Human error, brains and how agility helps
Download as PPTX, PDF
0 likes148 views
Human error is the leading cause of many technological disasters and project failures. The human mind has limitations like reduced working memory when multitasking or under stress. Organizational cultures and mindsets can also contribute to errors if they promote fear, lack of psychological safety, or compliance over generative thinking. Agile methods aim to prevent human error through small batches of work, frequent feedback, shared understanding, and a structure that supports the right cultural mindset for continuous learning and improvement.
Human error, brains and how agility helps
- 1. Human Error, Brains and How Agility/Scrum Helps ScrumDay Barcelona 2nd December 2021 Guy Maslen
- 2. Levi Strauss SAP rollout. Planned rollout for new ERP system Budget $5m Expenditure $192.5m Why Your IT Project May Be Riskier Than You Think by Bent Flyvbjerg and Alexander Budzier, HBR September 2011
- 3. Piper Alpha Platform built in 1976 In operation for 12 years Pump under repair Repair not finished at shift end Tag-out lock-out fails Another repair scheduled at the same time Pump assumed okay by new shift Pump started and fails Explosion Gas continues to be pumped 187 lives lost
- 4. Black Swan Events It is something that is assumed to be impossible It is something that will have a big impact It is something that happens (event) After the event, it is obvious Nissam Nicholas Taleb (2007)
- 5. Swiss Cheese Model We have processes to reduce risk Those processes can have holes The holes are due to human error Each stage is like a slice of Swiss cheese When the holes line up, a disaster happens Human Error James Reason - 1991
- 6. So what is human error? Unsafe Act Unintended Action Slip Attention Failures Lapse Memory Failures Intended Action Mistake Wrong rule Knowledge gap Violation Routine Violations Acts of Sabotage James Reason GEMS 1990 Skills Errors in execution Errors in planning Knowledge Culture and Mindset
- 7. Skills Domain Errors Operates over a few seconds Temporary storage Manipulates information Focusses attention We run out of RAM and: - forget to do things (lapse) - our attention fails (slips) Familiarity Increases our skills Familiarity makes us complacent - Context switching - Multi-tasking - Getting interrupted at a key point - Getting distracted by other activities - Being under stress or feeling threatened - Being hungry, angry, late or tired - Ultradian rhythm (40+ minutes) Conceptual model different parts of the brain in practice Factors decreasing working memory
- 8. Countering Skill Domain Errors Simplify Small slices Clear criteria Shared understanding Focus Less distractions Less interruptions Frequent breaks Structure Use a pattern Work in pairs or mobs Work to a cadence
- 9. Knowledge Domain Errors By Snowded - Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=33783436 Experts have hypotheses to test Experts have solutions to apply ANCHORING the first thing we hear, we hold on to CONFIRMATION we cherry pick data selectively to support our beliefs GROUP THINK we are reluctant to go against the group SUNK COST we have invested time/effort we are reluctant to abandon DUNNING-KRUGER misplaced confidence through inexperience
- 10. Brains, Knowledge and Belief Facts that contradict our beliefs drive a threat response(1) Being right increases our status in a group Being wrong decreases our status in a group(2) (1) Neural correlates of maintaining ones political beliefs in the face of counter evidence Jonas Kaplan 2016 (2) SCARF A brain-based model for collaborating with and influencing others David Rock 2009 Prefrontal Cortex Hippocampus Amygdala REWARD THREAT APPROACH AVOID We process information in parallel cognitive and pattern matching Pattern matching (threat/reward) is faster than cognition Threats reduce working memory loss of skill, emotional reaction
- 11. Countering Knowledge Domain Errors Discuss bias Discuss errors Discuss brains We might be wrong! Awareness Define hypotheses List assumptions Do research spikes Uncover what is right! Experiment Use groups Check in/out Avoid drama Get feedback Structure
- 12. Waterfall Failure Human Error Model ANALYSIS REQUIREMENTS DESIGN DEVELOP TEST DEPLOY Cognitive biases Assumptions Mistakes Slips and Lapses Mistakes, Slips and lapses Slips and Lapses Cognitive Biases Mistakes ???? Swiss Cheese Problem eventually the holes in each layer align Teams uncover problems time pressure and stress increases Complex documents combine with reduced working memory more errors
- 13. Agile Model : Bow-tie defense We Build The Wrong Thing We Build The Thing Wrong Requirements Errors Design Errors Analysis Errors Execution Errors Testing Errors Lean Canvas User Stories Walking Skeleton Pairing TDD, Automation1 (1) - Plus Bias Training Software Defect Prevation Using Human Error Theory Fuqun Huang, Bin Liu Cost Overruns Delayed Delivery Benefits Not Realised Incremental Delivery Iterative Delivery Highest Value First No longer a Swiss cheese chain of errors Barriers/Controls on likelihood of error and on possible consequences
- 15. What is culture? Paradigm Power Structures Organisation Structures Control Systems Stories Rituals and Routines Symbols Cultural Web Johnson and Scholes How we do things round here: - Formally and informally based - SCARF applies to change - Whats my Status? - What Certainty do I have? - What Autonomy do I have? - What are my Relationships? - Am I treated Fairly?
