2014/07/18 Creating High Velocity Orgs (Day 2)

by Gene Kim on

#devopsdays

Dr. Roberto Fernandez

  • MIT Sloan, Executive Education Faculty, William F. Pounds Professor in Management, Professor of Organization Studies
  • ex-Stanford

  • Ray Stata, founder Analog Devices: "The hard stuff is easy; it's the soft stuff that is hard."

  • Cleanup is the worst way to do things; get the soft stuff right

  • Experiment: 18 month old shown one of four posters

  • At MIT, logo shows people pointing outside: "and now you see my life at MIT"

  • Management is a contact sport

  • My goal: change your mental model of what it means to be effective

  • 18 month olds can be cued to be social; normally developing 6 week olds mimic adults

  • inattentional blindness: invisible gorilla: tunnel vision

  • our brain is a legacy system: side effect of visual cortex: blind spot and blurriness

  • Fernandez: "Our visual system's bandwidth is about 1 Mb/sec (slower than most cable modems; biological basis for inattentional blindness"

  • motivation

    • Steve's true north
    • design principles
      • systemic features
      • learning to learn
    • how do you get from here to there
  • 1st session: introduce some conceptual tools for thinking about the organizational change issuse

    • robust finding: most orgs struggle with firefinding
    • goals: lessons from the Repenning article
    • especially relevant for project based orgs
    • ask you to diagnosis
  • 2nd session

    • Hidden Traps in Decisionmaking"
    • consequences for active brain
    • implications for change
  • It's not about creating value; it's just as important to capture value: Xerox PARC

    • how do we create value, and then defend it
    • why do you need a strategy, because it guides choices: what to do and not to do
    • chronic overcommitment
    • Plan on 55 projects, do 18, 4 of which were not in the original plan
    • firefighting cycles
  • PreQuip has great control systems: they know exactly how over committed they are

    • Year 1: 289% overloaded, Year 2: 307% (got worse), Year 4: 226%
    • Project 26 is when they're at capacity; Project 26-30 is at overload
    • everything bogs down, people feel desperate that they can't get anything done
    • adult conversation: what is actually going on
    • Steve Spear talks about what should go on; I talk about what actually goes on
  • Why is it so hard to kill Project #26?

    • powerful sponsors: thinks his project is important
    • everyone knows that Project 26 will not getting done
      • zombie project: but everyone still gets a paycheck
    • elephant in room: Projects 1-25 aren't getting done either
    • no one has confidence that Project 1-25 are the most important things
    • the fear of the countermeasure is worse than status quo
    • Ideal, all work is prioritized in business priority: it should be worked FIFO
    • high level plan is good at PMO level: problem is when work gets decomposed
  • Discussion

    • political pressure
      • you should be afraid if you kill my project; you come after me with a knife, I come after you with a gun
      • you don't kill mine, I don't kill yours; we're one big happy family
    • overconfidence
    • invested
    • polarization
    • just say yes
      • we can stretch with consultants (and other stories we tell ourself)
      • when it comes to projects, we're a pro-life company
      • when numbers don't work, it has strategic value
      • more is good; there is so much good food there, ,we must eat it all; we must eat until we're so bloated, we can't get up
    • little commitments: this is what sales people do to you
    • Misc
      • we've already put effort into 26: it's too big too fail
    • you must follow thru on commitments
      • what do you call a project that you don't want to commit to? a pilot project
    • incentives
      • is killing your project a career enhancing move?
      • the shareholders forced me to pull a chair up to the buffet
  • Analysis

    • Blue: processes: orgs are just a set of processes
    • Red: competing demands: play politics
      • "I don't want to work in org that has politics"
      • we are human animals: politics come into play when we have scarce resources
    • Green: beliefs: orgs make assumptions, unquestioned beliefs
      • in mergers: that was different the day after, that wasn't true before: those are the beliefs
  • If you want to execute along Steve's vector, you need a multi-color palette

    • Blue: this is Steve's work: "if we had a better process, then all will be will"
    • There are also systems of competing interests, and there are collective assumptions
  • Projects

