Interview with Frank van Geffen, Process Innovator, Advanced Data Analytics, Rabobank (the Netherlands) and Chairman of the Special Interest Group on Process Mining within the Ngi-Ngn (Dutch society for ICT Professionals).
Frank van Geffen – LinkedIn: https://nl.linkedin.com/in/frvangeffen
Frank van Geffen is a process mining expert and one of the leading innovative business consultants and works for Rabobank in the Netherlands. Frank has introduced process mining in Rabobank and completed numerous process mining projects, resulting in millions of Euros in cost savings. It is a pleasure to have Frank here for a chat.
John: First, thank you very much Frank for being willing to do this interview and spending time with me on this exciting subject, Process Mining.
Frank: You’re welcome. I am very happy to share my thoughts and experiences.
John: Frank, can you tell me a little about your role at Rabobank?
Frank: I am a Business Consultant in the department for Advanced Data Analytics within Rabobank, specialising in process mining. The department specialises in advanced analytics spanning from data analysis, advanced analysis, machine learning, text mining, and process mining - all under the same umbrella; Big Data.
John: What do you use process mining analysis for in Rabobank?
Frank: Process mining is one of the tools used in the Advanced Data Analytics department, and we use it for a lot of things. We explore process mining tools as a new way of looking at processes. We found that process mining was leading us to results way faster than the usual approaches, such as machine learning or manual process analysis. In addition, process mining gave us many new insights we did not have easily at hand previously, e.g. overview of ineffective processes, bottlenecks in processes and costs.
John: That sounds interesting. What is your main driver when initiating a new process mining project?
Frank: When we started out doing advanced analysis and using process mining in the organisation we had a very firm success factor: Go out and create some real value. We were not interested in applying these new tools and techniques and waiting five years to see results. So, we gathered together with managers and identified their urgent challenges and from that we discussed what the main driver should be. It turned out that the focus was on cost reduction. Keep in mind that we started out in the middle of the financial crisis, so cost reduction was a very natural element.
John: That certainly makes sense. Is cost reduction still a main driver when doing process mining projects in Rabobank?
Frank: It is, but another urgent challenge not directly related to cost reduction is also taking advantage of process mining techniques, for example working together with LEAN Six Sigma projects in order to apply facts based improvement or simply getting a grip on a primary process.
John: Getting a grip on a process and really understand what customers are doing, can be very valuable insights, even though it is maybe not directly related to cost reduction. Can you give an example of a case?
Frank: Yes, for example working with our mortgage process. Here we used process mining to really understand the process and especially understand what the customers are doing in the process. We had known KPIs attached, but during process mining analysis we found that process mining was able to acknowledge/proof some of the KPIs - but the analysis also led to adjusting some of the KPIs.
John: Interesting. You mentioned that you were working closely with LEAN Six Sigma projects. Can you elaborate on that?
Frank: Sure. A very interesting lesson learned is that there is a high degree of benefit when using process mining in LEAN projects, because the improvement initiatives become facts based. We also found that different profiles are needed when applying process mining to LEAN projects - to put it in other words; in our experience process mining consultants should not do LEAN projects and LEAN consultants should not do process mining, the best results emerges when the projects involve both LEAN consultants and process mining consultants in a joint effort. We find that LEAN consultants and process mining consultants is a very powerful combination.
John: That is a very interesting point. I would like to go a step back and ask you, how you got started and how you managed “selling” the idea of process mining to Rabobank.
Frank: Well, when I introduced process mining some years ago, process mining was a very new technique. I introduced process mining bottom-up trying to present it for people involved in projects or processes that could benefit from it. As mentioned before things got going when I presented it for managers with urgent challenges, and made some proof of concepts for some of the cost reduction projects. Another important aspect was that about the same time of introducing process mining in the organisation, an initiative around Big Data evolved in the organisation. Back then (and also today) Big Data was a huge trend. A lot of attention was on Big Data and how we could improve our business taking advantage of Big Data. That led to Rabobank establishing a Big Data proof of concept initiative, and indirectly that also led to establishing the Advanced Data Analytics department, where I am still working.
