Return On Automation (ROA) - A Ratio To Prioritize RPA Projects

Summary

As part of our Robotic Process Automation (RPA) program, one of the most frequent decisions we must make is determining which robots to build. Though all work that can be automated should be automated, companies do not have infinite resources to invest in robots and therefore must choose automation projects that will deliver the greatest value as quickly as possible. RPA programs in the beginning stages must frequently prove that the value derived from automation will at least cover the cost of the investment in automation itself. Return On Automation (ROA) is a simple tool you can use to pick those automation projects that will deliver the most value as soon as possible.

Proving Value From Automation

The benefits from automation are extensive: costs can be reduced, risks can be avoided, quality of decisions can be improved, and higher value-added jobs can be created. However, for RPA programs that are just starting or are thinking about scaling, investments in RPA must be justified by delivering direct financial benefits that will at least offset the costs. If you oversee the RPA program at your company, you have the unenviable task of delivering benefits that must be captured in departments for which you are not responsible. Capturing benefits from RPA is extremely difficult: you can build the robots, but others must monetize those benefits by reducing spending. I run an automation program where we build robots that save people time, but I must rely on other leaders to translate those time savings into dollars. This article is not focused on the process for capturing benefits from automation but rather to help you increase the chances that the benefits from your automation projects will be easily capturable. The currency of an RPA program is hours, specifically hours automated per year. The more hours that can be automated through RPA, the better the chances are that those hours saved will translate into direct cost savings. However, if you chase automating high volumes of hours by itself, you might find yourself automating thousands of hours without saving a single dollar.

Return On Automation (ROA)

When we began our automation journey, we spent a lot of time identifying automation use cases that would automate the most hours. If an automation idea was less than one thousand hours, it was considered inadequate. I recall discussing an automation idea that would save 750 hours per year with a leader who expressed her disappointment that the idea just wasn’t “big enough.” However, looking at hours by itself ignores speed to value. Speed to value is the measure of time required to start experiencing benefits from an investment. Have a bot idea that will save 2,000 hours? Tell me how long it will take to put the bot in service and that answer is speed to value. You should consider hours automated and speed to value together to rank automation projects. To prove my point, I offer the following example. Let’s say we have two bot ideas: one bot (“Bot A”) will save 2,000 hours per year and the other bot (“Bot B”) will save 700 hours per year. If I gave you that information alone and I asked you which bot to pick, you might be tempted to choose the bot that will save 2,000 hours. However, what if the effort to build each bot is different:

We can conclude that developing Bot B will take less time than building Bot A. Now let’s say we ask one developer “John” and another developer “Mary” to spend their time building bots that resemble Bot A and Bot B (i.e. with the same hours automated and hours to develop). If John and Mary spent three years focused on building bots like Bot A and Bot B, here are how many bots they would create and the number of hours they would automate:

As you can see, if we focus on building bots like Bot B, over time we will automate 50% more hours than if we were to have focused on building bots like Bot A. If we were to pick a bot based on the hours automated alone, we may not automate the maximum number of hours over a time period. Hopefully, you can see why speed to value is so important to consider when choosing automation projects. Similarly, just as considering the number of hours automated alone should be avoided, considering speed alone must be avoided as well. For example, if the hours automated by Bot B were only 400 instead of 700, the total hours automated over three years would be 12,000 hours. Taken together, however, we can quickly calculate which bot project will yield the most hours automated over time.

Return On Automation (ROA) is the ratio between Hours Automated and Hours To Develop. For a given robot project, ROA can be calculated as follows:

ROA (for a single bot project) = Hours Automated By The Bot / Hours Required to Develop the Bot

Let’s calculate ROA for our two bot examples:

If you are managing a portfolio of automation project ideas, choosing the bots with the highest ROA will always yield the most hours automated over time. At our company, we choose a small portfolio of bot projects every quarter and use ROA, in part, to pick the portfolio of automation projects that will yield the most value over time. We calculate ROA for each automation project, sort the list, and use the resulting ranking to help pick the best projects.

Conclusion

Saving time is only one of the many benefits we receive from Robotic Process Automation. In fact, over time, the greatest benefit of RPA will be freeing up human ingenuity, allowing people to innovate and solve problems. In the early stages of your RPA program you will likely be asked to prove the value of automation by monetizing the hours saved. To maximize the chances of achieving your automation goals, do not consider hours alone when picking your bot projects. Instead, use Return On Automation (ROA) to maximize the benefits of your RPA program.

--

Andy Quick leads the Robotic Process Automation (RPA) program for finance and human resources at a Fortune 500 energy company. Andy is a certified Automation Anywhere RPA trainer and an adjunct instructor at Tulane University where he teaches a class in robotic process automation. He is also the winner of the inaugural Automation Anywhere Bot Games competition in 2018.