Systems Thinking workshop @ Lean UX NYC 2014

  • Published on
    17-Aug-2014

  • View
    1.281

  • Download
    2

DESCRIPTION

Slides with notes for my workshop at Lean UX 2014. This is an iterated version of my 2013 workshop - different exercise, slightly different content, but much is similar. Includes link to handout!

Transcript

Making sense of messy problems Johanna Kollmann @johannakoll ! Lean UX NYC 2014 Systems thinking for complex business models Illustration by David Wicks: http://www.ickr.com/photos/sansumbrella/467998944/ Intro about me: worked on a range of complex systems such as a voice communica;on system for the NASA, before learning more about systems thinking as part of my HCI degree in London. Interest in systems theory and organisa;onal structures remained when I was consul;ng, e.g. a large retailer who was reshaping their en;re business and data structure to enable mul;-channel. While geFng interested in business models and the startup world, I realised that systems thinking is also core to business models, lean manufacturing, and lean startup. The next 3 hours of your life: Introduction to Systems Thinking Tools for modeling systems (collaborate!) Systems behavior over time Change ASK: Whats your current understanding of systems thinking - share with neighbour What are your expectations for today Ask a few people to share Monitor changes in the system Understand peoples worldviews To reduce uncertainty NUTSHELL!!!! Systems Thinking? Why you should care about it !Increasing complexi;es and dependencies require us to think holis;cally. We need to think dynamic and over ;me rather than sta;c and short-lived Technology and business context changes. !ST is relevant to both UX and LS. http://visitmix.com/work/descry/awebsitenameddesire/ The systems we deal with in the world of a website Running a business is taking this to a dierent level - being a founder is taking the running around and coordina;ng to a dierent level! In the past the man has been rst; in the future the system must be rst. ! ~ Frederick Winslow Taylor (1911) father of scien;c management and eciency movement In the past the man has been rst; in the future the system must be rst. ! This in no sense, however, implies that great men are not needed. ! ~Frederick Winslow Taylor (1911) According to Eric Ries, forgeFng the human part has led to 2 problems: 1) overly rigid business systems that failed to take advantage of adaptability, crea;vity, and wisdom of individual workers 2) overemphasis on planning, preven;on and procedure, which enable organisa;ons to achieve consistent results in a stable world. At the root of every seemingly technical problem is a human problem. ~ Taiichi Ohno A system is ~ Donella Meadows a set of elements or parts o[en classied as its func;on or purpose. that is coherently organized and inter-connected in a pa]ern or structure that produces a characteris;c set of behaviors, Peter Checkland Human activity systems Soft Systems Methodology Examples: hard system = thermostat, motherboard. so[ system = game of poker, soccer game, mee;ng, healthcare. Human activity systems, on the other hand are essentially complex, indefinable and purposeful. !He developed the so[ systems methodology, sugges;ng that most problems in systems are caused because human beings are hard to predict. He did not think that there were things you could x with systems thinking, instead there were situa;ons you could improve. !4 ac;vi;es of SSM: - Finding out about the situa;on - Making purposeful ac;vity models based on par;cular world views. - Using the models to ques;on the situa;on - Dening ac;on to improve the situa;on. Peter Checkland Soft Systems Methodology Activities: Finding out about the problem situation Making purposeful activity models Using the models to question the situation Dening action to improve the situation Examples: hard system = thermostat, motherboard. so[ system = game of poker, soccer game, mee;ng, healthcare. Human activity systems, on the other hand are essentially complex, indefinable and purposeful. !He developed the so[ systems methodology, sugges;ng that most problems in systems are caused because human beings are hard to predict. He did not think that there were things you could x with systems thinking, instead there were situa;ons you could improve. !4 ac;vi;es of SSM: - Finding out about the situa;on - Making purposeful ac;vity models based on par;cular world views. - Using the models to ques;on the situa;on - Dening ac;on to improve the situa;on. ! Leverage points places within a complex system where a small shift in one thing can produce big changes in everything. are often counterintuitive. Systems Thinking & UX 1) Modeling 2) Behavior over time 3) Change 1) Modeling Models are tools for understanding complex situations. Models are tools for communicating complex situations. ! Only by building a model of customer behaviour and then showing our ability to use our product or service to change it over ;me can we establish real facts about the validity of our vision. ~ Eric Ries Personas from Design Jam London, by Je Van Campen http://www.ickr.com/photos/otrops/tags/designjamlondon/ This is where UX oers lots of tools: personas, customer journey maps; Lean Startups hypothesis-driven approach also is modeling. Flickr User Model by Bryce Glass http://www.ickr.com/photos/bryce/58299511/ Models help us understand how things work. 1) Modeling 2) Behavior over time 3) Change Rich Picture 1. Construction of the Humber Bridge (adapted from Stewart and Fortune, 1994) The Open University 2. Distance Learning Situation Wood-Harper et al, Information Systems Denition: The Multiview Approach, Blackwell Scientic Publications 1985 1) Modeling 2) Behavior over time 3) Change Rich Picture elements Stakeholders Worldview Connections Conicts 2. Distance Learning Situation Wood-Harper et al, Information Systems Denition: The Multiview Approach, Blackwell Scientic Publications 1985 Worldview is a concept for empathy !Consider: - roles that people adopt in the situa;on (which may be formally recognised or quite informal); the norms which govern peoples behaviour; and the values they espouse. - poli;cal aspects of the situa;on, in other words recogni;on of the dierent interests that are represented and how these dierent interests are accommodated. 