The process mining maturity model

How good is your team at process mining? What should its progress look like?

The process mining maturity model is a way to guide your journey from

‘I think we could benefit from process mining‘

to

‘I can predict the likely outcome of every case currently traversing my core processes.’

Working with clients, I’ve identified five levels of maturity which capture the typical evolution of a process mining capability.

Each level describes what the likely state of process mining is, key challenges to overcome, and realistic next steps to take.

Process mining is a capability, not a tool

First, let’s discuss why talking about process mining maturity makes sense.

Process mining is not as simple as installing software.

Every process instance is different. Your company, your IT landscape, your activity sequences – the way you create value – are all unique.

You must be wary of off-the-shelf process mining solutions that don’t take your context into account. A cookie-cutter approach will only get you so far.

I strongly suggest you experience the process mining learning curve yourself to lay a solid foundation for a high performance capability through all the levels. It might be harder – not necessary take longer – but will pay off as you scale.

To summarise how process mining works: You get analyses via event logs. You get event logs via data models that transform system logs. Therefore, the data model is the critical success factor. The better the data model, the better the event log, and the better the end analysis.

When starting out, I recommend that you take great care building the initial data model. Start small, select a reasonable sub-process, take your time on the model. When you have a working model, then expand its coverage and improve/enhance the data with each iteration.

This sounds slow and difficult, but you will then be able to take the lessons you inevitably learn from this effort and apply them to additional processes later. With this experience, you’ll be able to spot issues, understand how to resolve them, and iterate your data model faster. This all means better outcomes with process mining.

Maturity is a function of your capability and experience. To increase your maturity level you need a cross-functional team, working together to solve the inevitable challenges of process mining.

You need people who understand your IT and data storage landscape working alongside people who understand how your processes work and how they add value to your customers.

This close collaboration is vital and is very difficult to outsource. If you build this team yourself – let them learn and get experience, advised externally perhaps (contact me!) – and your process mining capability will be built on solid ground, allowing you move up through the maturity levels, which we’ll now explore.

Process mining maturity model levels 1-5

The process mining maturity model is a sense check at the different milestones of a capability. At each milestone you can benchmark:

• Where is your organisation in its process mining journey?
• What should you watch out for?
• What do you need to do to improve?

The model helps you recognise what you can expect of your process mining effort at each level and what a realistic path for progression is.

Here are the five levels:

Level 1 – Process mining aware
Level 2 – Partial process coverage
Level 3 – End-to-end process coverage
Level 4 – Live process data
Level 5 – Predictive process analysis

Level 1 – Process mining aware

Your status:
You are starting from scratch.
A few stakeholders are excited about process mining.
There is limited technical understanding of how to do process mining.

Challenges:
You need enough technical knowledge to understand what it takes to add value with process mining
You need to create a cross-functional team: getting process experts, mining experts, improvement experts, and IT experts in the same room
Make sure you aren’t doing process mining because it’s a hot topic; ensure there’s a legitimate business need

Next steps:
Build the team. Get people together who are invested in the outcome of improved processes
Upskill so there is enough understanding of process mining across the team
Locate realistic business problems to analyse with process mining

Level 2 – Partial process coverage

Your status:
You’ve built a first analysis
People are getting interested the process map visualisation you’ve discovered
You’ve offered some interesting insights

Challenges:
As your analysis gets shared outside of the initial stakeholder group, people are underwhelmed by the analysis coverage (‘it’s not a full process’)
Some people are underwhelmed by the initial insights (‘we already knew this’)
You need to be ready to deflect an initial reaction of a ‘shiny object’ critique (‘looks great, but how does this help me?’)

Next steps:
Continue to set expectations: Explain what has been achieved. Acknowledge the limitations. Talk about the remaining data gaps and how you’ll fill them.

Add value: Work with process experts to make realistic recommendations. Don’t get carried away with the possibilities of process mining, stay grounded in attainable insights.

Accept and encourage critique: Getting to this point has probably taken a lot of effort, and you are wanting acknowledgement from the wider business. However, you’re at your most vulnerable just as your stakeholder group is at their most skeptical (because it’s likely to have expanded following initial analysis).

