Predicting the Unpredictable with Six Sigma
Housing Association
This housing association a social enterprise, spun out from the local council, and a mix of type and condition of housing stock has been transferred from council ownership, resulting in a £14m operation.
Business Situation:
- Need to budget the spend to refurbish the whole housing stock on a annual rota, covering approximately 800 units
- Historically always overspent, despite trying to use control charts to predict engineering spend
- Used “average unit cost” in the past, although housing units varied widely in size and refurbishment needs
Business Solution:
- Split the type of engineering work done on the housing stock into 3 sections
- Model each section, category a, b, and c, including process capabilities studies
- Build control charts not on means and SD, but on percentiles linked to capability studies, and the appropriate non normal distribution, for each section
- Identify and model properties/units which will liable to have a higher than expected cost, and treat separately
- Bring all costs together for a final control chart and scorecard approach to managing engineering spend
Results:
- Improved budgeting tool, more accurate than historical average approach
- Better understanding of variation within the process, allowing tighter control of contractors
Monthly spend versus budget is closer than ever before
Categories: case studies, consultancy, private sector, reduce costs, six sigma