1) Modeling 2) Behavior over time 3) Change 1) Modeling 2) Behavior over time 3) Change 1) Modeling 2) Behavior over time 3) Change ! 1) Modeling 2) Behavior over time 3) Change Handout: http://bit.ly/OQzwza 1) Modeling 2) Behavior over time 3) Change Rich Picture elements Stakeholders Worldview Connections Conicts 2. Distance Learning Situation Wood-Harper et al, Information Systems Denition: The Multiview Approach, Blackwell Scientic Publications 1985 1) Modeling 2) Behavior over time 3) Change Rich Picture applications Framing the problem: Checklands root denition Understanding and communicating a complex situation Uncovering assumptions and knowledge gaps Research planning Stakeholder risk matrix CATWOE 1. A system Transformation ie a clear relationship between system inputs and outputs. 2. A system Owner ie someone who is ultimately responsible for the system. This person, or persons, can often be identified by asking the question who can stop the activity? 3. Actors those people who take action within the system. 4. Customers for the system ie the beneficiaries, or intended beneficiaries, of the system. 5. The system Environment within which the activity takes place. 6. The World view which enables all of the above to make sense. 1) Modeling 2) Behavior over time 3) Change Business Model Canvas Job seekers Recruiters Jobs Candidates Manage, promote platform Platform Manage and develop platform Marketing costs Job ads Hiring fee 2) Behavior over time 1) Modeling 2) Behavior over time 3) Change Flows inow outow information feedback, control stock Bath tub example - overow pipe !2 types of ows. First one is material and stock ows. Stocks change over ;me through the ac;ons of ow. Stocks act as buers or delays, and help a system to stay in balance. You can also apply this to people. Shows limits to growth if your resources arent endless. Key is to understand and monitor system behaviour over ;me. Do not focus on only individual events. !The second type are informa;on ows. While its hard to change physical structure, materials, resources, changing how informa;on is distributed and presented in a system can have major impact. "Informa)on holds systems together and plays a great role in determining how they operate. Most of what goes wrong in systems goes wrong because of biased, late, or missing informa)on." (Meadows) Adding or restoring informa;on can be a powerful interven;on, usually much easier and cheaper than rebuilding physical infrastructure. !Notes on John Seddon: interes;ng to consider how customer inquiries/feedback come in and ow through the system 1) Modeling 2) Behavior over time 3) Change Feedback loops Georges ability to solve problems Number of problems solved Number of remaining problems Time available per problem Project in trouble Management pressure to solve problems R1 R3 R2 Need to involve Paul B1 Reinforcing feedback loops A posi;ve feedback loop is self-reinforcing. The more it works, the more it gains power to work some more. Posi;ve feedback loops drive growth, explosion, erosion, and collapse in systems. A system with an unchecked posi;ve loop ul;mately will destroy itself. Usually nega;ve feedback loop kicks in, eg epidemic runs out of infectable peopleor people take increasingly strong steps to avoid being infected. Reducing the gain around a posi;ve loopslowing the growthis usually a more powerful leverage point in systems than strengthening nega;ve loops, and much preferable to leFng the posi;ve loop run. (...) control must involve slowing down the posi;ve feedbacks. !Balancing feedback loop A nega;ve feedback loop needs a goal and a response mechanism. Self-correct the system, o[en inac;ve = emergency mechanisms. Seem costly as inac;ve, removing them has li]le impact in the short-term, neglect the long-term impact. Here are some other examples of strengthening nega;ve feedback controls to improve a system's self-correc;ng abili;es: preven;ve medicine, exercise, and good nutri;on to bolster the body's ability to ght disease, pollu;on taxes. !The informa)on delivered by a feedback loop - even nonphysical feedback - can only aect future behaviour; it can't deliver a signal fast enough to correct behaviour that drove the current feedback. There will always be delays in responding. The loop that dominates the system will determine the behaviour. Consider the driving factors, how they might behave, and what drives them. ! Dynamic systems studies are not designed to predict what will happen, but to explore what would happen if... --> system dynamics models explore possible futures and ask 'what if' ques;ons. !Causal Loop Diagrams help reveal system dynamics. Crea;ng the diagrams involves more work than reading them, but can be done by anyone willing to take ;me to think things through and look for rela;onships. For example, what problems might arise by involving help? Is it possible that things will get worse before they get be]er? And why would that be? !Rela)ng loops to Eric Ries engines of growth word of mouth, side eect of use, paid adver;sing, repeat use S;cky - make me come back Viral - word of mouth Paid 1) Modeling 2) Behavior over time 3) Change Behavior over time graphs inventory days perfect informa;on scenario ! 