Well it’s a subject that wouldn’t go down well with any other group of folks but, here, I think it should be well received. University of Exeter (represented by Beth Reichwald and Sandy Day) presenting with their BI project and subsequent data quality initiative.
HESA and HEFCE are both drivers but problems in data quality exposed by BI. Phew – thank goodness our BI isn’t good enough to show us the cracks back at home 
At least, as far as I know, we don’t have students who are studying with us in 9999 like Exeter. I guess that’s planning ahead for you though.
Exeter seem to have grasped data quality at a more fundamental level than most of us, though, talking about appropriateness of coding structures etc as well as just whether data is ‘right’. I think this is one for us all to take on board. I’m pretty sure that every institution has some degree of craziness around their coding structures. Usually it’s down to “it made a lot more sense at the time…er.. when my predecessor made the decision”. Maybe we need to revisit data decisions as well as data itself on a more regular basis?
A very good point is made that data quality matters more now than ever – with ever decreasing barriers between customers/prospective customers and your data. Data now flows from source to web to external person (prospective customer, press, funding council etc) with ever increasing speed and if the data is bad then the institution looks bad.
It really looks like Exeter took a robust approach to identifying data issues, giving context and prioritising corrections and getting things in order. A well-published set of core data seems to have been critical to their success. Ongoing prevention of problems have been focused on staff development – educating staff regarding data quality. This is really critical and I definitely recognise a lack of data training (rather than systems training) and the issues that arise as a result. It’s a tricky nettle to grasp, though, when everyone has to learn how to use the systems as well. Worth it in the long term though…
And, if you need to be convinced about how much damage bad data can do – here’s the light-hearted video used in the presentation: http://bit.ly/QZeZj