Around 6 weeks ago I wrote the article Logging in to Trip which provoked much negative comment. I have clearly not rushed to reply properly but that’s because I’ve wanted to properly think through the issues.  In short, the reason for asking people to log-in meant we knew more about the users and could therefore improve the service we deliver.  As I see it there are two connected issues of concern:

  • Logging in – it’s a pain for users.
  • Profiling users and altering the results accordingly.

One suggestion (from Paul, see comments in previous post), in relation to the second point, is to focus on our refine by clinical area feature.  This is an already available system which tags documents by clinical area.  So, an article titled ‘Cholesterol and the elderly‘ would be tagged as cardiology and geriatrics.  If a user did a search for, say, cholesterol the above document would be returned in the results, alongside many others.  But if they decided to refine by geriatrics, the above document (alongside others tagged with geriatrics) would be moved to the top of the results.

Below are two images showing how it currently works:

The differences are clear and shows the potential for the system.  For me, it’s about helping users find the answers they need really quickly.  An anaesthetist, interested in ‘awareness’, would find the results of the ‘normal’ Trip disappointing but by selected anaesthesiology all the results are relevant.

As mentioned above this system is available already and can be accessed as shown below:

It’s not particularly developed and we could definitely improve it:

  • Machine learning to improve the document tagging
  • Better user interface so it’s more apparent and more intuitive to use.

Ultimately, it might serve a similar role as profiling users in improving the search results.

Advertisements