Our Methodology
Methodology is more than just a diagram and a description. It’s the experience to know when to go “off script” and modify elements as needed for a given project.
Along with our experience in project management and knowledge of the tools we use, our methodology allows us to avoid many of the common pitfalls of model-building or business analytics projects (such as getting lost in the data, building a model that only a few people understand, using a tool that no one will use, solving the wrong business problem, providing an answer that fails to convince).
Some example activities in each phase
Hypothesis
• Conduct facilitated sessions to develop hypothesis
(what is a good hypothesis? what is it used for?)
• Identify the boundaries of the business problem
• Develop related graph(s) and diagrams to describe hypothesis
• Confirm scope of model or analysis
• Develop related graph(s) and diagrams to describe hypothesis
• Confirm scope of model or analysis
Qualitative Model
• Confirm stakeholders and sources of data/information
• Collect any relevant pre-existing documentation
(process maps, transaction records, policy documents, management reports, etc.)
• Conduct interviews and/or facilitated sessions to develop the qualitative model
(interdependencies and relationships across the different elements of the business problem)
• Identify and document relevant business rules
• Identify and discuss scorecard and/or evaluation criteria
• Develop qualitative model document(s)
• Test early concepts of the quantitative model (using mock-ups or previous examples)
• Collect any relevant pre-existing documentation
(process maps, transaction records, policy documents, management reports, etc.)
• Conduct interviews and/or facilitated sessions to develop the qualitative model
(interdependencies and relationships across the different elements of the business problem)
• Identify and document relevant business rules
• Identify and discuss scorecard and/or evaluation criteria
• Develop qualitative model document(s)
• Test early concepts of the quantitative model (using mock-ups or previous examples)
Quantitative Model
• Collect, clean-up and integrate data
• If necessary, build data model or database
• Confirm underlying mathematics or calculation rules
• Identify and confirm tools to utilize (algorithms, UI tools, data and calculation tools, etc.)
• Discuss and identify the scorecard metrics (and how the simulation experiments will be evaluated)
• Build and test the simulation model (algorithms, code, calculations, visualization, UI, etc.)
• If necessary, build data model or database
• Confirm underlying mathematics or calculation rules
• Identify and confirm tools to utilize (algorithms, UI tools, data and calculation tools, etc.)
• Discuss and identify the scorecard metrics (and how the simulation experiments will be evaluated)
• Build and test the simulation model (algorithms, code, calculations, visualization, UI, etc.)
Analysis
• If necessary, develop analysis framework
• Conduct sessions to describe the model or analysis
• Review original hypothesis
• Conduct experiments and collect results (often in a facilitated session)
• Deploy/deliver model
• Conduct sessions to describe the model or analysis
• Review original hypothesis
• Conduct experiments and collect results (often in a facilitated session)
• Deploy/deliver model
Throughout the phases, we use various tools along the way.