The analysis phase is purely mathematical. It is during this phase that we set up a set of tools to identify the real cause of the problem.
For this, the mathematical tools at our disposal are numerous. It is a question of choosing them wisely according to the problem encountered, the possible measures, the means, the context … these tools, go simple tools of the problem solving: diagram of correlation, diagram Ishikawa… to the more complex tools below.
Do you know what a statistician says with feet in the embers and head in a bucket of cold water ?
” On average I am well “
1. Descriptive statistics
The methods of descriptive statistics allow studies to be carried out on the basis of exhaustive data, ie concerning all individuals in the study population1. Simpler than inferential statistics, we find notions of Mean, standard deviation or Variance.
2. Inference statistics
Les inferential statistics refer to the procedures by which we generalize the information of a sample to the population from which it was drawn. It is in this type of statistics that we find tools such as :
3. Validate the cause
Set up the case verification plan by testing. The important thing is to be able to reproduce the problem wisely.
Good questions to ask at the end of this phase
- Do we have the root cause of the problem ?
- Is it root causes or only symptoms ?
- Do we know how to reproduce the defect ?
- Are we in tune with the expectations of the project ?
1 – E. Bressoud, J. C. Kahane (2010) – Statistique descriptive
J.N. Gillot (2012) – La gestion des processus métier
C. Frechet (2011) – Mettre en œuvre le Six Sigma
W. Fox (1999) – Statistiques sociales
N. Volck (2011) – Déployer et exploiter Lean Six Sigma
R. Basu (2011) – Fit Six Sigma : A Lean approach to building sustainable quality beyond Six Sigma
T. Burton, J. L. Sams (2005) – Six Sigma for small and midsized organizations
M. Pillet (2013) – Six Sigma : comment l’appliquer