How to get started with a BayMiner QVM project
This summary describes the steps needed to start use the BayMiner QVM method to identify the root causes to punctuality problems in manufacturing of equipment. It suits exceptionally well engineering companies that deliver and install complex equipment in small and mid size volumes and have multiple manufacturing plants. Unlike conventional load levelling methods based on classic statistics and expert team evaluation, BayMiner QVM utilizes the whole organisations’ experience. The knowledge model is easy and fast to update, increasing productivity significantly. This ensures that the root causes of up to 75 % of delayed deliveries can be identified saving much cost with very small investments.
We recommend that you proceed in four stages, two first stages are for a feasibility study, each of which takes 1-3 calendar months and BayesIT’s share of work is 1-3 work days/ phase. The realization of the project takes place during the two later phases. We recommend that in the feasibility study participants from your side are at least:
- The problem owner (the manager responsible for the result of the factory).
- A person that knows the delivery process.
- A person that knows your databases.
About the realization
- BayesIT takes care of the data analysis.
- Typically your work content is some days per phase.
Your tasks are:
- Delivery of data to BayesIT experts and to participate in data selection.
- Testing.
1st phase BayesIT finds out does your company have enough useful data
Contains:
- Review of data collected by the presently operating IT-systems.
- Pre-selection of the relevant variables.
- Creation of the first model and test analysis of it.
- The package includes
- Consulting 1-2 half-days.
- Training basics for 1-2 key persons.
- User licenses for 1-2 test users during the period.
Results of 1st phase:
- A simplified knowledge model about delivery punctuality.
- Basic understanding of knowledge modelling methods for your staff.
- Basic skills to analyse using the BayMiner method.
2nd phase
BayesIT verifies together with you that the data contains enough information
Contains:
- The search for the relevant variables (approx.20) and testing of them continues.
- If necessary, we agree on collecting additional data and the testing of it.
- The package includes
- Consulting 3-4 half-days
- Training for 1-2 key persons.
- User licenses for 1-2 test users during the period.
- Co-operation concerning variable selection and data collection continues until the knowledge model is good enough.
Results of 2nd phase:
- A simplified off-line punctuality model that gives first indications.
- Solutions to fetch and combine data from company databases.
- Determining of the relevant variables.
- Understanding of knowledge modelling by your staff
3rd phase
We verify together that everything works as it should. You prepare for the utilization.
For more information, see BayMiner QVM project description.
Both the work and the invoicing of the fourth phase are subject to the data being good enough. If it does not materialize then the project is cancelled and the final part of the bill is not charged.
