How To Define and Prioritise Processes for Robotic Process Automation (RPA)?
Process Assessment Matrix for Your RPA Journey
By Arnold Kinzel
Consultant | Advisory
Jan 03, 2020 | Frankfurt, Germany
Before starting one’s RPA journey, one must take a high-level view and classify the processes to be automated before diving deeper into a specific process. This is usually done by using the Process Assessment Matrix shown in Exhibit 1.
With this, companies prioritize their processes for the automation journey in order to realize RPA’s ROI potential earlier and free up resources.
The matrix is traditionally clustered in four categories stemming from the differences in their respective complexities and benefits. Categories include Quick Wins, Must-Do Improvements, Low-Hanging Fruits, and Long-Term Improvements.
The Complexity factors are:
- Number of screens that are involved in the process since they approximate the number of steps.
- Application types such as SAP, Web based applications, MS Office, Java, etc.
- Numbers of Scenarios possible in the process
- Type of inputs whether easily readable and digital or scanned PDF documents, free flow text in Emails (structured, standard, free text, image based)
Statistical validity and re-evaluation
The matrix above serves as a good approximation and helps to imagine what different types of RPA automation processes there are, however, it might create the illusion that organizations have processes that are of equal complexity but one of them bears much higher or lower benefits. This would be too simple.
A much more accurate picture is given in Exhibit 2 where the dots represent RPA automation projects within the matrix including the four classifications seen above. There is strong correlation between complexity and benefits which implies that with great complexity, the benefits become greater. Outliers in the form of Long-Term Improvements or Quick Wins are rather rare, as presented in the red circles below.
Adjusted process assessment matrix
It was established previously that one must not interpret the process assessment matrix too strictly but as an approximation and support in planning. A more realistic diagram can be derived from ‘statistical experience’. The outliers to the right of the trend line are to be prioritized when planning the RPA journey since they will have similar complexities to those on the left but have higher benefits e.g., quick wins. From this, we can derive the red rectangle below where most or all the processes will be found. The processes to the left will be automated the last.
From this analysis, it becomes apparent that there will be rarely processes found outside the above illustrated rectangle. If there were any, those to the right would be prioritized compared to those to the left of the trend line. Thus, the common assessment matrix from Exhibit 1, although a good illustration for basic understanding, needs to be reconsidered and the analysis extended to a more detailed framework for strategic RPA prioritization. The more realistic matrix would look much like in Exhibit 4 replicating our experience and the observation of a complexity-benefit correlation.
Why do all this?
Why can we not simply just take the original matrix and observe the respective processes for each category. An old management wisdom explains that greater time allocated towards early planning will save project costs compared to allocating greater time towards more advanced stages. Thus, a granular approach will enable companies to prioritize in an effective way and realize returns on investments constantly throughout the entire project since quick-wins generate positive ROI early on, while the project progresses to the next stage.
BCH process assessment matrix for RPA
For more effective planning, Business Consulting House has developed an improved and more detailed process assessment matrix with nine categories including action recommendations. The matrix evolved from the above analysis and is filled using an extensive questionnaire about the processes. The processes are evaluated using our own algorithm that runs in the background using a weighted grading system so that each process falls into one of the nine categories shown in Exhibit 5. Each category provides a recommendation to make it more intuitive.
Categories to action according to its automation benefits and complexity:
- Start immediate implementation
- Compare and prioritise
- Invest for medium-run development
- Defer further actions and put last
Processes that fall into the categories “Start immediate implementation” should be automated immediately. In the next stage, management must compare the two “Compare and prioritize” rectangles and assess their individual situation before deciding on processes which are most suitable. In any case, the matrix narrows down the choices to a few processes that presumably have the similar or even the same utility levels for the organization. Therefore, it breaks down the analysis to allow for the most objective evaluation possible. After red processes have been automized, green processes are considered until the project team reaches the white rectangle in the top-left corner. These processes must be deferred until there are no other processes left and are therefore considered long-term improvements.