ITSM and the Theory of Constraints Stuart Rance @StuartRance James Finister @jimbofin #bcssmsg
About Us Stuart Rance Optimal Service Management Consultant, Trainer, Author ITSM and Infosec James Finister ITSM Strategist Thought Leader ITSM, Audit, statistics 2
Questions we promised to answer. Why don t ITSM projects produce long term benefits? Why projects miss deadlines despite contingency time? How to get the Head of Ops and Dev to agree? How DevOps differs from Agile? The Theory of Constraints is a thinking tool that helps answer all these questions 3
But we ll begin with a really important question Have you ever asked yourself why might it be quicker to stand than walk on an escalator? And why that might not even be the right question? 4
Key Points Global not local optimisation Always a single constraint Elevate the constraint Drum Buffer Rope - Constraint dictates cadence - Buffers buy time - Rope pulls in work OK, so that works for widgets.but what about in ITSM? 5
ToC for ITSM We fix bottlenecks that aren t the constraint We fix process when we should think people We fixate on the wrong end of the queue We don t fix, we expedite? 6
ToC for Projects We design around tasks not the constraints We build contingency into every task - and then use it up at the start of the task (student syndrome) We should Look at critical resource contention Put the buffer at the end of the whole project Protect the critical resources 7
ToC for SIAM Critical to understand how complex value streams impact business level measures Suppliers build other suppliers risk into their contracts Constraints are out of sight Contract models need to equate to real world risk Remove risk from individual suppliers Allocate outcome risk equitably Grey box monitoring 8
Challenges Unpredictable external demand sets the drum Bottlenecks occur across multiple workstreams Difficulty of making to stock in service environments Faith in downstream suppliers More applicable in some areas than others? How do we make it work. 9
The First Way - FLOW Look at the end to end flow of your work Pull not push Global optimization, not local 10
The Second Way - FEEDBACK Need local AND global feedback Toyota Production System stop button Test often, not just at the end Understand flow (1 st way) to encourage feedback 11
The Third Way EXPERIMENT AND LEARN 12
Measurement and rewards Tell me how you measure me and I will tell you how I will behave Eli Goldratt 13
Questions needed for improvement What to change? What to change to? How to make the change happen? How to manage conflict? 14
Questions needed for improvement What to change? Current reality tree What to change to? Future reality tree How to make the change happen? Pre-requisite tree (and ambitious target tree) Transition tree How to manage conflict? Evaporating cloud 15
Ambitious target tree 16
Evaporating Cloud Common goal Needs Wants I need I want We both agree They need They want 17
Evaporating Cloud Common goal Needs Wants???? Customer to agree that meeting the SLA is sufficient High quality service at a reasonable cost???? Service provider to agree that meeting the SLA is not sufficient 18
Current reality tree 1 - Agree the scope of the system The service desk in London All services offered by IT to customers Management system of the entire organization How our team allocates work My personal skills and development 19
Current reality tree 2 - List five to ten undesirable effects 20
Current reality tree 2 - List five to ten undesirable effects Many incidents are poorly managed There s a high turnover of service desk staff There are lots of repeat incidents End users are unhappy with the service Service desk agents don t know how to manage incidents Incidents take a long time to resolve 21
Current reality tree 3 Analyze cause / effect relationships 22
Current reality tree 3 Analyze cause / effect relationships End users are unhappy with the service There are lots of repeat incidents Incidents take a long time to resolve 23
Current reality tree 4 Tie everything together 24
Current reality tree 4 Tie everything together End users are unhappy with the service Incidents take a long time to resolve There are lots of repeat incidents Many incidents are poorly managed We identify very few problems There s a high turnover of service desk staff Service desk agents don t know how to manage incidents 25
Things to Read 26
Thank you @StuartRance StuartR@OptimalServiceManagement.com @jimbofin James.finister@tcs.com