How To Measure RPA Success? A Guide to Robotic Process Automation Metrics

One of the coolest things I find about Robotic Process Automation development and implementation is that you can measure virtually everything.

I am not talking about the performance of bots or other highly technical metrics that can be understood only by RPA developers. I have in mind actual business- and money-related metrics that can help you measure your ROI, calculate business gains, and get a clear answer to how much (or: if) RPA automation is profitable for your company.

That’s the true beauty of RPA: being able to generate a data-based report showcasing the exact effects of automation implementation (be that saving $$$, shortening the payment cycle, or decreasing the time needed to perform processes) ready to share with your CEO, board or other stakeholders.

Below you will find the collection of RPA metrics worth considering following from the business project management and technical perspective.

Should you know all of them? Yes. You should get familiar with all the metrics below to get a better understanding of what and how can you measure in regard to RPA projects.

Should you use all of them? No. Over-measurement can be worse than no measurement at all. When you track too many metrics, you lose focus on what is essential and suffer from information overload. You should select a handful of metrics and stick to them over a more extended period of time.

It can also depend on the type of RPA project you are managing, e.g., for some projects, time-related metrics are entirely irrelevant (because the process can be executed overnight or there is no set deadline). For others, timing may be super important (e.g., due to the tight SLAs.)

Value of Time Gains (VTG) tells the difference between the cost of the process carried out by human employees versus the cost of the process carried out by RPA bots. It can be calculated with the following equation:

VTG = (EC – AC)/AC x 100%

Where:

EC = costs of the processes delivered by employees: payrolls (with leaves and holidays), taxes, costs of office and workplace (rent, utilities, etc).

AC = costs of the processes delivered by bots: license fees, cost of development, virtual machine/hosting, and maintenance.

VTG is your base metric for calculating RPA ROI.

Robotic Process Automation Return on Investment is a key business metric that should tell you how much money you saved (in other words, earn) by implementing RPA in your business.

RPA ROI is calculated using the Value of Time Gains (above).

To build a complete picture of RPA ROI, you need to take two factors into account:

Value of process acceleration

The value of process acceleration tells you how much money was gained by implementing the automation.

For example, this can be represented by the effects of shortening the client onboarding cycle: a difference between the number of clients processed before and after the automation multiplied by the average gain on one client.

Value of error reduction:

“Value of error reduction” shows how much money was saved on eliminating mistakes. It should be aligned with what’s essential in your organization.

For example, it can represent the total cost of deals lost due to errors or lengthy processing or the cost of fixing the errors manually (represented by payroll costs, e.g., 30% of 1FTE), the improved percentage of cases under SLA, improved quality of data, or anything necessary for a given process/organization.

After calculating VoPA, VoER, and VTG, we can calculate the RPA ROI as follows:

RPA ROI = [(VTG*AC + Value of process acceleration + Value of error reduction) – AC]/AC x 100%

Want to see an example? Read how our client from the finance sector achieved the 300,000% ROI (yes, you got the zeroes right!) with one RPA implementation.

Expected Business Value is an RPA metric that combines all other RPA business-related KPIs. It represents the sum of all the cost savings (from increased performance, resource utilization, and error reduction) multiplied by the average FTE costs over a given period of time.

“Gained Productivity” is one of the most common metrics companies use to measure RPA gains. It tells you how many FTEs were gained by automating a given process (or multiple processes). For example, if automation has replaced the manual work of 4 people, who would spend 2 hours of their time each day carrying out the process, the gained productivity here would be 1 FTE/month.

Business Value lost in Downtime tells you how downtime hurts your Expected Business Value.

You can calculate the Business Value lost in Downtime by subtracting the quantified downtime from your Expected Business Value over a given period of time.

This metric tells about the average cost of an error in a given automation. It should be the sum of the value lost due to downtime and the cost of working hours needed to fix the error (Break-Fix Person Hours).

It shows the difference between the estimated budget before starting to work on the project and the final budget. In RPA development (as in other types of software development), estimations usually vary from reality due to the unpredicted software issues or changes in the project during the development phase. Everything over 70% accuracy is considered acceptable.

Process velocity tells you the average time needed for a single case/task of the automated process to be executed.

Looking for an example? Here is the story of how we decreased the time needed to execute a user registration process by 30.000%!

These metrics tell you about the average time (expressed in FTEs or the number of working hours of RPA developers) that is needed for the development and implementation of one RPA bot to execute one process.

Automation uptime is the percentage of time your automated processes can be executed. It is the difference between total bot availability and automation downtimes (times when bots were broken or unavailable).

The downtime rate represents the percentage of time your bot could be running a process that is not due to errors or bot maintenance.

This metric shows the % of cases that have been covered by the bot on a given process. The good success metric for RPA is 80% and above.

Of course, building an RPA bot with a 99% accuracy rate is entirely possible. But you need to weigh the longer cost of the development and actual business gains coming from the increased success rate.

This is a very important metric from the business and technical perspectives. The exception rate tells you what is the % of cases that have failed. There are two subcategories of errors here:

Business exceptions, such as missing data, case-specific problems, but also out-of-scope items)

System exceptions (all the things robots should do but couldn’t for any reason, like., timeouts, login errors, changing webpage layout)

For the RPA automation to be considered good, the exception rate shouldn’t go over 20%.

Break-Fix Person Hours metric details how long it takes to fix a given error. It should be expressed in FTEs or in man-hours needed to repair the automation and get it back on track.

Process utilization represents to what extent a given process is utilized in a company’s operations. If a bot in a given process is used 50% of the time, the utilization rate is 50%.

This metric can help you validate which automation is critical to your business and which is playing more of a supportive role, as well as decide which you should scale up and which you can resign from in the future.

It can also be a great indicator of process implementation: consider the situation when 30% of employees are scared of using brand-new automation and are still doing the job manually. This creates a great space to improve bot efficiency.

This metric shows how easy to scale a given process is. It should represent the cost and time needed to add a new machine/duplicate the bot to perform the given task vs. the time and cost needed for recruitment and onboarding the human processor to perform the same tasks + the payroll costs (e.g., 30% FTE).

As a rule of thumb, if the RPA bots are written with all other good practices in mind, they should be bound to scale almost instantly.

The above metrics on measuring RPA success will help you understand your RPA setup better and provide a deeper perspective on what and how you should measure to have complete control over your Robotic Process Automation bot network.

You can quickly implement RPA in your finance, banking, real estate, or any other industry.

They can also come in handy while collaborating with the RPA development agency (such as Flobotics) to better understand what is going on and if your business is actually profiting from RPA implementation.

And if you still feel a bit lost and unsure how to measure RPA success, which metrics should you choose – as I mentioned before, you better avoid measuring EVERYTHING so you don’t get overwhelmed with the data – reach out to us.

Our team, sourced only from the “Big Four” of accounting, will analyze your processes and current RPA setup and recommend the metrics and KPIs to measure best suited for your business goals and needs.