Deloitte and Celonis collaborated on a project that combined Process Mining and Task Mining to drive transparency into the manual effort and costs behind the global Order-to-Cash process at a large international toy manufacturer.
The project led by Deloitte, was designed to enable the client to understand their baseline productivity and identify bottlenecks and complexities in the way their employees worked across the end-to-end global Order-to-Cash (O2C) process. Using the power of Celonis to drive transparency across multiple systems, teams, and geographies, to provide unparalleled insights into where time was being spent and why exceptions were occurring throughout the O2C lifecycle.
We caught up with Craig Sumner, a Senior Manager at Deloitte UKI, to walk through this project and understand how they were able to quantify the true cost of operating the O2C process and where the client could drive simplification and productivity across their business .
The Background: The customer built a program to standardize their global shared services. To date the customer has rationalized their ERP Systems into one global instance and have begun to implement Intelligent Automation. The customer has also deployed Celonis Process Mining across its O2C process with the aim of simplifying work.
"The key challenge for them was how to take the data that they had identified within Celonis Process Mining and drive actionable insights or identify value pockets across their end to end O2C process where they could simplify, standardize or automate business processes, with a key focus on employee productivity, customer engagement, and employee satisfaction," said Sumner. "It was about improving people's jobs, reducing the mundane, repetitive tasks they are doing, whilst also driving that simplification and standardization to improve quality."
Finding gaps in processes. Process Mining was able to show transaction data from their core ERP in an easy to absorb format, however lacked visibility into the true nature of human work because employees were frequently using applications outside of the ERP system, such as Microsoft Outlook, Excel, and Teams. The Deloitte team designed a task mining pilot that would cover the broad scope of the global O2C process and complement the process mining data already available. This was with the aim to shine a light on this ‘dark data’ to better understand the employee experience, ways of working and productivity
Process mining data was used to map out the teams and users that were working on the end-to-end global O2C process. That move "helped us design and scope a sample size that would give us a representative view of the activity being performed across these different teams in the end-to-end global O2C process," said Sumner.
Best practices: Sumner added that a task mining study requires collaboration with key groups in a corporation including IT, Security, HR and Legal to review data protection, privacy and anonymity requirements.
"We agreed with all of the relevant teams that no personal data would be captured, that everything would be anonymized, and one of the important points of the study was that no data would be traceable back to individual team members. All of the data was aggregated at the team level, ensuring that the analysis did not pinpoint individual performance issues or individual people," said Sumner.
Change management and communications were critical to educate senior leadership and team leads, who in turn educated employees about the project. Employees were even given the opportunity to manually stop recording every day if they needed to use their desktops for personal use or work outside of the O2C process. "The employees were quite keen to see some of the results, which we played back to individual teams and users," said Sumner.
Sample size for task mining: Sumner said the customer had about 220 employees working on the global order-to-cash process. Deloitte identified users based on the total number of transactions being processed. "We were able to work out approximately how many unique users were touching the end-to-end business process, and then within that, which groups of those users were relevant to the work that we were performing," said Sumner.
Deloitte ultimately had a sample size of 40 users willing to be in the task mining study for a minimum of 10 working days. "For the different groups that we recruited, we obtained at least 10 days’ worth of data. Some groups, we got slightly more because they were recruited early and they were happy to continue recording task mining data as the project progressed, but the most we had was 15 days," said Sumner.
Volume vs time spent: In many ways, process mining demonstrates the volume of transactions based on system data, while task mining addresses the manual effort and time spent on executing tasks and activities across a process outside of core systems, Sumner explained:
"The easiest way to differentiate between process mining versus task mining is that process mining looks at the system-based data from one or many core systems that are used in an enterprise process. It's the actual system logs, the activity logs, within that system, that you're mining to show you what activity is happening and when. The key challenge is that this information doesn't necessarily represent the time that a human user spends performing an activity within that system. If you see an action log with one time stamp on it for opening an order, for example, and then another time stamp for the order being submitted, that's not reflective of how much effort has been expended processing that order, because it's just within one system. Task mining enables us to capture all of the activity happening on the employee’s desktop, providing the insights into how much time is really spent processing a transaction between those time stamps and where the work is being done in the application stack."
Traditional consulting methods such as time-in-motion studies and questionnaires were used previously, but task mining based on actual data recorded from the human user activity on the desktop provides detail down to individual fields which far surpasses these methods for accuracy, quality and speed of implementation. These insights can be delivered in as little as 8 weeks covering 100’s of users, more than 50% quicker when compared to traditional consulting methods.
The findings and data: Sumner said the real value delivered was transparency into workforce productivity and a full 360-degree view of an individual transaction through the end-to-end process. The customer found that employees had productive time of 5 hours a day with 20% of that time spent in the ERP system.
"The remaining 80% of their time was being spent across other applications outside that core ERP, including 54% of the time being spent in the Microsoft stack. So that's things like Excel, Microsoft Teams, Internet Explorer, or Edge and Outlook," said Sumner. "We were quite surprised when this particular client is supposed to have a highly optimized ERP, that a lot of activity was being spent elsewhere. And you can then use that data to ask informed questions. Is this the right thing for those users to be doing? Should this activity be performed outside the ERP? Or are they not fully aware of the features, functionality, and value that the ERP could deliver? And is some of this activity redundant?"
Celonis applications were able to map the process journey including transitions between applications and screens to uncover multiple copy and paste transactions as employees hopped between applications to address process problems. When the process mining and task mining data was combined, the customer was able to quantify the real human cost of processing these exceptions.
"We found that 60% of the time across the global O2C process was being spent on order and fulfillment, and the majority of the time was being spent on changing orders. This data enabled the client to quantify exactly how much these challenges were costing and where in the end-to-end O2C process to focus and drive change. The next step would then be to do detailed root cause analysis, based on the new data that was available to them, to reduce the volume of order changes upstream,” said Sumner. “This data can even be used to underwrite existing investments in new systems and process changes or build new business cases for transformation."
Transparency and next steps: The customers ambition was to replace static HR and Finance estimates for productivity with objective task mining insights to truly understand the costs for executing their global order-to-cash process. With continuous improvement, the customer can find hot spots in this process and justify investment to tackle these issues.
"When it comes to prioritizing future work and identifying next steps, the customer now has a full, transparent data driven assessment that they can use to justify investment and prioritize where to go next. A lot of the next steps might involve simple changes. It could be process improvement, retraining the workforce. It could be optimizing the ERP further, or it could be more transformational changes, like implementing a new platform, making major changes to their order and sales cycle, or using tools like Celonis EMS to automate decisions and actions across the end to business process."
Innovation: Celonis supported Deloitte by creating new data models and dashboards in Celonis to produce this output focused on productivity and simplifying ways of working. The study was designed in collaboration with Deloitte and Celonis Data Engineers to achieve these unique and valuable insights across a broad, end to end process with a large sample size. Performing the data interpolation and mapping to the Deloitte O2C process taxonomy enabled the client to visualize where most of the time was being spent across the end-to-end process.
If you are interested in learning more about how this type of pioneering Task Mining work can deliver value to your organization, please reach out to Deloitte and Celonis.
Craig Sumner – firstname.lastname@example.org Intelligent Automation Lead, Deloitte UKI
Snaedis Walsh – email@example.com Client Partner, Finance and Performance, Deloitte DK
David Wright – firstname.lastname@example.org Intelligent Automation Partner, Deloitte UKI