From Reactive to Predictive Process Management
5 likes•1,000 views
The document discusses the evolution of business process management (BPM) from reactive to predictive methodologies, emphasizing the importance of governance, standardization, and the integration of processes across organizations. It uses case studies, such as a financial services provider and a global investment bank, to illustrate challenges and strategies for embedding BPM into organizational culture. The text also highlights the significance of real-time and predictive analytics in improving processes, particularly in contexts like college admissions and operational excellence.
![9
“Processes are actions on objects. […]
The true nature of an organization
follows the flow of its processes.”](https://image.slidesharecdn.com/2016-09-23fromreactivetopredictiveprocessmanagementfordistribution-160921151054/85/From-Reactive-to-Predictive-Process-Management-9-320.jpg)
From Reactive to Predictive Process Management
- 1. Dr Michael zur Muehlen Center for Business Process Innovation School of Business Stevens Institute of Technology Hoboken, New Jersey [email protected] @mzurmuehlen The Mirror and the Road Ahead From Reactive to Predictive Process Management
- 2. 2 “If it doesn’t make three people angry, it is not a process” Beyond Reengineering (Michael Hammer, 1996) Corollary: We tend to pay attention only after something broke
- 3. What is the Price of a Cup of Coffee? 3Source: Starbucks
- 4. What is the Value of a Coffee Lover? 4Source: Starbucks US$ 14,000
- 5. 5
- 6. 6 1. Capture Best Practices 2. Reduce Complexity 3. Standardize Work
- 7. 7
- 8. 8 1. Specialization 2. Hierarchical Structure 3. Centralized Control
- 9. 9 “Processes are actions on objects. […] The true nature of an organization follows the flow of its processes.”
- 10. 10
- 12. Premise Financial Service Provider with 5 Lines of Business Operating in 3 Countries Various levels of maturity across LoB, Technology, and Location Goal: Realize efficiencies by creating one holding company, shared back-office services Question: How do we stand up a BPM organization? 12
- 13. Where do we start? 13 Governance Method IT Support People CultureStrategic Alignment Process Roles and Responsibilities Process Design & Modeling Process Skills & Expertise Process Values & Beliefs Process Improvement Plan Decision-making Processes Process Implementation & Execution Process Education & Training Process Attitudes & Behaviors Strategy & Process Capability Linkage Process Management Standards Process Improvement & Innovation Process Knowledge Management Leadership Attention to Process Process Output Measurement Process Metrics & Performance Linkage Process Control & Measurement Process Collaboration & Communication Responsiveness to Process Change Process / Business Architecture Process Management Controls Process Project & Program Management Process Management Leaders Process Social Networks Process Customers & Stakeholders Business Process Management Maturity Process Design & Modeling Process Implementation & Execution Process Improvement & Innovation Process Control & Measurement Process Project & Program Management Additional Facets (can be considered under Method) Business Rules Management Process Data Management Process Taxonomy & Ontology Process Standardization & Harmonization
- 14. Initial Focus 14 Governance Method IT Support People CultureStrategic Alignment Process Roles and Responsibilities Process Design & Modeling Process Skills & Expertise Process Values & Beliefs Process Improvement Plan Decision-making Processes Process Implementation & Execution Process Education & Training Process Attitudes & Behaviors Strategy & Process Capability Linkage Process Management Standards Process Improvement & Innovation Process Knowledge Management Leadership Attention to Process Process Output Measurement Process Metrics & Performance Linkage Process Control & Measurement Process Collaboration & Communication Responsiveness to Process Change Process / Business Architecture Process Management Controls Process Project & Program Management Process Management Leaders Process Social Networks Process Customers & Stakeholders Process Design & Modeling Process Implementation & Execution Process Improvement & Innovation Process Control & Measurement Process Project & Program Management Additional Facets (can be considered under Method) Business Rules Management Process Data Management Process Taxonomy & Ontology Process Standardization & Harmonization Business Process Management Maturity SharedInboundProcesses SharedOutboundProcesses Enabling Processes (e.g. Accounting) Enabling Processes (e.g. IT) Enabling Processes (e.g. HR) Governing Processes (e.g. Strategy) Governing Processes (e.g. Risk & Compliance) Business Line 1 Business Line 2 Business Line 3
