In my SAS Press book Business Statistics Made Easy in SAS® I place a strong focus on the skill of extrapolating analytics/statistical outcomes to key business implications (similar techniques can be used to link statistics to other key societal outcomes). Unfortunately, business analytics often stops short of defining the impact in terms that key organizational stakeholders care about.

One key business outcome is Return on Investment (ROI), i.e. dollars of profitability per \$1 of cost. We might use market research to estimate the ROI of an insurance promotional campaign, operational analytics to estimate return on changing a production process, and so on.

The Figure above shows the step-by-step statistics to the ROI process from Chapter 17 of Business Statistics Made Easy in SAS®. Using the marketing example:

1. Statistics gives the level of or change in focal variables. Example: analytics shows change in insurance purchase intention based on promotional campaign (say average improvement of 5%).
2. Estimate revenue/cost of one unit of the focal variable. What is one insurance contract worth? Say customer lifetime value of \$3,162.
3. Combine statistics with per-unit financial values. Generally, multiply the statistical result by the per-unit financial implication of the focal variable(s). Per potential customer, value is .05 x \$3,162 = \$158.10.
4. Include scope. Extend step 3’s result by number of units/people involved and timeframe. Say your promotion reaches 100,000 potential new East Coast customers this month. Then potential revenue from the promotion in this group is .05 x \$3,162 x 100,000 = \$15,810,000. Say you want to do another promotion to a similar Midwest group next month. Potential benefit is .05 x \$3,162 x 100,000 x 2 = \$31,620,000. Note: time extensions can be a lot trickier!
5. Net profitability: Add in other financial information to finish. For ROI we’d need the cost of the campaigns. If cost is \$10 million then ROI = Profit/Cost = (\$31,620,000 - \$10million)/ \$10million = 2.162 = 216%. Financial analysts also compare this to a cost of capital (something to read up on).

Various versions of such techniques can go a long way to making analytics useful to end users.

Share

### About Author

Research Director

Professor Gregory John Lee is currently the Research Director and an Associate Professor in Research Methodology and Decision Sciences at the AMBA-rated Wits Business School, Johannesburg, South Africa. He has authored books on HR Metrics, including his newest title, Business Statistics Made Easy in SAS. Lee focuses on issues in human resource management, notably HR metrics, in which he has established himself as a leading expert, and other areas such as training, employee turnover, and the employee-customer link. He has served in many capacities within the international academic field and has sat on the GMAC Advisory Council, the editorial board of the Journal of Organizational and Occupational Psychology, and he engages in frequent reviewing for many journals. In addition, Lee is a well-known consultant, writer, and speaker in the corporate and practical management arenas, notably in the area of HR metrics but extending to other areas such as human resources strategy and foresight.

### 1 Comment

1. Hi Gregory,
I've always thought about how to translate the benefits in a way stakeholders understand better, and i would love to try your approach for a change.
Thank you for sharing!