⚠️ Important update: Workfellow is now part of ProcessMaker
Read update

Optimizing processes: why the "process mining first and task mining next" approach doesn't work.

Work is never straightforward and is never done within a single IT system or business application. Try to recall your latest workday - how many different tabs did you have to open? How many IT systems – software, databases, and platforms – did you have to use? An average white-collar worker has dozens of documents, third-party portals, apps, or government websites, which they use to do their day-to-day tasks. 

Processes = railways? 

The old way of thinking about processes has been like that of the railways. Railways are very simple and standard. The train runs along the railway and then stops, leaves some people and picks others and then runs again till the next station. Much like the trains go, people used to think of business processes the same way: let’s say, the invoice approval process follows a particular logic and happens inside a single train – the invoice approval software. 

However, that idea does not work for a vast majority of processes that go very differently from a straightforward railway logic. We assume that there’s a structure and logic to things, but in reality, most of the processes are complex, don’t follow any particular logic, and have lots of variations every single time they’re executed. 

Let’s have a look at this example. A customer service team receives hundreds of thousands of customer tickets about thousands of different products and services across tens of different regions. In this scenario, there are certainly many different processes involved. For a certain product, the employee will have to go to a certain system and check certain details and retrieve necessary information from contracts, databases, customer reports, etc. Those processes might involve tens to hundreds of different business applications. This path can be different every single time for every single product or service. 

90% of the work happens outside the core IT system 

Different functions like finance & accounting, payroll, and procurement have core IT systems that they use to manage daily business activities. It could be a customer service platform like Zendesk, a CRM platform like HubSpot, or an ERP software like Oracle. It’s believed that most of the work is supposedly done there, and therefore, process mapping and process mining activities are usually done only within these main IT systems. However, around 90% of the process happens outside that core system, and only the remaining 10% is done directly within the main system. 

Reality hits different

90% of the work happens outside the IT system

As an example, let’s look at the typical Finance team. A rough approximation would bring us to the following division of work time: around 30% of the time is spent on Microsoft Teams, 30% – on communication tools like Outlook, 30% – on Excel, and only 10% – on the main system Oracle. What is done on the core system is always very straightforward, fast, and oftentimes highly automated, thus it doesn’t take much time. 

In an invoice approval example, one would receive the invoice, investigate it based on the pretty straightforward and set process, and approve it - all within a few minutes. However, it takes much longer when one has to leave the core system and use some other tools and applications. For example, if there’s a tricky invoice and you have to ask your manager about it, then writing an email, sending it, and waiting for the response is what might slow down the process considerably. The actual button to approve the invoice takes only 2 seconds on the core system, but the emails from both sides considering the waiting time might take up to a whole day in total. 

Why does traditional process mining need to be transformed?

Process mining is like a lottery where you only know 1 of the numbers and the other 9 are unknown. You can get some limited understanding of the process and compliance and set up the rules, but it would be a lot of guesswork on a larger scale. This is especially true for work-intensive processes with many small tasks that require work outside the core system. 

Workfellow looks holistically at how business is executed, not focusing solely on individual IT system events. Its analysis aims to answer the challenges such as: How much capacity do we need to run our operations? How much time is spent on each IT system? Which IT systems might need to be substituted? Which systems should be used more to drive efficiency? Which new systems should be integrated? Workfellow looks at the business holistically, delivering what is expected from process mining and going much beyond that. 

Why is task mining not enough to cover the shortcomings of process mining? 

If the process mining is not able to get visibility over the tasks, task mining seems to be a reasonable way to cover the shortcomings of it. But is it so?

Task mining was initially born with the goal of employee monitoring and surveillance. It is especially used by team leads that want to monitor the performance of their teams and how efficiently they work. However, task mining fails to comprehend the entire process and understand how to do process transformation. 

Task mining can provide data on how much time people have spent on SAP or Excel or other applications. However, it fails to understand what people did and why they were there, what process they were engaged in, and which transactions they were working on. Task mining can only tell if the work was efficient based on what the person was doing on the computer, but not if it was truly productive.

Efficiency ≠ productivity. An employee can be really efficient and for example, type really fast, construct emails fast, and do all kinds of tasks highly efficiently. However, that does not say anything about how much valuable work they have done or how productive they were. This happens when only the time factor is understood, while transaction-level understanding is left behind. 

Workfellow helps track end-to-end processes

Process mining is very good for processes that are done in systems like SAP, for example. However, even the most SAP-centered functions will have some other systems, software, and applications involved. 

Whatever IT landscape and chaos you have – whether that’s custom systems, fragmented processes, or complex landscapes – Workfellow’s Work API can map out the reality of the processes. Process mining can see 10% of the process - the parts that are done in the core IT system, and Workfellow can see this 10% and the remaining 90% as well. 

Written by

Kazyna Turdibayeva

Marketing manager at Workfellow

Interested in enterprise automation?

Check out Workfellow’s solution which is your one-stop-shop for discovering all the automation opportunities.

Get a Free Trial