Lea Patterson

What’s the benefit of defining a model’s scope before beginning a model build?

On Posted on by Lea Patterson to Data-Driven Decision Support

What’s the benefit of defining a model’s scope before beginning a model build?

Defining a model’s scope before beginning to build it may be the difference between developing a useable, defendable and pertinent model that can answer multiple operational and management questions, and that of a model that may struggle to be relevant and robust 12 months down the track.

What does a scoping study involve?

A scoping study would normally involve an on-site visit to meet with model owners, model users, and source data managers to determine:

  • What are the goals of the project?
  • What will the model be used for? Immediately? Short term? Long term?
  • Who will use the model? Will there be specific roles for specific users?
  • What is the structure of the organisation?
  • What source system data is available?
  • Can data be obtained from a central source (BI platform etc) or will it need to be sourced from each individual source system?
  • What data fields can be obtained? What format will each data set be in? What is the quality of the data set?
  • What does the organisation want to cost? Measure? Track?
  • Are there other sources of data outside of source systems such as KPIs or other metrics?
  • Does the organisation want to survey staff, or do they want time to be allocated via a series of business rules?
  • How often will the model need to be updated?

It is also important to obtain sample data sets from each of the major source systems.

What are the benefits of a scoping study?

Prior to the model being built, the soon-to-be model owners and users will have a vision regarding what the model will look like and how it will be used. It is not uncommon to find out that this vision differs remarkably depending on who you talk to.

So the primary purpose of undertaking a scoping study is two-fold: firstly that there is a clear (and single!) vision regarding the outcomes of the project and what the model will be able to do, and secondly, confirmation that the data that is available is able to support the model requirements based on that vision.

Where the data doesn’t support that vision, the scoping study will provide management with possible alternative solutions, or where feasible, a migration path between what is able to be incorporated into the current version of the model and that in the future – this can occur where there are other longer term projects underway, such as the implementation of a new enterprise finance or HR system etc.

In summary, a scoping study will confirm the outputs of the model prior to the start of the project and minimises any future misunderstandings regarding the model’s scope and boundaries. It also provides management with a more granular specification of the structure and capabilities of the model, and the resulting reporting functionality.

Tags