surface the content they need. Business terminology can vary wildly between different departments, let alone organisations. Thanks to the preeminence of search on the wider Internet, users are increasingly skilled at formulating a search terms. However, the intranet is not the boundless container of endless forms of content that the intranet is. Failing to achieve results in an intranet search may reflect either a poor search or that the content that does not exist. How can the user know, without performing more than one search?

The more insidious issue with intranet search is one of overconfidence. Reliance on search can encourage both intranet maintainers and users to ignore or neglect both traditional intranet navigation and organisation and newer social ways to aggregate content for discovery and reuse. This is unfortunate, since these are the very tools that can compliment search and provide for the limitations and deficiencies I mention above.

With this in mind, it is my opinion intranet search is usually limited and that you certainly should not rely on search to meet the end user's needs. Thinking about how users can take part in Discovery and understanding how content context and authority can be communicated should be a priority. Search can be improved, but equally important, other forms of content navigation should be reviewed and buttressed.

All intranet search is not created equally. There was a time when full-text search was a major feature, and now it is simply the starting point. Does your search provide any context or authority for the results? Can the user understand how old the document is, who wrote it, who maintains it or how valid the content is? Does the search support word stemming, synonyms or the domain language of your business? Can it be tuned, or even better, learn from user behaviour? Does it leverage intranet structure and document links (ala Page Rank)? Unfortunately the average intranet search facility is a black box that is difficult to understand and harder to customize, especially for the maintainers themselves.

Intranet Search is a useful tool, and you should maintain and enhance the functionality to reduce the limitations I have described. But very often the potential for a particular search package is tapped early, and replacement  too costly. Sometimes investing in a dedicated search appliance can be advantageous to allow a decoupled upgrade path. But regardless of how you can improve search, search is fundamentally only one way of connecting users to valuable content, tools and collaboration.

Reviewing and improving the intranet features that complement search, such as intranet organisation, navigation and customization are essential for avoiding an over-reliance on search. A well designed intranet will compete strongly with search in allowing fast access to frequently required content. Where users can navigate with speed and efficiency in an intranet to the content they need the limitations of search do not apply.

Navigating the site hierarchy communicates context and authority excellently. Any content hosted from the Sales area of the intranet can be assumed to apply specifically to the Sales department. Any content specifically linked from the the Sales homepage is more likely to be authoritative on a subject, at least for the Sales department. Each page, which may be navigated to organically or via a search result, should also give a clear indication about the context that accompanies that page or piece of content.

There are many opportunities to enable content discovery through traditional navigation. Content discovery can  based on the area of the intranet, the user, the activities of other users and other signals. Home pages can be customized with frequently accessed content or provide a staging point for accessing the content, tools and collaboration around a specific community or project.

Where intranet organisation, navigation and customization are working well, search should only see use intermittently for occasionally accessed content. This should be a user preference. If it is not, I would take that as a sign the intranet organisation, navigation and customization can still be improved. You can also perform a simple test. If you disable the search facility, is it still possible to find all of the information required for a specific user?

What do you think? Do we rely too much on our intranet search. Is over-reliance on intranet search, like the user bookmarks I discussed in my previous post, a symptom of poor intranet design?


  1. A visit to the intranet is simply a journey. Starting with the user and ending with the content or task they desired. How they take that trip is up to them. Navigating through menus, searching, using A-Z index pages or sitemaps or even from a bookmark that they saved previously. Using any one of these methods of “transport” to complete their trip does not reflect on the quality of the intranet. It's a simple user preference. Different users get to pages in different ways.

    What would reflect poorly, in my opinion, is when the user has to use several different methods of transport to get to their destination. Rather like taking the tube home only to find that it terminates halfway, then having to ask someone directions to the best bus stop and then taking the bus.

    Not all intranets are full of user-generated content (yet!) So the notion of surfacing content that we wouldn't otherwise see is not the same as when we are on the web. Some intranets have a limited amount of content. Users need to get their job done, hopefully using the intranet as an aid, but few have the luxury of surfing the intranet, hopping from blog to blog to news article. Fewer have the desire to go looking. When staff come to the intranet they've got a job to do, a task to complete, information or data to find. They do it and get on with their job. The average staff visit to our intranet takes under 2 minutes. Time is money. And using a bookmark to get to a page instead of wading through navigation or search results is sometimes just a reflection of time and efficiency, a reduction in the number of clicks. But again, a personal choice.

    Intranet managers should concentrate on users and be aware that users will have their own preference for finding content. I don't look on search as a cop out. Yes, users will resort to search if other preferences fail. But 30% of visits to our intranet content are initiated from the search box. These users actually prefer to search (like they do with Google and incidentally we use Google as our intranet search engine) instead of click menus. Just because they searched doesn't mean the navigation structure is wrong.

    I don't think we should rely on any one method of transport, be it search engine, navigation or bookmarks, but strive to offer success to our users through all them.

  2. Hi Luke. Thanks for your comments. When a user knows exactly what they want - and that it exists - search can be an efficient way to find content, tools and collaboration opportunities.

    It is too simple to say that users searching for content implies poor navigation structures, and this was not my point. In my opinion, an over-reliance on search, either by the users or more significantly by the intranet managers, can limit the potential of the intranet to deliver business value, for the reasons I discussed.

    Thanks again for commenting!