KPIs instead of gut feeling Scaling and optimising automation with metrics

Related Vendors

Weissenberg Group

HASCO Hasenclever GmbH + Co KG

Renishaw plc

DIPROFIL Diamantprodukter AB

Measuring the value of automation is crucial to the long-term success of any automation initiative. Decisions based on gut instinct are not particularly effective in this regard. Only through a data-driven approach can automation leaders identify where automation graves have been created and where automation can be effectively deployed to deliver the best business benefits.

Meaningful automation ROI ensures the cross-functional support needed to scale automation projects. For each automated process, KPIs (key performance indicators) or metrics such as hours saved, cost savings and service delivery improvements can provide the necessary visibility into which automation candidates should be prioritised next to achieve the greatest gains. In addition, aggregating these metrics across the automation ecosystem also measures the overall impact of automation on business outcomes.

Metrics that companies should be aware of

There are a variety of KPIs that can be tracked as indicators of automation success that go far beyond cost and hours. Examples include mean time to repair (MTTR), the number of incidents or requests handled solely through automation, or the time it takes to complete service tasks. Customer satisfaction and employee satisfaction are also important KPIs for many companies. Which KPIs companies ultimately consider important depends very much on the strategic objective pursued with automation. Without KPIs, however, automation projects are like poking in the fog.

The essential KPIs that companies should know and have on their radar include:

Number of automated processes: The sum of all automated processes serves as an operational metric that indicates how RPA programmes are developing and growing, how adept staff are at identifying automation opportunities and prioritising them through to development.

Speed of automated processes: The speed metric refers to the average time it takes to complete an automated process. It quantifies the time and cost savings that result from a bot doing the job faster than an employee.

Utilisation of automated processes: Workload provides information on when and how often an automated process is executed. It indicates whether bot availability is being used around the clock — one of the main benefits of RPA. The more hours a bot is in operation, the more value it can deliver.

Average uptime of automated processes: Average automation uptime gives an indication of how often bots are available to do what they were designed to do. This is different from utilisation as it is a measure of a bot's ability to contribute to the expected business value at any given time.

Accuracy of automated processes: The metric of how often the automated process is executed with errors indicates whether the current automations justify continuing with automation projects. The question behind this is: Has the quality of the process output improved such that fewer errors occur than with manual execution?

Average processing time: Average processing time is about the time it takes a robot to process a transaction compared to a person. Digital workers are 3 to 5 times faster than a human. The average processing time of a process should be significantly reduced by using robots.

Volume variance: This metric refers to the difference between the predicted volumes used for ROI calculations and the actual volumes handled by the bots.

Break-fix cycles: Break-fix cycles indicate how often an automated process is interrupted and when it needs maintenance. Failures of bots, whether they need maintenance or repair, have a direct impact on ROI. When a bot is no longer productive because it is not running, it does not reduce costs or help increase operational efficiency. Break-fix person hours indicate how much manual effort in FTE hours is invested in fixing the bot.

Number of service desk tickets: Each time a bot breaks down, a service desk ticket is usually created. The more tickets received by the service desk, the greater the vulnerability of the technological environment, the data collected and processed, or the programming of the bot. Many tickets usually indicate a faulty environment that is less efficient and optimal than it should be.

Causes of failure: This metric gives companies an overview of why bots fail in the first place. It thus opens up the possibility of identifying gaps in automation practices that prevent bots from scaling uptime for maximum returns.

Scalability of automation: The cost and time required to add or duplicate a new bot to complete a specific task is compared to the cost and time required to hire and train a new employee to complete the same tasks, including labour costs.

Scalability of development: Development scalability is about capturing what percentage of the existing automation code can be reused in the new automation. In most cases, adding a new bot is easier and takes less time.

Expected business value: The expected business value metric essentially consolidates all the other KPIs. Basically, expected business value is the sum of all cost savings resulting from increased speed, utilisation and improved accuracy multiplied by the cost of an FTE over a period of time.

More does not always equal better

The range of metrics can vary and should reflect support for the different functional strategies of the organisation. The priorities set will inform how organisations design their automation projects and the importance they place on the performance and/or speed of bots in achieving business goals. In general, it is not useful to track too many metrics. A handful of meaningful metrics, collected over time and reflecting change, is often sufficient. They should be consistent with and support the overall goals.

* Dr Norbert Niemeier has been managing director of Weissenberg Intelligence at the Weissenberg Group since 2021.

(ID:49720381)