- 15. Problems of Procedure-Making 15 clarifying policy intent; publicizing generalizing useful information that was formerly the possession of a few employees; standardizing costs and man-hour requirements; and training new person- nel by the use of procedural manuals. Problems of Procedure-Making. The central problem in procedure- making of any kind is how to combine experience in procedure per se with [knowledge about and an adequate "feeling" for the operations which par- ticular procedures govern. To hard-working operating officials, the "proce- dural analyst" may seem an uninformed busybody he cannot possibly know as much about the operation as those-on-the-job, it's none of his business anyway, and if he really wants to be helpful as he says he does! let him pitch in and help with the work that's piling up. To the inquiring procedural analyst, in turn, operating officials may seem short-sighted, self- centered, narrow-minded each interested only in his task of the moment and utterly lacking in organizational perspective. Failure to solve satisfac- torily this central problem of combining diverse outlooks has two results. Either the procedure-making organ, if bolstered by formal authority, promulgates procedures that are useless and often disregarded; or no pro- cedural changes are made except through tedious evolution or violent revolution. Let it be admitted that the task of the procedural analyst is frequently
- 16. 3 Rules of Process Analysis 16 Don’t be Judgmental Don’t Solutionize Don’t Overstay
- 17. The Art of Writing Procedure 17 procedural by organizations doing similar tasks? Can the job be done effectively with the present organiza- tion or should change be recommended? If a particular method appears efficient, can it be reconciled with stautory requirements? Is the job one that can be done best by using ten skilled or fifteen unskilled employees? Such are some of the questions for which answers must be found. Intimate knowledge of the facts is essential, but hardly less important is "writijig^up" the procedure. For this, some of the art of the playwright is necessary, for good written procedures are rather like the script of a good play. The "characters" must be introduced and identified, and they should have an organic role in the production. Entrances and exits must be planned with purpose, stage directions must be given, and so forth. Pro- cedural materials should be brief, clear, and concise. All the tricks in the writers' and printers' trades should be employed, for to learn their lines quickly the players must be induced to read them. Unfortunately, proce- dural manuals are frequently as dry and forbidding as the Sahara. 7 Procedures must be installed. Even the best-written materials do not suffice of themselves. Educational campaigns must be undertaken, incentives offered, sanctions devised, methods for apprehending violators worked out, test runs made to discover "bugs," and follow-up inspections planned. All possible devices for breaking old habits and creating new ones must be
- 18. Types of Knowledge Tasks 18 Task Type Description Context Analysis Focus Typical Labels Gather Obtain data, often from multiple sources Typically found prior to decision tasks. Mirror image of Inform tasks Data sources, storage format, transmission medium, relevance, currency, signal-to-noise Receive, look up, retrieve, ask for, request, get, etc. Decide Combine inputs with business rules to either assign outcome value or select path of action Often (but not always!) precedes branching moments in process Input data, rules, output data, default values Approve, review, sign, check, authorize, determine, etc. Inform Notify process participant of data value or decision outcome Often after a decision or at the end of a process. Mirror image of Gather tasks Communication medium, message format, message timing, role of recipient Send, email (verb), post, etc. Update Persist data, change stored values, synthesize data, transcode/transform data Often after a decision or at the end of a process Output format, default values, data quality checks Enter, update, save, store Manage Organize the work of others, allocate resources, ensure process progress Meta-tasks that remind participants to act, chase external parties for data, and handle escalations Affected process performance metrics, efficacy of management actions Remind, chase, assign, staff (verb), monitor
- 19. Suggestions for Process Improvement 19 Task Type Description Context Example Eliminate Remove an element from the process Identify wasteful tasks, redundant or irrelevant information Eliminate a participant from the process, eliminate a task, eliminate a data element Reduce Reduce the weight of a process element Identify things that are currently mandatory that could be made optional. Reduce dependencies on outside parties, data sources Create triage points, move from mandatory controls to statistical sampling, bring data sources in- house Increase Strengthen a process element Increase the weight of value- adding process elements Allow real-time monitoring by the process customer Create Create new process elements Add processes, tasks, resources, information where they can add value to the business Create new contact channel that enables self-service by the customer Modeled after Blue Ocean Strategy: Mauborgne, Kim (2004)
