Process mining software is designed to analyze logs and other data from processes in order to identify process improvement and automation opportunities. For content-centric processes in particular, the new process mining tools can be invaluable in helping you analyze, optimize, migrate, and monitor them. This post explains how to evaluate and select the best process mining tool for your requirements.
What Questions Should You Ask?
We recommend an efficient process mining evaluation approach that asks and answers the following questions:
What business areas and processes are you trying to address? Are you interested in targeting core line-of-business processes for e.g. banking, insurance, manufacturing, or utilities? Or processes in corporate functions like procurement, sales, accounting, finance, or HR? Or information systems like your document and data capture, ECM, or BPM/case management systems? Or your ERP, CRM, claims processing, or other business systems?
What do you want to do with those areas and processes? Are you most interested in process discovery? Conformance analysis? Process improvement. There are several options within each of these three. Determine which of these you want to prioritize.
What are your process mining functional requirements? Given the above (1 and 2), you will need particular process mining capabilities for data collection, process discovery, process enhancement, and other features and functions.
What are your process mining technical requirements? Given the above (1-3), and other factors about your organization and IT environment, you will have particular technical requirements regarding architecture, performance, security, etc.
How do I test and select the right solution? Finally, given the above (1-4), how do you test a very short list of candidate vendors to assess if they can get the job done for you?
1. What business areas and processes are you trying to address?
Start by thinking about what kind of organization you are. Are you in financial services, insurance, manufacturing, life sciences, utilities, or something else? Are you interested in targeting vertical line of business processes and functions, or horizontal corporate processes and functions? Within any of the processes and areas, do you want to focus on the information systems like your document capture, ECM, or BPM/case management systems; or your business systems for ERP, CRM, claims or underwriting?
To provide an example: we at Doculabs find the line of business processes in insurance and financial services to be the among the best targets for improving with process mining. In particular:
Insurance: sales and underwriting, policy administration, claims
Investment Management: account opening, servicing, money in/out
Banking: consumer lending
Within the above use cases are a mix of specialized business systems along with process automation and content management systems from IBM, OpenText, Pega, Sungard, Appian, Kofax, and Hyland.
Other examples of good targets for process mining: horizontal corporate processes within areas such as procurement, sales, accounting, finance, and HR. These involve business systems like ERP and CRM from e.g. SAP, Oracle, Salesforce; along with many of the same process automation and content management systems mentioned above.
2. What do you want to do with those areas and processes?
At the most general level, process mining lets you do process discovery, process conformance, and process improvement. There are several options within each of these three. Determine which of these you want to prioritize.
Process discovery: you can get a comprehensive understanding of your as-is processes, identify how your actual processes deviate from the desired ones, or discover the best processes to automate. The use cases in this category focus on structure of the processes, how frequently work follows one path rather than the other paths, and the distribution or percentage of cases following each discovered path.
Process conformance: you can assess whether the process has the intended behavior in practice. For these cases you need an additional input besides the event log of the process to be analyzed. You need some form of process documentation, process map, or rule. Typical use cases are for GRC (governance, risk, and compliance) and evaluating a current state process against a future state design or requirements. The use cases in this category focus on exceptions from the normal path, degree to which the rules and regulations are obeyed, or comparison of the actual process with an explicitly documented process.
Process improvement: you can monitor performance by tracking KPI measures, identify process bottlenecks, and use AI-based capabilities of process mining tools to get specific recommendations to increase the performance levels. Most process improvement use cases relate to time, e.g. the throughput time of cases, the slowest activities, or the longest waiting times.
3. What are your process mining functional requirements?
Given the above (1 and 2), you will need particular process mining capabilities. The following are some examples of the standard and most important capabilities:
Collection - Data Extraction and Transformation
What volume of events and cases can the tool import, and how is data loaded?
What systems can it import from? (e.g. document capture, BPM/case management, ECM, RPA, ERP, CRM, etc.)
Does it support client side importing? (e.g. spreadsheets and documents, copying and pasting data, importing from desktop systems)
Can it connect to multiple data sources and combine their data?
Discovery - Visualization and Analysis
Does it provide standard analyses and components, and can you customize them?
Can you create new analyses?
Can it visualize and analyze many processes and variations simultaneously, with control over different metrics?
Does it provide process notation capabilities, customer journey mapping, and a documentation repository for process maps and documentation? (This is essential for conformance evaluation)
Does it suggest root causes for process deviations?
Does it use AI/ML to recommend ways to improve process performance? (AI/ML capabilities are essential for process improvement)
Enhancement – Insight Operationalization
Does it identify actionable patterns, trends and changes?
Can recommendations be executed within the solution?
Can it identify processes and tasks for automation?
Reporting and Monitoring
Does it offer integrated reporting capabilities?
Does it support setting specific objectives for KPIs and automatically measuring and reporting on progress? (KPI capabilities are essential for process improvement)
Does it provide live reporting of KPIs?
Usability
Does it provide an intuitive user experience for all users, including business users, analysts, and developers?
4. What are your process mining technical requirements?
Given the above (1-3), and other factors about your organization and IT environment, you have particular technical requirements regarding architecture, performance, administration, security, etc. The following are some examples of the standard and most important technical requirements:
Architecture
Is the deployment approach cloud, hybrid, or on-premises?
Is the product containerized?
How does it integrate with other IT systems?
Are any installations necessary?
Performance
Can it support multiple simultaneous users?
Can it handle large data sets?
Administration
Does it define user roles? If so, are user roles customizable?
How does it handle data access permissions? How granular is it?
5. How do I test and select the right solution?
Finally, given the above (1-4), how do you test a very short list of candidate vendors to assess if they can actually fulfill your process mining objectives? Besides evaluating whether the product adequately does the job, you’ll need to assess the vendor’s overall stability, fit for your company, professional services and partner network, expertise, strategy, and ability to support you in the short and long term.
The field of solutions is rapidly growing, with both product expansion and consolidation. Currently the top vendor, Celonis, owns more than 60% market share and Doculabs tracks about twenty offerings, including Celonis, Software AG, ABBYY, Nintex, UiPath, Minit, iGrafx, myInvenio, Signavio and Everflow.
The most important thing you can do after selecting a small short list of one to three candidates is to do a trial or proof of concept. This involves selecting a good use case or two and using the tool to prove out the impact and value while evaluating the tools’ viability and practicality for your team. Please contact us if you’re interested our help in evaluating your options.