Intelligent Automation: Do you use an "automation index" to measure and report value?

Measuring and reporting the success of an intelligent automation program is often done in a simple way. If you have an RPA-based program, then you might be doing something along the following lines:

1. Estimate the number of hours that you have freed up from each task you automate, add them up
2. Multiply by some blended cost rate to get a dollar value of the time freed up
3. Maybe you have additional measures around, for example, reductions in time to execute a process, increases in customer satisfaction, increased accuracy for a task.

How satisfying (to business people) are these measures? Did your cost-saving translate into a P&L impact that you can draw a direct line to? Is that increase in customer satisfaction directly attributable to your automation initiative? (and not other reasons?) How real is that increase in task accuracy you reported? Are you able to use these measures to justify scaling your automation program?

On one hand, reporting back "hours saved" is a single measure that can be used to track across automation initiatives. On the other hand, if it does not show up in the financials - with a direct P&L impact - then the measure itself can lose its value, or worse be ignored.

One way to track the progress of an automation initiative is through the creation of an "automation index". 

 A standard use case for RPA-type programs is the automation of processing invoices. You could report back on hours saved, but perhaps reporting an automation index based on the human capital leverage - invoices per person processed - may be a better measure:

A team of 10 people processing 500 invoices in a month gives you an index of 50. After automation, a team of 2 people processing 1,000 invoices per month increases that index to 500. This demonstrates a 10-fold increase in performance and in a single number encapsulates time savings, scalability, and an impressive improvement in performance. 

Alternatively, the index could be based on sales orders processed per person month, customers managed per account manager, or issues resolved per person. Another example of an index, this time from the insurance industry, comes from the lemonade.com blog found here. 

One of the great advantages of using an index like this is that it can capture value from scalability in a process, that simply recording hours saved cannot track. It is also simple to understand, especially if your index is always tracking "up". An index can also be used to compare yourself to competitors in the same industry - the Lemonade.com blog article linked above also illustrates this.

While a single index to cover every single automation initiative might be the most desirable, this may not always be practical, and a proliferation of different indices is also probably not ideal. The "sweet spot" may be to find a limited number of performance indices that are clear, simple and can be traced to a financial impact. Often the index will reflect "work performed per human" at a team, department or functional level.  

You will probably need to (and should) continue to track hours saved or whatever other metrics you have been tracking, but an automation index can become a great way to communicate the success of what you are doing and can still be quantified in financial terms when required. Most importantly to a practitioner -you might be able to use it to help scale your own automation program even further!