Maximizing Business Process Automation ROI
The point of business process automation is to save money, or improve consistency by having a machine perform a process instead of a person. The idea is straightforward, but it can be hard to figure out which business processes to automate and how to prioritize them, especially when developers are so expensive. It is easy to fall into the trap of automating the business processes that stakeholders identify as causing the most pain. While this may endear you to your stakeholders, which has its benefits, you will likely not be delivering the most value.
When choosing business processes to automate, keep in mind what computers and people are good at. Computers are fantastic at rapidly and repeatedly performing the same exact action. People are fantastic at performing context reliant, complex, variable actions. The best business processes to automate are simple, mature, time consuming processes where data integrity is important.
Have you ever tried to write out the complete logic of a process that requires little thought for you? It always shocks me how quickly the if/else statements stack up. People are great at context dependent decision making. Computers require exact instructions at every decision point to do the same.
The more complex a process is, the less frequently its code is run. If there is no variability within a process, then all of the code to automate the process is run all of the time. Every decision point within a process introduces code that will sometimes go unused, decreasing the return on investement.
Complex processes also take more time to build and test and can lead to higher maintenance costs. Stakeholders are more likely to discover edge cases that were missed in the original request and to reach out with questions about the automation. Since they are not performing the processes anymore, they forget the nuances of it. Even if the automation is working correctly, stakeholders will reach out because they have forgotten the details, requiring developers to look into non issues.
The code in simple processes is run more frequently and requires less developer time to create and maintain.
New, or changing processes are top of mind for stakeholders. They are unfamiliar and create additional work for the team, making them likely to be identified as a high priority to automate. These are, however, the worst processes to automate.
New processes are not well defined and change frequently. A large amount of discovery will be required before automation can begin. If development resources have been dedicated to the project, they will either be idle, or involved in discovering the process instead of automating it.
It will also be impossible to know the value automation. The process may take a long time to perform manually now, due to its novelty. As it becomes familiar it may end up being a quick process to perform. Similarly, it may look like a simple process now because it has not been fully built out, but as discovery progresses it could become complex.
The maintenance costs of changing processes are higher. Instead of being able to complete the project and move on, developers will be tied to it making updates.
Another consequence is that the owners of the process become less familiar with it than the automators. Ultimately the automators are forced to own the business process, increasing developer costs.
Mature processes make it possible to assess the benefits automation and reduce the development and maintenance costs.
Time Consuming Processes
The longer a process takes to run manually, the more human hours you save by automating it, so the return on investment of automating it is high. There are a two different factors that make up the total process time: how often the process is run and how long it takes to run. The most time consuming processes are ones that run all the time and take a long time to run.
This is the most obvious component, but important to mention because it is one of the most beneficiel components of automation.
High Accuracy Processes
Computers are more consistent than people, so it can be beneficial to automate processes that require high data integrity. For high accuracy processes to be beneficial to automate, the automation has to perform them more accurately than a human.
It takes a lot of work to make sure a complex process is automated correctly. If the appropriate resources are not allocated to the development, automation can end up producing lower data integrity than the manual process. In the worst case you can end up investing more in the development of the automation than the data integrity is worth, or scrapping the automation and reverting to the manual process.
Automation can improve the accuracy of processes, but only if enough resources are dedicated to the project.
When choosing business processes to automate, it is easy to understand the benefits – saving human time and improving accuracy – but often hard to understand the costs. Looking at the simplicity and maturity of processes can help identify the costs and lead to better returns on investment.