    • 7500 people: self-inflicted wounds
    • Nelson's article has 2 answers on curing firefighting
      • start fewer projects: put moratoreum
        • let rabbit work it's way out of the python
      • Injection of new resources (one time) for old business
        • avoid temptation to expand projects as capacity is added
        • Surge in Iraq: overwhelm the problem, right the ship, everything is ine
      • both have "worse before better" property
        • how appealing are these?
        • how easy is it to implement these?
      • first step in any diagnosis is to look in the mirror: "there is gambling going on in this casino"
    • if we take on so much work and can't get anything done, we don't have a strategy
      • it's like: "i'm a vegatarian, but wow, the shrimp and ribs are so yummy."
  • 3 types of orgs: 60 years of research

    • functional: tall, hierchical, allows for career growth
      • problem: they're tall, optimized for one market
    • organize around customer, faster
      • cross-functional: get close to customer
      • problem: lots of duplication
    • matrix: came from aerospace
      • airliner is a customer segment
      • you need areas of excellence: engines
      • that pain that you feel is result off not wanting to choose between functional or market
      • we have downloaded the ambivalence of not choosing
      • everyone has two bosses who think they can control all of your time; the result: we all work harder
      • IBM study: shape of curve: # of engineers on projects vs. productivity
  • Strategic design

    • organizations are machines
    • action comes thru planning
  • Political

    • organization are contests
    • action comes through power
    • paper: "power failures in management circuits": power is the ability to mobilize resources to make people do these things in order to get the job done
      • ideas, no matter how brilliatn, alone changes nothing -- xerox parc
    • premium on ability to get others to agree
    • good ideas are worth fighting for
    • power should be a complement, not a substitute for other talents (!! TODO: Erik quote)
    • Red with polite language
      • stakeholder, champion, sponsor,
    • who gets to make a decision: functional or market manager
    • video: john thompson, symantec ceo; paul ortellini, intel ceo
      • thompson story was about veritas merger: "they didn't get the strategy and blah blah blah. it wasn't about strategy, it was about buy-in. ship in or ship out"
  • Cultural

    • organizations are institutions: hard to see
    • action comes thru habits
    • as humans, we like to act without thinking, we are creatures of thought
    • we economize on thought where possible
    • we see culture when we observe
      • language
        • profit center (strength)
        • cost center (weakness, depressed): people are trying to cut them all the time
          • Apple: no one has profit or loss responsibility, except CFO: everyone charge to "increase value": increase value or reduce cost: no myth that one has more value than the other
      • cost vs. profit
      • front vs. end
      • customer vs. support
      • Mr. Profit always tells Mr. Cost that they're always working harder to make up for your sins
        • Cost Center people create value too
        • Accounting system is not the representation of reality (holy crap: !!)
    • "picture of cat" vs "cat" -- "it's not a cat, it's a piicture of a cat"
  • every class customized brass rat ring

    • we're all members of cult: "wear ring with beaver shitting on you until you until you graduate, so you can shit on everyone else"
    • no one taught it to them: but everyone knows that
    • "imagine me, showing up from california, and i see someone with brass rat: 15, 68; 17, 72; they nod, and walk away"
    • "what is your major and year graduate; {x,y}; very efficient ways of communication"
    • that's what siloes are: shared values
    • in science, articles get shorter and shorter, because we know what the terms are
    • except in interdisciplinary papers, where we have to define everything
      • in science, articles get shorter and shorter, because we know what the terms are; except in interdisciplinary papers, where we have to define everything
    • it feels efficient to work w/people we have shared cultures; cross unit coordination is hard, precisely because we don't share culture (WOW)
  • strategic design, political, culture comes up in all orgs

