Aligning Your Automation Plans with Clear Objectives
So, you’ve decided to automate some or all your business processes. We will walk you through how to measure whether you’re automating the right process or process step and meeting your objectives. This article uses some common terminology and acronyms related to automation and process analysis. We find that those terms aren’t always used consistently, so we’ve provided a set of definitions as a reference for the enterprise automation space here.
Know Your Business Objectives
As with any product management and development project, the entire project team and supporting executives must align on what the business objectives are. (Need help with that? See this post) Objectives typically relate to reducing costs or increasing revenue in some way, but how you get to that will differ from project to project. Each objective must have a measurable key performance indicator (KPI) that you will use to gauge the success of your automation project. Example objectives with KPIs may include:
- Reduce average time to resolve customer support inquiries from 28 hours to 8 hours.
- Achieve 100% accuracy when data moves from your customer database to downstream systems, such as order fulfillment and external carriers.
- Reduce manual interventions by bank employees for new consumer bank account opening from 40% to 20%.
Pick the Right KPIs
Let your objectives guide you to the right KPIs. Below are some metrics that you may want to track as part of your automation project.
Turnaround Time (TAT) is a measure of the time it takes to complete a process end-to-end correctly. The higher your TAT, the longer it takes to complete the process. Good automation should decrease the TAT. You might just measure parts of the process in time, even just one step of it.
Costs are a measure of staff costs to perform a process and technology costs related to automating a process. Staff cost might correlate to TAT or opportunity costs of staff doing other tasks. Technology costs could be associated with maintaining or implementing new technologies.
Error Rates can tell you how error-prone your process is. This is especially important when considering processes that require humans to swivel-seat or do manual data entry. Ideally, your error rate should decrease with automation.
Deviation Rate measures how often the user must stop and deviate from a particular workflow. Deviations are usually confined to edge cases, but if you have a high deviation rate, something is wrong. Either you need to re-engineer the process, develop a new process, or implement something for edger cases.
Net Promotor Score (NPS) measures how satisfied your internal business users or external customers are with your service.
Bounce Rates allow you to measure how often a user abandons a process. High bounce rates indicate that something is amiss in your process, and your users don’t feel it’s worth their time to try to complete it.
Know Your Budget and Timeline
Anything is possible with unlimited money, resources, and time! Unfortunately, we rarely get to work in such an environment. An executive may not be interested in an automation project that fulfills a time objective if the cost dwarfs the benefits of a speedier process. Because automation tools can range from simple robotic process automation (RPA) scripts to advanced artificial intelligence (AI) and machine learning, you must keep your solution aligned with the project and timeline. (By the way, you can read more about those terms here). Taking the time to document your requirements and evaluate possible solutions before selecting one ensures that the project will cost-effectively meet your organization’s needs. This analysis may even show that it is cheaper to leave manual processes in place rather than automate them.
Know the Industry Standard for your Process
Most likely, the processes you have in place are similar to others in your organization or those at other organizations within your industry. If industry-standard metrics exist, they are a source to use for setting performance improvement goals. Either way, set attainable goals for improvement.
Measure Your Current Process
Understanding objectives and KPIs is essential for any project and should be set based on current performance and realistic targets for the processes being improved. The following suggestions are adapted from Software Requirements, 3rd Edition (Microsoft Press, 2013) by Karl Wiegers for Process Impact and Joy Beatty of ArgonDigital.
It’s important to understand which business performance metrics are most important so that the stakeholders can prioritize the development work. Understanding specific metrics against the current processes will help highlight bottlenecks in the current processes.
KPI models (KPIMs) associate business processes with their important performance metrics. KPIMs are drawn as flowcharts, swimlane diagrams, or activity diagrams with KPIs overlaid on the related steps. This figure shows an example KPIM for a project to automate a loan risk-profile calculation process.
The most important business processes to automate are those that would make the greatest contribution to achieving the goals your metrics are indicating. Determine a current baseline value for each metric, so that when you automate the process, you can tell if they are improving as desired. Keep in mind that you might degrade certain business performance metrics to improve others. You might have to make trade-off decisions to favor one performance metric over another.
Tracing requirements to the process flow steps, which in turn are mapped to KPIs, allows you to prioritize which automation requirements to be implemented. Additionally, you might need to build functionality into the system to periodically measure the relevant KPIs to evaluate the effectiveness of the newly automated solution, raising a warning flag if a measurement is not improving. In the spreadsheet example, the system should measure how much time it takes to aggregate data inputs to determine if the system is achieving the two-minute goal. If not, further changes might be needed, or the goal needs to be revisited.
Business users often think it’s ideal to automate every manual process if and when possible. However, there are costs associated with all automation projects. Business analysis helps you determine which processes are worth automating and which are not.
Scope of Measurement
You don’t have to add KPIMs to every single process. However, you might need a baseline measurement for the overall process you are trying to automate, even if you are only planning to automate a portion of it. The baseline must correspond to the KPI you are tracking. That is, if your KPI is to reduce cost, your baseline must be the current-state cost. In addition to a baseline for the complete process, you might need to baseline each step in the process, even if your goal is to automate the entire process. The baseline information provides:
- Assurance that you are addressing the right process. You may have a good hunch that a particular process needs improvement, but the baseline measurement will confirm your hunch, especially if you can compare it to an industry standard.
- Verification that whatever automation you implement has improved your baseline measurement. If your measurement is not improved, then you need to start digging into the individual steps of your process to make sure you’ve identified the actual bottlenecks that warrant automation to improve performance of the overall process.
Remember all of these metrics are lagging indicators of automation success. Proper business analysis will help ensure the right things are automated and measured, well before you automate anything. See our next article for more on that.