Why your intranet looks so dull…

Check out this interesting article around the value of branding intranets; Why Your Intranet Needs Its Own Personality.The article points out that the weak governance most organisations have around their intranets makes developing and applying branding difficult. I found it interesting that the author describes a familiar three way tension for branding:

1. Users don’t really care about corporate branding
2. A lot of managers don't want to follow corporate branding, what they actually want is the ability develop sub-branding for thier division or business unit intranet sites
3. Typically it is ‘only a few lonely soles at the top of food chain’ who seem to be adamant on strong corporate branding being applied internally

Sound familiar? The article goes onto describe various reasons why branding the intranet is of value. For me the key point was keep a strong focus on creating value for staff who use the intranet.
Why can’t the intranet be funky?
How many times have you looked at your corporate intranet and despaired that it is another fine example in all that is dull and lazy in web design? Surely for many large ‘steady as she goes' organisations the internal web space is place where a few more risks can be taken? So why can’t the intranet be funky and engaging or at least well designed and easy on the eye?
Try shifting the focus to users and at all cost avoid branding becoming an expression of the power of the designated cooperate owner. How many times have you heard ‘we need to enforce the branding guidelines developed by marketing/internal communication/corporate affairs’…

Tag clouds, life events and showing relationships between data…

Last week I had a colleague phone me and ask me what I know about tag clouds. He’s had an idea that this might be a good technology for highlighting relationships between services offered by the organisation. A sort of Amazon model of people who brought X may also be interested in Y & Z. This arose from a more general discussion within the org about using a “life events” taxonomy to classify and then combine online services from several agencies.

It strikes me that what all of this is really about is being able to infer and illuminate thematic relationships between discrete pieces of information. In my view this is one of the greatest opportunities afford by the web and more specifically by semantic technologies. So for this post I’m going to try answer the initial question and cover some options and ideas for ‘semantic lite’ web apps.

Starting with Tag clouds…
There are lot of different approaches to generating tag clouds. However for the purpose of this discussion the most interesting is:

1. Classifying or ‘tagging’ content using some form of taxonomy and adding these to the metadata, typically using subject or functional keywords


2. Then displaying these tags in a visual way on the page, with each word being a link to keyword search for each term.

The ABC uses this approach for their hot tags function. For a good overview of tag clouds see Tag Clouds Gallery: Examples And Good Practices.

And now onto inferencing...
Consider the following simple example of an inference:

• If “A is related to B” and “B is related to C”
• Then it is possible to infer “A and C are related in someway”

So how would this work in multi agency Life events approach? If someone is registering the birth of a new born child in Agency A. And this service is part of a group of the ‘becoming a parent’ life event, then this person may eligible for other services in this life event group. Now consider the following list of actual birth related life events from various government service providers:

• Birth (http://www.healthinsite.gov.au/)
• Having a child (Service Tasmania)
• Having a Baby (Connecting NSW)
• Becoming a Parent (Citizen Ireland)
• Starting a family (State of Jersey)

So if these were mapped as meaning the same thing, i.e.

• Birth = Having a child = Having a Baby = Becoming a Parent = Starting a family

Then it would be possible to infer that someone eligible for a service related to ‘having a child’ from Agency A could be eligible for services classified as ‘starting a family’ in Agency B.

Combining this visualisation…
So taking the tagging and the inferencing as foundation and adding some visualisation technology I think it would be possible to create some very useful ‘visual maps’ of services.I’ve done a quick trawl and found some interesting sites to give an idea of how powerful this approach can be.
Live Plasma is really cool as it re-use data from Amazon API and adds a great interface to help users to see relationships between music genres, artist etc. So it is simply representing exiting publicly available data in a visual way!

And have a look at this online thesaurus . This tool searches a number of online thesauri. Look out for the results from the Visual Thesaurus.

A few more sites if you’re interested in visualisation
http://www.visualcomplexity.com/
http://www.orgnet.com/twitter.html
http://www.ivy.fr/revealicious/demo/spacenav.html

How much should an organisation invest improving the intranet search?

Last week I began grappling with the following question:

“How much should the organisation invest improving the intranet search and will this be recouped in timed saved finding information on the internal web channel?”

There is general agreement that poor search cost organisations money in lost time. Quantifying this is tends to be based around simplistic extrapolation. For example: X Staff spending Y seconds each on searching for information equals ten zillion dollars in lost productivity. Therefore we should invest money in a new search application. Sort of fails to discuss the complexity of finding information or recognise that there are plenty of possible approaches in addition to just focusing on improving the search application.

As I ponder this within the context of where I work some thoughts have begun to emerge including:

• Quantifying the performance of an internal search engine is difficult and will probably be based around fuzzy qualities measures


• Without these it could be difficult to establish a benchmark in which to measure the effectiveness of various enhancements or optimisations


• Building a search based around lots of high quality metadata is resource intensive. It could end up costing the organisation more then it returns in productivity improvements.


• Is it more cost effective to rely on the built in power of a good search application and put the resources into improving content, fixing the IA, archiving out of date pages etc?

I’ve found a couple of interesting articles and I'm sure I’ll be posting more on this topic.

• Metadata Schmetadata, Relevance and Reality


• How to Measure Findability in Enterprise Search Solutions