  • effective action requires a more complete read

Exercise

  • the curse of expertise
    • "autopilot" responses
    • in BP, mining disasters... overriding automatic alarms
    • normalize deviance.. no longer mindful
    • who are the more mindful?
      • new people: people who don't share assumptions
      • NSA: hires best programmers on planet: those problems who could solve problems were... high school students... different mental models
  • brain shortcuts
    • "our brain is powered only on 20W; which is why mental shortcuts & biases are needed; to economize on power consumption"
    • night scopes and helicopter pilots
      • professor of cognitive neural science: it was a disaster, b/c it was monocular
      • do you change the scopes? or just find one-eyed pilots?
      • most orgs look for one-eyed pilots, instead of design orgs around human capacities
  • 2nd framework
    • Thinking Fast And Slow: Dan Kanneman: Princeton cognitive professor,
    • fast: automatic processes
    • slow: learning, effortful, put into basement (can't multitask)
    • oldest part of brain: limbic system: fight/flight
    • there's also emotional processes
      • controlled vs automatic
      • id vs. ego
    • cuddy, fiske, glick: warmth is perceived first, competence is second
    • oxytocin increases trust in humans (created when nursing young)
      • trust game: investments are higher when oxytocin is adminsitered
      • cinnabon moved ovens from back to front
      • mcdonoald's: managers trained to keep parking lot clean: why? create perception that inside is welcome, clean and safe
        • that's where disgust lives; dirty and food do not go together
    • subtle role changes and horomones
      • ted talk: amy cuddy short video
      • 2 min high power poses: 20% inc or dec changes in cortisol, assertive, powerful and comfortable
    • engineering shrine in tokyo: 747 fuselage
      • mba student: tom sanderson: sloan graduate: that's why we go to pilgramages to narita airport from seattle; it's sacred ground
      • now moving an example to seattle
    • that emotional basement is powerful: storytelling
    • RAF: 1500 engines per month, rolls royce derbe works: hugh easton: not making aircraft engines, we're making the tools that saved England
    • "work until it hurts" to unionized workforce

Kahneman and Tversky

  • 2002 nobel prize: economists gave prize to psychologists to show that humans are not rational
    • we keep dragging irrational things into our decisions: like emotions (available online)
  • traps
    • anchoring trap
      • not baysian updating
    • status quo trap
    • sunk bias trap
    • confirmin evidence
    • framing
    • estimation
  • MBAs
    • smart, young, aggressive

small group exercise

  • anchoring efffect
    • setting targets: "in the next 5 years, do you think we can halve lead time?"
    • negotiation: important to put out number first
    • as change agents, deault anchor is status quo
  • endowment effect
    • everyone's children are always beautiful
    • can even overcome their own anchor: coffee mug experiment (people who receive coffee cups immediately increased assessed value, now that it's theirs; can take less than 1 hour to take efffecct)
    • countermeasure: walkout, give car for weekend
    • change agents should give people ownership
      • better yet, have them come up with idea themselves
      • if you design it, you're more likely to implement it
    • hardest to sell: complacent orgs who are fat and happy
  • sunk cost fallacy

    • especially difficult for pharma, extractive industries, with $1b+ of upfront cost
    • experiment: do we spend last $1B of R&D budget to build plane? (B side: we're 90% done, and competitor has cheaper, better version)
      • first case: 70% say yes
      • second case: 67% say yes (check picture)
    • careers are viewed as investments:
      • objection: "I've spent 30 years learning how to be a DBA. you're asking me to throw it away, and do completely different!"
        • countermeasure: " it's not totally different: it's a 15 degree change"
  • the further behind you are, the more risks you take -- the more corners you cut

    • is it about saving $4MM, or making $4MM
    • BP disaster: they were 6 months behind:
    • if you ask people to gamble: frame it around losses: "look what happens if we don't change" -- that's why the CRT is so useful (TODO);
      • >> you can double your odds of success by framing goals in terms of losses (loss vs. gain bias)
    • Nokia burning platform memo: "we're in deep trouble"
    • talking about the gains won't get you there
  • Social proof

    • when in doubt, you do what others do
    • mirror neurons: contageous sociality
  • ultimatum game

    • I get $7, you get $3; rational people will always say yes
    • in general, experimental break point is 7/3
    • in general, strategy is "how do we make them whole"
    • get the plan for this

Conclude

  • we need all three colors
  • expect resistencee to change, but it has to be guided
  • commitment map: where are people, and where do they need to be?
    • strategy: get high influencers and leaders first
    • 1 strategy: start with high influencers with low resistence; but then where? (use commitment map to figure out where soil is most receptive)
  • we all want to do things quickly -- but quick & dirty (w/angry amigdalas) can take longer than slow and deliberate

Dr. Steven Spear

  • most people are good, standouts are great; hugely disproportionate
  • story: reporter from Stars and Stripes about VA, what would you do?