Level 3 – End-to-end process coverage

Your status:
You’ve created your first full, end-to-end process analysis
Your analysis is linked to value stream SLAs/KPIs
Root cause analysis is providing actionable insights

Challenges:
There are still data gaps (The Pareto effect: The last 20% of process coverage takes 80% of the effort).
Trust is still lacking in the data. This makes it hard to inspire action without executive sponsorship.
Add in business intelligence. Enrich the event log with additional case dimensions as interest from business units grows. Make the analysis relevant to them.

Next steps:
Keep talking to process owners: Engage with your end users. Where are they skeptical? Where are they interested?
Drive and demonstrate value: Don’t just flag process issues, suggest process improvements.

Level 4 – Live process data

Your status:
You’ve operationalised ETL after confidence in your data model has reached an inflection point: People believe what you’re saying about their process
Live data supports day-to-day process management
Process experts now rely on your analyses for their operations

Challenges:
Maintaining the data flow is now a vital task. This might be boring for your team, but stability is now crucial.
You need to continue your data model iteration: The process isn’t static, nor should your understanding of it be.

Next steps:
Don’t set-and-forget, maintain stakeholder enthusiasm. Don’t let dashboards go stale, keep them relevant.
Actively support end users. Ask them how process mining is helping them. Be proactive. Understand process goals so you can help it improve.
Encourage skeptical views and additional questions from SMEs. Allow them to continually stress-test your data set.
Evaluate event logs: What dimensions are redundant which should be added?

Level 5 – Predictive process analysis

Your status:
You offer predictive process analysis
Stakeholders use process mining for alerts and forecasting
Live process data triggers situationally aware automations

Challenges:
Staying strategic: there are different altitudes of analysis available now. Tactical reliance can pull you away from the bigger picture.
Building resilience for the process mining ecosystem and team. The business now relies on you, so short term operational robustness and long term knowledge management are crucial.

Next steps:
Can process mining trigger automations to take preventative action based on root cause analysis?
Can cases be triaged via Decision Model Notation (DMN) derived from process mining, and orchestrated via automation platforms?
Can process mining be expanded into your BPM efforts as part of a continuous improvement cycle?

To build maturity, move through the levels

The maturity model outlines a realistic path to create a capability that offers all the benefits of process mining. It’s difficult to skip steps; if you jump steps, you might fall down, biting off more than you can chew (to mix metaphors).

Obviously, you can build a capability in different ways: target live data of a key sub-process as a first step. But I’d avoid that, if possible. Each level represents a new challenge – technical and collaboration-wise – and mastering the previous level makes things easier overall.

The early lessons – overcoming painful experiences of data extraction, transformation, and event log preparation – lay the foundations for the later steps.

A recommendation: Keep a core team in-house, don’t outsource everything. It’s a false economy. If you work with vendors/consultants, don’t let their knowledge walk out the door when they inevitably move on. Remember: If all you are left with is interesting dashboards, what happens when the context changes or the process evolves? Who knows how to connect a new system and update them? These issues are certain to happen.

Regular reflection means continuous progress

Don’t get swept up in the process mining hype. It’s not magic.

Yes, it’s true that process mining gives you live, insight-rich analysis of your processes. But that doesn’t mean you instantly get that when you subscribe to a process mining tool; you have to start small and iterate towards the full capabilities of process mining.

As with any capability, it’s better to think of process mining as a mountain to climb, with level 5 at the top. Plot your path from where you are right now. Think of each step as a milestone. Set expectations accordingly; within your team and with your stakeholders.

Document your learning. When you start process mining it’s the start of a Centre of Excellence. You are establishing your business-specific best practice with each level you progress through.

Most importantly, foster a self-aware team. The secret sauce is the data model that creates your event log: that is the engine of process mining, and it is created collaboratively. That cross-functional team is vital, so establish it early, allow it to learn, and formalise it within a Centre of Excellence as soon as possible.

Process mining isn’t as easy as some suggest. But with the maturity I’ve proposed here, you can move through the levels and get the most out of this exciting and incredibly insightful capability.