Three fundamental factors should be considered when developing a LOE versus ROI plot, which is used to make decisions regarding when to address specific data science questions.

1) Focus on the actionability and value of the questions results.

The following factors inform which data science question(s) best serve them:

• Who will use the results?
• What level of influence do they have?
• How can they influence the business?

2) Gauge the quality of the data and its alignment to the question.

• How well will the available data support the development of a model that will produce meaningful results?
• Is the data sufficiently clean and organized to support model development?

3) Assess the statistical feasibility of developing a model that addresses the question.

• Consider the reliability of potential solutions. In other words, how likely is it that the results will be reliable enough to drive business decisions.
• Is the available data, when applied to potential statistical approaches, enough to address the desired business question?

Finally, outline how the questions fit into the broader data science landscape.

Often data science questions build off each other to provide insight into additional business questions.

• How do the model results of a given question support the development of a more complex or difficult question?
• What other data science models would make the solution to a question more valuable before processing?

Want to develop a LOE versus ROI plot for your organization?

• Schedule a call with us now.
• Download our free predictive guide.