Skip to main content

Nottingham Trent University Case Study

Business Intelligence Project

Efficiency Exchange article interviewing James Lacey, the Director of Finance at Nottingham Trent University

An interview the Data Capability project undertook with James Lacey and his team has been transcribed in summary here:

What were the most important ‘relationships’ in the project (between business units) and how were they forged/maintained?

The original team was built ‘organically’ based on the desired outcomes from the three projects. Subject matter experts were brought in as and when needed. The core project team remained constant throughout the initiative.

The project was managed using Prince2 and was specifically business led and technology enabled.

How were the changes/goals communicated?

The top down nature of the project created a lot of senior management visibility. Providing the data early – even when the quality was known to be suspect – did much to demonstrate transparency and and end state where ‘the single version of the truth’ would be created from this dataset.

By creating a central shared service with report writing capability this became the ‘go-to’ place for data. This reduced the need/effectiveness of cottage industries within individual departments and faculties. Also bringing a wider audience to understand the importance of data driven outputs (such as the HESA return) showed the value of setting quality metrics for data, managing it with good practice and having a single ‘data mart’ for the outputs.

What training, if any, was required for the use of the new capabilities?

‘A mixed economy’. Specific training for the report writers, super-users in the business units and faculties. Some resistance to change although as the institution transitions to business as usual this has become less of an issue.

Interestingly within data operations the skills have moved away from hard-core data analysts to something more business focused. Very much like the change from a systems analyst to a business analyst. There is no less ‘work’ for these individuals but the outputs are now very much aligned to wider business objectives, not difficult ETL or SQL scripts looking for problems.

Within the ‘data users’ community it was a case of using data to be more effective in the role rather than having to ‘wrangle’ data of dubious quality to perform the role. Now the trusted data is mostly ‘live’ much of the confusion around versions and ownership have gone away.

If the project were to be done again what would be done differently?

Potentially more consultative. Although difficult to know at the time how hard to push against ‘how hard to ask’. Without the project having a very clearly defined outcome and the most senior sponsor, this would likely have been more of a problem.

What lessons were learned/what advice would you give to someone starting this project?

  1. Don’t be afraid to publish data early. Creates transparency and early results. It doesn’t matter if it’s not perfect, it gets people talking about the right things – i.e. increasing the quality of a single data source not all of them.
  2. Be single minded. Accept it is a difficult road and there will be ups and downs. Expect criticism and setbacks. Consider a Project Board that is very inclusive of the more difficult stakeholders.
  3. Hold your nerve. Be robust and consistent. Keep an eye on the end state but be consultative and open to new ideas.
  4. Dedicate 100% resource to the project. Back fill those roles. Otherwise there will be conflicts over priority between the project and the day job. Individuals within the project team must be allocated full time and their costs need to be factored into the original project/change budgets.
  5. Create an end state (see 3) – specifically so there can be a common agreement about what is being achieved. This needs to be a moving target because as business changes so will the end state, but it needs to be rich enough in narrative and objectives to obtain approval from different stakeholders.
  6. Find a good partner. You can’t do all this yourself. Many of the ‘big’ players in the data space (e.g. IBM) do no consider organisations such as universities as large customers, and subsequently do not provide sufficient support. Smaller, local re-sellers with similar HE experience are often a better choice.
  7. Don’t get hung up with frameworks. They can help, but they shouldn’t drive or constrain the project. Focus on best practice for the organisation, not on a generic set of activities.