John: Why do you think there was such a huge attention towards Big Data?
Frank: First of all it is a huge trend, and the financial sector is indeed entering the internet age with a shift of business towards digital business. We need to stay ahead of competition, so one direction was to see if all of our data could give us new insights; thus, new competitive advantages.
John: How did you approach performing the Big Data proof of concept?
Frank: We formed a small group of people. All enthusiasts. Then we established a set of success factors. The goal was to explore these new techniques and tools, and see what advantages and insights they could bring. The proof of concept was executed in around a year. Process mining turned out to be very valuable, bringing new ways of looking at processes.
John: I like the idea of forming a group of enthusiasts that get time in the organisation to prove that data and process analysis is a way forward. What is the most important advice to others, who would like to do the same?
Frank: Well, one good advice is to start out small. Focus on areas or processes you can get a grip on and where the data is accessible. Don’t start analysing a process where the process data is scattered over multiple systems. Also, don’t focus on technology. Technology should never be the main driver and get all your attention. Instead focus on what value you can bring.
Another advice is to keep going. You will definitely face obstacles from time to time - but always keep going, keep pushing and benefits and rewards will emerge.
John: What findings did you identify in the process?
Frank: One thing was that the process is explorative and as you go you will discover, that you will need more information than first anticipated, for example. Another finding was that having good enthusiastic evangelists inside in the team is very important. In addition, I found, as mentioned before, that LEAN people together with process mining people is a very powerful combination.
John: So, you are suggesting forming a team, that should lead the Big Data / process mining initiatives?
Frank: Yes, absolutely. Both that but also forming teams when approaching projects. We formed a group of enthusiastic evangelists with a passion for Big Data, process mining and data drives business in general. That turned out to be powerful. Create a small team that can perform, and be aware that it takes time to get things up and running. You need to create a movement.
John: I definitely agree. We have already talked about how to approach management and “sell” the idea / open up their eyes for the importance of process mining, but what obstacles were you facing when performing the actual process mining projects?
Frank: There are many obstacles to overcome, but a major one is of course on how to get the data. Even though most process steps are already stored in the IT systems, it can still be challenging. One thing is the extraction of the data, but sometimes the data exist in various formats and sometimes some of the attributes are missing. Some good advice, when trying to extract meaningful process data from the systems, is to start identifying the case-ids; unique keys identifying the cases for which you would like to analyse the process, and then as a next step try finding the corresponding activities, timestamps, and attributes.
John: That is clear. You mention a typical process mining obstacle, namely not having enough data or not having data in a sufficient quality. Our Process Mining Manifest addresses that as a need for on-going data improvement - treating your data as first class citizens. Do you have such initiatives as well in order to continuously improve the effect of process mining analysis?
Frank: We do, and it is certainly another obstacle. We try to address this by lobbying inside our organisation for more data and better quality. Hesitation from IT is sometimes, that adding logging will have a cost on performance and storage, but frankly, nowadays storage costs so little and enterprise architecture can easily overcome our logging needs.
John: Sounds like the way forward. What tools are you using for process mining?
Frank: We use a lot of different tools. For process mining, Disco is our favourite tool because of its usability. We also use ProM, but that is primarily for explorative use, as it is more cumbersome and less user-friendly.
John: I agree. Some BPMS suites incorporate process mining functionality, whereas Disco is a stand-alone process mining tool. Do you favour approaching process mining analysis with stand-alone tools?
Frank: At the moment we favour stand-alone tools. We try to integrate process mining analysis in ways of working and, as mentioned earlier, we try not to focus on technology, but to focus on business value.
John: That makes sense. Frank, it has been a pleasure talking with you and hearing more about how Rabobank got success with process mining. Thank you very much!
Frank: You are welcome!