1) Modeling 2) Behavior over time 3) Change Behavior over time graphs inventory days what really happens !What came before? What might happen next? !Focus on trends over ;me rather than single events. Learn if the system is approaching a goal or limit. Inventory = stock (could also be informa;on) 1) Modeling 2) Behavior over time 3) Change Cohort analysis 1) Modeling 2) Behavior over time 3) Change Cohort analysis Eric writes: Cohort analysis: This technique is useful in many types of business, because every company depends for its survival on sequences of customer behaviour called ows. Customer ows govern the interac;on of customers with a company's products. They allow us to understand a business quan;ta;vely and have much more predic;ve power than do tradi;onal gross metrics. p 145 Cohort-based reports are the gold standard of learning metrics: they turn complex ac;ons into people-based reports. 1) Modeling 2) Behavior over time 3) Change Custom tools to monitor interactions by @lukew 1) Modeling 2) Behavior over time 3) Change Photo by Anders Zakrisson http://www.ickr.com/photos/anders-zakrisson/4982281184/ Talking to people, empathy, intui;on DATA MEANING humanise the data tell a story !Informa;on ows enable other things in the system to happen Consider the feedback loops Observe customer behavior over ;me Use qualita;ve ndings and your gut 3) Change 1) Modeling 2) Behavior over time 3) Change inventory days Flows and loops Donella Meadows also says that its quite tricky to properly monitor a system and react appropriately, because the delays in observing, and then the delay in ac;ng means that by the ;me your change goes into place, the system is probably in a dierent state. Its easy to over compensate. It seems to me that you need to try to get both stats as real-;me as possible, and gain a good understanding of natural ows over ;me. shi[ a]en;on from the abundant factors to the next poten;al limi;ng factor. layer of limits. !If a decision point in a system (which can be a person) is responding to delayed informa;on, or responding with a delay, the decision will be o target. Ac;on taken too fast can cause unnecessary instability. !When there are long delays in feedback loops, some sort of foresight is essen;al. To act only when a problem becomes obvious is to miss an important opportunity to solve the problem. !genchi gembutsu from Lean: understands that a small change can aect the overall system. the person close to the problem is trusted with solving it. You have to 'go and see for yourself'. dont change your strategy on a whim! 1) Modeling 2) Behavior over time 3) Change systems with dierent users: consider how role changes will impact everything. Some of this is quite hard to implement! Understand the system structure youre building! Work with developers who draw diagrams about the so[ware system, so you also understand technical legacies and ripple eects. 9. Numbers (subsidies, taxes, standards). 8. Material stocks and ows. 7. Regulating negative feedback loops. 6. Driving positive feedback loops. 5. Information ows. 4. The rules of the system (incentives, punishment, constraints). 3. The power of self-organization. 2. The goals of the system. 1. The mindset or paradigm out of which the goals, rules, feedback structure arise. 1) Modeling 2) Behavior over time 3) Change Leverage points Mention how we often find ourselves as consultants in a situation where we are working on one level - eg improving information flows - but effectiveness of the solution we are implementing is constrained by leverage points of a higher level (eg rules). 1) Modeling 2) Behavior over time 3) Change Disruptive startups change existing systems Behaviour at scale Emergence of culture Environment readiness Why certain businesses emerge from certain locations, contexts Take-aways The worldviews that people and elements in the system hold The processes that are necessary to deliver value to customers ! How to gather and visualize information holistically How user-centered design and empathy help to reduce uncertainty ! What is the right level for the impact you are aiming for? What enables the change, where are conicts, who can be your change agent? This matters because Business trends. Humane systems. The intuitive mind is a sacred gift and the rational mind is a faithful servant. We have created a society that honors the servant and has forgotten the gift. ! We will not solve the problems of the world from the same level of thinking we were at when we created them. More than anything else, this new century demands new thinking: ! We must change our materially based analyses of the world around us to include broader, more multidimensional perspectives. ! ~Albert Einstein Resources The Lean Startup by Eric Ries ! Systems Thinking, Systems Practice and Soft Systems Methodology by Peter Checkland ! Thinking in Systems: A Primer by Donella Meadows ! Business Model Generation by Alexander Osterwalder and Yves Pigneur ! Donella Meadows article Places to Intervene in a System can be found at http:// www.developerdotstar.com/mag/articles/places_intervene_system.html ! Peter Senge is a key systems thinker, I havent included any of his material directly, but read about this perspectives especially on organisational change. Check him out. ! For the design geek in you, read up on Buckminster Fullers Design Science. Handout: http://bit.ly/OQzwza