My first RNIB project, and one I was dumped straight into on my first day, was an intranet re-launch.

The old intranet was dated and creaking. It was powered by Stellent which was soon to be unsupported. Applications  produced by different teams were styled as a completely different site (with some lovely palette choices). The main navigation had reflected the old org structure and had been removed and not replaced when it became out of date(!)

SharePoint had already been chosen as the content management systems and we had to overlay overlay an accessible interface on top of SharePoint rather than using the out of the box interface. The big challenge for the project (and where most of the effort went) was around delivering an accessible front and back-end.

The scope for the project included lots of nice ideas and even functionality that I would have considered standard for most intranet projects. But as the scale of the accessibility work became clear we rapidly descoped.

We ruled out taxonomy-based tagging, faceted search, blogging,and wiki functionality. We focused on getting the absolute basics right, on things that would deliver clear business benefits. Even Best Bets got descoped, a decision I made but did so rather sadly.

One of my main tasks was to define the SharePoint content types and the metadata fields. I’ve discussed reasons to define a SharePoint content type elsewhere on this site. For the reason stated there we ended up with only one. We weren’t likely to do much clever with the content so we tried to keep the content creation overheads low. The spreadsheet below shows content types that were considered and abandonned (or struck-through).

The search was widely derided but the main problem was it was running unforgiving AND queries so if a single word in your query was ‘wrong’ then you got no results. ‘Sick form’ returned no results because you needed to search for ‘absence form’. Staff had learned to simply search for one word you were confident of (like form in this example) and then scroll through the results. These problems are a bit of a gift to IAs as they are so frustrating for users and so easy to fix!

A big improvement was that we introduced primary navigation. The navigation was topic based rather than organisational. The organisation had to go through that standard IA journey of understanding that org structures aren’t the primary way that even staff think about finding (most) information or completing tasks. Generally staff only understand the organisation structure in their immediate area and often they don’t even understand that, particularly if your organisation is prone to re-organisation.

We introduced a chunk of new content around IT topics like ‘How to change your password, login to the network from home etc…’ Simple stuff but just plain useful. Again could have been done in the old system, although wouldn’t have been as visible.

Discussion forums  were entirely new and so the business benefits were a bit of a leap of faith. They’ve been highly successful for the right topics, particularly those around the strategy and campaigning. They work well as an indicator of what is occupying the minds of staff.

We also used them for getting feedback on the intranet. I asked for feedback around search in particular, which was a useful way of getting some context to the search logs. Surfacing the latest discussions on the homepage helped drive traffic and also helped keep the homepage interesting.  Forums is a new concept for the business but there’s been no bad behaviour, everyone adjusted fine.

We also introduced collaboration spaces (or TeamSites in SP-speak). These are private areas with webpages, document libraries and discussion forums. They are useful for our distributed teams, with home-workers and with problematic access to existing shared drives.

Just don’t get me started on puns in intranet names.