- 16. Fear Fear invites wrong figures. Bearers of bad news fare badly. To keep his job, anyone may present to his boss only good news W Edwards Deming Fear of censure or punishment. Fear of missing out on a reward. Fear of loss of status. Fear of loss of relationships. Fear of loss of autonomy. Fear of being treated unfairly. Fear drives violations we do not act as we should out of fear
- 17. Psychological Safety and Teams Teams that reported the most mistakes were also the best performing this led Google to Amy Edmondsons paper Psychological Safety and Learning Behaviors in work teams Not the same as TRUST If I speak up, it will not impact negatively on my STATUS or RELATIONSHIPS - Willingness to speak up about slips, lapses, mistakes and violations - Open to feedback on things that impact on our status, ego and identity Its NOT about safe spaces; it is about behaving in a safe way The Scrum Guide suggests courage; this is how we apply it
- 18. The Safety Culture Ladder No belief or trust; environment of punishing, blaming and controlling the workforce Failures caused by individuals; no blame but responsibility. Workforce needs to be educated to follow processes. Command and control environment; lots of metrics and graphs flow upward, but do not represent what is happening. Management walkabouts; workforce involvement is promoted/ruled by specialist staff who are obsessed with statistics Management in respectful partnership with workforce. Management has to fix systemic failures, workforce has to identify them After Hudson, 2001 Safety Culture Theory and Practice
- 19. Gilbert Enoka All Blacks Mental skills coach We use STRUCTURE to recover our MINDSET to apply our SKILLSET Cabin crews, arm doors and cross-check Scrum (or XP, or Kanban) provides the STRUCTURE that helps to preserve the agile mindset - are we being generative? - are we being psychologically safe?
- 20. Countering Mindset Domain Errors Ownership Psychological Safety Behaviors Generative Servant Leader Culture Ritual and Routines Stories and symbols Structure
- 21. Wrap Up As coaches, we can teach our teams about human error, but thats not enough As teams, we can build processes to trap errors, but thats not enough As leaders, we can model behaviors, but thats not enough The safety world has walked this path already We can only preserve the right mindset with structure; without that, our culture and performance will erode, because we are human.