- 20. Takeaways Building maturity involves many moving parts Technologists sprint faster, but business people run longer Interesting things (e.g., architecture) may not be useful (at first) You don’t know your gaps until you’ve analyzed your first process Be nice Process analysis is not a tribunal Your job is to show understanding, and a path forward Your sources are your target adopters. Don’t lose them. Simple heuristics make us smarter Adopt language of business Then make it simpler You can teach ERIC in 5 minutes. Try that with Six Sigma. 20
- 21. 21 Case: Redesign the Redesigners
- 22. Premise Global Fortune 500 Investment Bank Scrutiny of Back-office Operations: Want to be like an industrial enterprise Process is one pillar of the Industrialization effort Goal: Embed BPM in the Culture of the Organization Question: Tool is selected, what do we do now? 22
- 23. Progress and Reality 23 Core Team Chartered Tool Selected Governance Designed Design Group Chartered Pilot Projects Started BPM Training Roll-Out Design Group Starts Work Training As SPOC BPM Team Reorganized (again) Design Group Reorganized (again) First Processes Automated IT launches Enterprise Task Service BPM Training Part of Academy Design Group Reorganized BPM Team Reorganized 10,000 Models in Repository IT resists Process Automation
- 24. Car Insurance Claim Validation Document Completeness Fraud Likelihood Insurant Claim History Evaluation Claim Reliability Fault Calulation Insurance Coverage Decision Amount of Damage Exception Applicability Police Reports DMV Accident Report Former Claim Reliability Claims per Year Coverage Limit Claim Report Medical Report Premium Payments Insurant Negligence Car Condition Counterparty Negligence Road Condition Weather Conditions Fig. 18: Example of an insurance claim as a combination of layering and clustering (hybrid structured) decision model 24 Beware of Shiny New Objects
- 25. First Mover Dilemma 25 Core Team Chartered Tool Selected Governance Designed Design Group Chartered Pilot Projects Started BPM Training Roll-Out Design Group Starts Work Training As SPOC BPM Team Reorganized Design Group Reorganized First Processes Automated IT launches Enterprise Task List Decision Modeling Starts 8 mio Decisions/Day Automated Business Rule Project Starts Rules Platform Changed DMN Training Roll-Out IT launches Decision Service Decision Projects Launched BPM Training Part of Academy Design Group Reorganized BPM Team Reorganized 10,000 Models in Repository IT resists Process Automation
- 26. 26 Agility
- 27. 27 Beware of Shiny New Objects
- 28. 3 Process Analytics without a Process 28
- 29. College Admissions from a Process Perspective College Admissions is a highly manual process Think: Commercial Sales/Customer Relationship Management Prospect > Qualified Prospect > Applicant > Admit > Student > Alumni Highly dependent on external (often dirty) data sources Applicant Data (self-provided) Test Scores Letters of Reference Transcripts Every Applicant is a Process Instance Lots of exceptions, manual intervention We just graduated from the “Excel and PDF over Email” age Different Stakeholders have different objectives (Quality, Revenue, Reputation) 29
- 31. Starting with the Rear View Mirror How many applicants did apply last semester? Metric: Applications How many applicants did we admit? Metric: Selectivity How many new students enrolled? Metric: Yield Not as easy as it seems… …how to count applications to multiple programs? …how to count redirected applications? …how to count dual degrees? Agreement on measurements more important than technology 31
- 33. Beginning to Look Ahead Which applicants are currently in the system? Real-time Pipeline What is the current selectivity level? Real-time Admission Decisions Will we meet enrollment targets? Predicted Yield 33
- 34. Process Analytics without a Process 34 Process Analytics without a Process 7
- 35. 35
- 36. 36
- 38. What We Need BPM For How quickly are we responding to applications? Cycle Times / Queue Depths Which outbound messages affect yield? Process Personalization How far can we raise our admission standards this year? Process Optimization* * requires magic pixie dust 38
- 39. Understand the Process before the Process 39 Student enrolls Admit is nurtured Applicant is admitted Application is reviewed Application is completed Applicant is rejected
- 41. What the Challenge Comes Down To Want to understand the behavior of process participants But: They are often not aware they are part of a process Therefore: Create Transparency about the Process and their Role Note: This does not mean “make them BPMN experts” Want to influence the behavior of process participants But: Incentive systems are difficult to unearth Externally: Experiment, measure, experiment, measure (changing 1 parameter at a time) Internally: Show success. Explain connections. Be patient. 41
- 43. Takeaways Process Analytics starts with asking business questions - need agreement on metrics Historical (retrospective) analytics Reports, KPIs Use Process Mining if you don’t have a model (or even if you do) Real-time Analytics Dashboards, BI-type environments Need to understand decision-making: Synced with Reporting Cycle? Use Business Rules (IFTTT) to effect actions Predictive Analytics Holy Grail Need lots of historical data to build and test models Tools of choice: Simulation, Machine Learning, Optimization 43