    • what do you measure and why?
    • they were measuring "time to appts", and people were fudging data (sometimes subtly, sometimes illegally), and was using evaluatively
    • rewards or penalties for good metrics; even if they weren't inclined to do wrong thing, they still were effected
    • what is agenda of person running running VA? is that person really concerned about what is going wrong, or what is going right? It's the bad stuff that matters...
    • if it's evaluative vs. diagnostic: diagnostic allows people to tell bad news
    • then it's about the reaction piece: if you're head of huge agency, do you yell and ask "why am I getting bad news?'
    • or do you ask, "what can I help because you're having bad news"
    • congress thought: "we have to measure, for purpose of evaluation; that means reward good, and punish bad"
    • what if they took same measure, "what would happen in VA sites instead?"
    • what do leaders to to get high performance
    • must make it easier to see problems and solve problems
    • speed and range of action is hugely expanded
    • complex rules always less preferable to simple rules
  • other background

    • how can we lean in to the problem space?
    • system failure modes
    • low freq/high consequence vs. high freq/low consequence (alarm here)
  • Bayside company

    • derelict company
    • design and build the whole contraction
  • Bayside

    • plus: market share, new market, maybe fastest ramp (apply lessons learned)
    • minuses: should do housecleaning first, capital may limit ability to control acqusition, acquisition risk, you're acquiring another crummy company, replicating culture hard (family business)
  • Overseas expansion

    • plus: take existing systems and product they know well, demand, transport systems, known product, compete on cost and quality, higher margin
    • minus: supplier support, cultural and political landscape
  • New product development

    • plus: existing ramp skills, large market share (67%), strong culture and ability to execute, leader in niche, leader passion
    • minus: no standards, if done wrong, drag on everything else
  • Rockland

    • plus: work underway
    • minus: market unknown
  • do you trust pepsi to build airplane?

  • bayside innovations

    • themes
      • dynamic change (grounded in stability: howard is stability [ice], avi would test/innovate [fire])
        • stabilize and then innovate
      • empowerment
        • competition between cells, like running their own business
        • (what is empowerment?
          • when I delegate, I now work for them, removing obstacles, make her feel comfortable telling bad news, you have permission to do what is right;
          • when you see improvement opp, take it; freedom from approvals
          • Im leaving for a week; don't call me too often; just do what you think is right
          • other people make decisions on my behalf, b/c they become my multiplier
          • give someone responsibility and authority, resources, boundaries (so we don't get everyone in a whole mess of trouble, where it's safe to play); purpose is to get this multiplier
        • what are they empowered to do?
          • resequence production schedule
          • go to get cup or glass from upstream step
          • hire and train their owner workers
          • moved 18 cells: cell owners? no, Tyler
          • (the point: only 3 things?)
  • revisit strategies

    • acquire: no
    • overseas: no
    • new product:
    • rockland product:
  • what did they actually do?

    • they baysided someone else? spin vs. x/y stuff. leverage channels.
    • split time between bayside and 20 miles away on Long Island; just has hard as Frankfurt
    • everytime they leave, no one is there to run the business
    • move the business back to bayside into local space, and everyone quits
    • now you lost everything you bought (tech, relationships, customers)
    • two division: spinners vs. sliders
    • no one ever was trained to run their own business
    • succession planning is insurance
      • succession skill is a multiplier
    • rickover: involved thousands of people
      • Avi and Howard were great at dynamic change, but terrible at empowerment
      • A/H might do it to get insurance so he could take long vacation, get hit by bus
      • Rickover mobilized 1000s off people
    • to get multiplicative effect, you must learn how to teach other people
  • what was missing?

