Selecting the most suitable Process mining Solution

Gartner defines Process mining as " designed to discover, monitor and improve real processes) by extracting knowledge from event logs readily available in today's information systems". -Gartner's Market Guide for Process Mining, June 2019

When I was first introduced to the term 'Process mining', just the reference and context- gave me a reminiscence of my days in the years 2007-12, as a Procure to Pay specialist, going to different client locations and leading Continuous Improvement oriented workshops. Many times I used Microsoft Visio, tools for modeling - many other times, it was a pen-and-paper workshop in which there used to a preparatory phase, where the participants were asked to capture sample data for a couple of weeks, so we can model the data on the process workflows, to simulate and see the outcomes/KPIs the process is generating , identify steps which are acting as bottlenecks and build an improvement plan to remove the bottlenecks and enhance process agility. This used to be a very tedious exercise, not just physically draining during the weeks of the workshop, but the preparatory phases as well. Since the participants, used to struggle a lot -capturing and collating the data points, like Throughput for each step, Cycle time, Exceptions/ Error rates, etc.

The accelerated pace of adoption of 'Digital Technologies' in the last decade (2009-2019) fundamentally changed the way IT assets are leveraged, managed and utilized. From a Process mining subject perspective, this happened primarily due to the following developments:

• Emergence of Advanced Process modeling tools- which are cloud-native, got simulation capabilities and are able to produce analytics on demand. What we see today in Process mining space, is a matured evolution of the initial prototypes (e.g ARIS, Lombardi, etc. from the early days). While some solution providers sold their capabilities to large companies (like Lombardi did with IBM), others decided to enhance their native capabilities to effect and develop an 'End to end' offering.
• Evolution of a Connected Application universe- the last decade also saw a shift from ETL (Extraction Transformation Load) data economy to a connected data economy- primarily enabled by open connectors, API ecosystem and exchange and finally concluding towards the strong capabilities offered by 'In-memory computing'. Mobility Apps only provided a robust use case for these developments. An Open interconnected Data universe, only meant- its easier for different tools to access the same source of data and realize what we call as 'One version of Truth'. This had repercussions much beyond the IT services domain. It opened the doors for simplified and real- data capture, which became the foundation for the development of Process mining.
• Developments in Analytics - as the world started shifting towards 'Data economy', the relevance and Innovation in Data Analytics -both at a Solution and at a Service level, opened up numerous possibilities, to unlock trapped value. This development had a trickling effect on Process mining too- as the commercialization, lower costs and easy availability of Analytics offerings, ensured this can be in-built in process mining solutions, to make it a holistic package. The ongoing wave of Predictive analytics is taking the evolution, further.

Current Market Landscape

Market analysts like Gartner, in majority of their reports talk about nearly 20 Process Mining solutions, which constitute the landscape. Each has got its own capability framework, with some having a heavy tilt towards a particular industry or business context, than the others. It is recognized that this market is evolving, as more and more organizations are adopting Process Mining, as a strategic lever to define their continuous improvement strategies and capture opportunities. At a practical level, to cause fitment into the core Process mining category, the solutions available must have the following basic capabilities:

• Process Discovery and Description- ability to perform Process modeling, Capture data, Identify which processes to mine and determine the KPIs required for relevant analysis.
• Process Execution - this is where the systems and event logs come into the picture. It also deals with customized data extracts, transformation scripts, data model and dashboards, along with an ability to compare process execution to original process design. Questions like 'How far the process is deviating', 'How much of variability exists' are answered here.
• Process Analysis- this segment of capabilities, deals with identifying bottlenecks and drilling down to find the last level root causes, Identification of improvement opportunities and automation potential etc. using Process analytics as the key driver.

Key criteria for selecting a Process mining Solution

Before answering this question, it is pertinent to recall that the primary asset, that Process mining relies upon is the IT event logs, across different systems, which capture the time-stamps. This provides a useful and relevant context to the overarching question- which solution to pick it up. In general, the following points can act as a guide in selecting the most appropriate Process Mining Solution:

• Are there any IT landscape dependencies- For example, SignaVio and ARIS have close affiliation towards SAP based landscapes. Needless to say, with the onset of HANA, this affiliation will only get stronger. If there are certain specific IT landscape changes in progress or planning, those have to be aligned with the overall Process mining solution strategy.
• What kind of end to end analytics platform is being used at the enterprise level- For example, Mehrwork Process Mining is more aligned towards Qlik, Process Analytics Factory has a similar affiliation with Microsoft Power BI. This is not an absolute condition, since using APIs and connectors, it is possible to have exceptions, but considering the fact - a lot of the value that Process mining releases, is dependent on the Analytics templates and Industry best practices, this gains the stature of a crucial consideration.
• Enterprise level Digital Strategy and Road-map- how much disruption you are anticipating from the Emerging Technologies? What kind of changes do you anticipate in your core processes and operating model? How do you want to leverage Data as a strategic asset in the future? These are relevant questions- more so relevant from the context of selecting a Process mining Solution.
• The choice of Managed Service partners- the Process mining market is still in its growth stage, so different managed service providers have built and developed mature capabilities using a variety of solutions, addressing different industries and markets. This is the factor that gives birth to potent use cases and subsequently credentials, which are proposed at the time of Solution discussions. Some technologically advanced companies like Accenture, developed their own Process Mining Solutions like MyConcerto -based on years of experience and skills their teams gained while developing solutions on other vendor platforms. It is a 'Best of both World's' approach, where the demerits of market existing solutions are eliminated, while enhancing the strengths encountered across variety of solutions. Mergers and acquisitions, further end up strengthening the value propositions, by establishing 'Centers of Excellence'. Different providers have different solutions as their preferred solution and have very strong competencies into the same.
• Pricing Model differences- of course for any selection, budgeting related decisions and constraints cannot be eliminated. Currently available market solutions have different pricing models and variants. For Example, the most popular one (almost synonymous with the term 'Process mining') Celonis offers pricing at a Process level. You end up paying a fixed price for the entire Process, regardless of number of users. Further, the price depends on complexity as well. On the other hand, Solutions like QPR offer by count of Analysts, Dashboards required at a monthly level. One interesting dimension, and quite impacting one for this factor is - do you want it On Public/Private Cloud or On-Premises? This will impact the overall TCO as well.
• Tactical considerations- in addition, some tactical considerations are important as well, while selecting the right solution. Examples- What kind of Process conventions are acceptable in your enterprise, Which are the Geographies under scope, What kind of language the tool needs to support, How do you cover activities which do not have a structured transactions or Event log files (like Excel-sheets, Emails), What kind of Data cleansing support is required, How much integration is 'Good integration' across processes, What kind of references you are seeking, Which industry expertise is required, etc.

It is important to note- that the above is not an exhaustive list of considerations, and the final decision will depend on a combination of lot more factors. But this can serve as a conversation starter, as a primary guide to define and identify those factors specific to an enterprise. The decision to select a particular Process mining solution, must be based on the strengths/ weaknesses of the existing application landscape, Data architecture and overall Enterprise strategies related to Digital evolution using Data as an asset.