Modelling to Measure Return on Investment

Be it a staff training program or a new computer system, the return on any investment in your business is very difficult to assess. So many other factors affect business performance that it is not possible to monitor the benefits without building a model which incorporates other major factors which drive business profitability.

This study concerns a retail chain who were installing a point-of-sale monitoring system in all their 2000 outlets. In theory, the system offered a higher standard of service to the customer and a more efficient transaction process for the retailer. When a small proportion of point-of-sale systems had been installed, the retailer wanted to review the process to make sure the investment was paying off.

The boost to gross profits due to the new point-of-sale system would be swamped by a variety of factors affecting profitability such as grade of manager, regional growth, and recent investment in the outlet. The only way to isolate the contribution of the point-of-sale system to profit growth was to build a single model of profit growth based on all these different factors. Only the neural network modelling approach, with it’s ability to see patterns in poor quality data, can reliably achieve this.

Data was collected on all the outlets: region, grade of manager, outlet type, etc., including the two factors of greatest interest, growth in profits and number of months a point-of-sale system had been installed.

All the different factors were used to model profit growth:

The above graph obtained from the neural net model shows that having the point-of-sale system installed for the full 12 months (over which profit growth was measured) boosts average profit growth by nearly ten percentage points—in this case, clear evidence that the $50 million investment will pay for itself.

The neural network took many factors into account when building the model, and it is possible to find out more about what drives profit growth in the outlets. For example, the benefit of manager training can be pinpointed with a similar profit growth analysis graph.

Thus, it is now possible to measure the benefits of "soft" investments such as staff training.

Additional Cases

• Case 1: Modelling for demand forecasting

• Case 2: Modelling to improve sales force efficiency

• Case 3: Modelling advertising and promotion effectiveness

• Case 4: Modelling for direct mail targeting and customer profiling 

Back to 4Thought overview