    • what was output of avi and howard?
      • you would have found dspite 18 changes
    • who did the work
      • avi and howard
      • all yellow post it notes were in front of avi and howard
    • conclusion:
      • all the work was done by 2 guys
      • hypothesis: if we asked "by whom", maybe that would separate successes from failures
    • "do you have a question, Captain Rickover?"
      • "No. I don't understand what the hell you're talking about. I don't even know enough to ask a question."
      • Had tutorial. Everyone showed up
      • Rickover probably knew more than anyone else
      • He needed to show what to do when you don't understand
      • He needed to amplify his knowledge and behaviors by 1000x in order to succeed
    • how do people learn? video of baby learning to walk
      • 12 month old; cruising brings more of the world within his reachh (vs crawling)
      • 3in for crawling; 20 in for crawling
      • how?
        • trial and error
        • encouragement
        • persistence
        • time was allowed
        • intrinsic reward
        • learning tools (couch)
        • help
        • examples (brain models internally before imitating)
        • fun
        • independent practice
        • novelty
        • physically safe environment (everything was padded: no stairs with jagged glass at bottom)
        • emotional safety (mom didn't yell "again?")
        • incrementalism: cruising is a precursor: 3 points of contact; pulling up, roll over, crawling
      • you never learn to ride a bike:
        • training wheels w/o seeing your feet
        • steering
        • more and more balance
        • enough velocity to stay up
        • backward to brake
    • big themes
      • incremental
      • trial and error
      • motivated
      • safe
    • who was the best adult learner I know?
      • who?
        • me: learning to skateboard; I'm a terrible athlete; scared to death
      • what did they learn?
        • learning to skateboard
      • where did it happen?
        • at home, got so jealous watching my 3 kids on scooters
      • when?
        • 3 months ago
      • how?
        • watched an 8 yo neighborhood kid
        • asked for help
        • each day, skateboarded with my kids
        • every day for 4 weeks
    • Jae Hwang examples
      • motivated, some qualiication/fitness, want to learn new styles
      • not too skilled (black belt who wants to integrate into their own style, doesn't fare too well)
      • chinese national got trained in physically distant (safe environment, safe from scrutiny)
    • discussion
      • even though I look foolish, I don't care; they create bubble around themselves
      • self-creating that safety
      • pain is part of learning; most of valuable lessons in life is probably the most painful
      • fear of public failure
    • Goal
      • come up with plan: who are you going to teach, what are you going to teach, balance frustration and reward, and do it in a way that's safe
      • who/where/how

conclusion

  • Neil Armstrong: 1961: We're going to commit that we're going to put man on the moon
    • Neil Armstrong land on the moon, 1969
    • giant leap or small step
    • actually a gazillion leaps
    • why didn't they launch in 1961: they didn't know how?
    • lots of steps
      • Mercury 1962, Gemini (2 astronaut, spacewalk, launch 2 rockets + roundezvois), Apollo (get to moon, orbit moon,
      • before Apollo was Explorer
  • even apollo program and rickover program took millions of small steps
    • can I come up with a bunch of solutions very, very quickly
  • One more picture and example
    • Toyoda Loom
    • problem was that people using loom, they don't know if the thread is broken; it would stop, so it could be rethreaded
    • it saved a huge productivity, got patent roytalties from british, and seed money for Toyota Motor
    • loom is in the lobby; in every wing, bathroom, you have to walk past the loom
    • sign: this loom was invented by Mr Toyoda himself
    • you pass it a lot into the description, rationale for circular loom; and it's the only of its kind. why? mr toyoda said this loom will run even if thread is broken
    • why did they put failure loom in lobby -- like edison bulb that didn't work
    • for 5 years, he wrestled with why. there was never a second loom -- he could never develop a loom that didn't stop with broken thread
  • one more story
    • chapter 9: friend of mine: bob dallas: transition from chrysler to toyota: meets with person of all toyota North America.
    • "how's it going?"
    • what we do is just like golf; they're both simple
    • everytime you play hall, hit with club, put in hole, repeating as necessary
    • it's difficult, but simple
  • who: A

  • where: weekly call

  • how:

    • how to handle inbound requests
    • review how handled each transaction
    • convince myself that it's not mission critical; I cant do the work anyway
    • first step: recontruct the steps
  • discussion

    • who: core team
    • where:
    • how: morning meetings; introduce topics as handoffs; do the post-it experiment, to see what we do, see what we do; use airplane experiment, to see how we can shuffle the work around (
      • make the team learning exercise (as opposed to whispering exercise)
      • show me how your team work, show me the post it notes
  • D2

    • share biases document, but now the framing traps, estimating and forecasting; implications to internal and external commitments; even sales proposal
      • losing money is far more powerful than making money
  • D3

    • IT project manager
    • haven't been provided with safe environment for trial and error
    • conflicting expert opinions lead to deadlock, then bring in vendors w/o making a decsion
    • give project managers a safe environment to do experiments
    • "any project we don't want to do, we call a pilot"

TODO

  • lookup kahneman nobel video

  • Love it! RT @jeffsussna: "There are no unicorns; there are trained horses. Netflix and Etsy took a journey" - @patrickdebois

  • RT @jeffsussna: "There are no unicorns; there are trained horses. Netflix and Etsy took a journey" - @patrickdebois

  • @lilydalekid: Guess who's getting a charge out of @Target at #devopsdays? #makeawesomehappen http://t.co/pvihG0pvQ4