This article was initially published in December 2018 and has been updated in September 2020.
Clients usually look for a straightforward, ready-made solution to their questions about measuring the ROI of their RPA deployment. The truth is, no two implementations are the same. Some companies start by automating a few processes which we call “low hanging fruit”, and some automate an entire program.
The benefits of implementing RPA become clear sooner in the first case, but most companies also realise that they will inevitably need to scale at enterprise level. UiPath’s Chief Strategy Officer Vargha Moayed recommends building a business case for your RPA program over a 2–3 year period to properly identify costs and benefits. This is something your RPA partner can help you with.
To accurately determine costs and benefits, you will need to assess both operational and business insights. Operational insights will offer information about the execution of your RPA implementation — for example, the number of hours a robot worked the previous month. Business insights refer to how the RPA implementation has impacted your company (business) — for example, the number of invoices that have been processed the previous month.
Keeping this in mind, let’s take a closer look at how you can measure the ROI that your robotic process automation deployment produces.
Read on to see our new infographic that summarises the 5 most efficient ways to measure the ROI of your robotic process automation deployment — placed at the end of the article.
— — — — — — — — — — — — — — — — — — — — — — — — — — — —
The changes brought about by robotic process automation generated reactions and fears similar in some ways to the Industrial Revolution. It’s natural for people to be be afraid of emergent entities that seem ready to accomplish the same objectives that they do. This kind of thinking gives voice to the basic human need of feeling needed.
Robotic process automation deployment activates this biologically rooted manner of thinking. However, things might be a little less dramatic than that when it comes to the alleged human — robot competition on the labour market.
A recent Accenture research report, whose title we deem relevant — “Intelligent Technology Meets Human Ingenuity to Create the Future Workforce”, emphasizes that not only the revenues of companies that ressort to automation are expected to rise by 32% by 2022, but, more importantly, employment will also increase by 9%.
Apparently, facts corroborate arguments that, although natural, the “robot phobia” is nothing but a myth. Not only will software robots not steal our jobs, but RPA deployment will lead to more available jobs.
Worries of job elimination are thus not justified, so let’s look on the bright side of RPA: the rise in return on investment (ROI) that it brings.
The ROI of robotic process automation deployment
What underlies those higher revenues expected from RPA? The example we’ll use as a road opener connects with the job anxiety that we touched upon at the beginning. PwC mentions a reduction in workforce costs of up to $2 trillion as a result of automating almost half the work activities in companies across the globe.
The right choice of processes to be automated is a precondition of obtaining a fast and significant increase in ROI. To begin with, you should first consider highly repetitive, tedious processes, which, as a consequence, are error prone and frequently call for re-work.
The short story reads “the simpler the process, the higher the savings, the quicker the ROI”. But of course, there are many more aspects that are worth considering when picking the most suitable processes for automation; you can read more on the subject here.
Suppose you’ve asked the right questions and, thereby, you started RPA implementation after having optimized your wisely chosen processes. So perhaps you now ponder what robotic process automation deployment will bring you in the long run.
Your customers might value your higher security risk-mitigation capabilities and entrust you their confidential data, as well as the benefits of extended uninterrupted service and much less “errors” in the system. Read, competitive advantage, read, faster and higher revenue.
Scaling is the safe way to go after implementation, because the financial gains will be most important when software robots are used extensively in a company. If you now consider other additional savings in areas such as workforce (e.g., reduced holiday payments, training, and team-building sessions), or economizing on time by not needing to double-check for and correct human errors, it shouldn’t be difficult to visualize its effect on the rising shape of the ROI curve. (These considerations are relevant for (1) in the list below; they form a solid basis for ulterior ROI calculations.)
Let’s see how you can operationalise the theoretical part and assess the concrete operation cost reductions. That is to ask, how can you actually calculate the ROI of RPA deployment in a comprehensive manner, beyond the financial impact? What kind of metrics allow for insightful evaluations?
1. Be clear about expected benefits
For efficient and accurate ROI measurement, this is a precondition. Be specific so that know exactly what you are aiming for. By being realistic and keeping your feet on the ground, you actually maximize the possibility of reaching your goals.
Pay special attention to the qualitative benefits (e.g., data and process quality — see (3), compliance, employees’ job satisfaction) that may affect ROI. These are by definition more difficult to evaluate quantitatively, so you might be tempted to dismiss them. Don’t.
A comprehensive ROI estimation, including qualitative as well as financial benefits, is more likely to provide an accurate overall evaluation of robotic process automation deployment in your company.
2. Estimate process speed
This metric is particularly suitable for back-office processes. The idea is to compare the time it takes to run a complete cycle (from input to output) before and after the use of software robots, expecting that RPA increases process velocity.
Robots do not need brakes or time-outs. In fact, you may expect processes to complete somewhere between 20 and 110% faster when software robots are involved. Service level agreements, for instance, provide a consistent example of increased processing speed because of RPA deployment.
When doing the calculations, you should take into account the effect of varying parameters, e.g., the monthly volume of transactions. In order to stay safe from potential artifacts, you should average over such fluctuations.
3. Evaluate output quality
This measurement is also the result of subtracting a quantitative representation of process accuracy when performed solely by human employees, from the same results when the process is run by robots. You expect a positive difference, the higher the better.
The main reason is that bots are not only error-proof, but also the noise that might result in deviations from the original design is absent. Therefore expectations of improved accuracy are legitimate, and 100% process accuracy is no longer wishful thinking. Some concrete ways to measure output quality is by looking at the costs imposed by output dissatisfaction, or by the need to rework.
4. Estimate compliance improvement
As we previously argued, robotic process automation is a trustworthy ally in managing the paperwork nightmare of compliance. By including automatic compliance check ups at critical points in the workflow, you can ensure that regulations are followed and compliance criteria are met.
Software robots can handle the basics of regulatory reports production as well. Built-in compliance measures can also take effect on the previous two metrics, by improving both process speed and process accuracy.