- 22. Selected References Papers and Articles: Why Your IT Project May Be Riskier Than You Think - Bent Flyvbjerg and Alexander Budzier, HBR September 201 Rest breaks and accident risk - P Tucker, S Folkard, I Macdonald, The Lancet, March 2003 The impact of rest breaks upon accident risk, fatigue and performance: a review- P Tucker, 2003 Impact of Task Switching and Work Interruptions on Software Development Processes A. Tregubov, N. Rodchenko, B. Boehm and JA. Lane, 2017 No Task Left Behind? Examining the Nature of Fragmented Work - Gloria Mark, Victor M. Gonzalez, Justin Harris, 2005 SCARF : a brain-based model for collaborating with and influencing others David Rock, 2009 Neural correlates of maintaining ones political beliefs in the face of counter evidence Jonas Kaplan, Nature 2016 Psychological Safety and Learning Behavior in Work Teams Amy Edmondson, Administrative Science Quarterly 1999 A Typology of Organisational Cultures Ron Westrum, Quality and Safety in Health Care January 2005 Safety Culture Theory and Practice- Patrick Hudson, January 2001 Ethical Dissonance, Justifications and Moral Behavior- Rachel Barkin, Current Opinion in Psychology, August 2015 Books: Human Error James Reason, 1991 Out of the Crisis W Edwards Deming, 1985 Exploring Corporate Strategy - Johnson and Scholes, 1990+ The Black Swan: The Impact of the Highly Improbable - Nassim Nicholas Taleb, 2007 ,
Editor's Notes
- #2: Introduction : Guy Maslen Journey into how and why things go wrong, why thats largely down to neuroscience and our brains, and what we can do about it. Using the journey that health-and-safety has been on over the last 20 years which is all about risk and applying this to agility which is also all about risk. Empiricism and Scrum; outcome of helping mere humans to do complex work or Scrum through the lens of human error. Linking the strands of my career offshore geophysics, software, geohazard monitoring
- #3: - 11 years ago asked by CEO to look at IT; deep rabbit hole into catastrophic project failures. - how can these things happen? - how do you set out to spend $5m and overrun by $192.5m? - other big failures Deutsche Post-DHL wrote off 345m of costs and abandon the ill-fated New Forwarding Environment - example from a paper on IT and black swans a long tail of disastrous overruns ; advice was to slice projects small, and lots of medium sized risky projects made for a lot of risk - still curious about why>
- #4: A different kind of disaster; NS Oil place in my life, did my BOSIET training just after recomednations brought in. Unlike IT projects we have a public enquiry, and theres a chain of events. What has this got to do with IT? Its all about human error. Why things go wrong, just that the HSE world has been their first, and we can learn from it.
- #5: These are called black swan events obvious after the fact there will be a big issue, but before hand no-one seems to see it coming, or the advice and leading indicators are ignored. Named for the Australian blackswans all European swans were white, referenced by HBR paper
- #7: Routine violations at piper alpha breakdown of safety culture (tag out lock out)
- #8: Example driving car to the office and otherhtings on my mind, turn to the gym instead Wing mirror checking when driving Walk through doorway and forget why we came in (working memory wiped) Stress lowers working memory glucose and oxygen go to pattern matching part of brain Errors 2x more likely in the last half hour of a 2 hour period (Tucker, Folkard, MacDonald) Flow state can take 25 minutes to recover (Gloria Mark)
- #9: Pomodoro technique (5 mins every half hour), 20 minutes every 90 minute sessions.
- #10: We all suffer from cognitive biases; tricks our brain plays on us, that reduce how objectively we think. Many biases, just a few key ones here. Dave Snowdens Cynven presents the same challenge how do we know if we are in the complex or complicated domain? Are our biases at play? Deepwater Horizon and the bladder effect talk themselves into a disaster, via anchoring
- #11: Some of this again is about brain function. We parallel process incoming information, through the pattern-matching centres (amygdala/hippocampus) and the cognitive centres. The cognitive centre is faster so the emotional response (threat or reward) is triggered first, reducing working memory, limiting our openness and creativity. We are strong but wrong example Reason gives is Chernobyl. Two under pinning phenomena related to our status.
- #12: Raise team awareness of thee types of error; understand we have a hypothesis to test, discuss leading indicators Work in groups aware of biases, and use structure to recognize when we have an emotional attachment to something.
- #13: So applying these ideas to a Waterfall development model, what can we see? Theres scope for human error at every handoff.
- #14: Bow tie is a model from aviation industry around risk. Look at contributing evenets or triggers that form a hazard, then the consequneces of that. Then explore barriers we cam put into place. So how does Scrum help us?
- #16: Were into culture and mindset driven errors. Diliberate violations we know what we *should* do, and we do a different thing. Ethical Dissonance (Rachel Barkin 2015) we do the wrong thing and justify it to ourselves or others. So what drives these deliberate violations of process?
- #20: GNS Science conference 300 scientsts and engineers, 10 minutes all taking notes. What coaching really is, and how effective it can be, Asked question
- #22: Car analogy seatbet