Dean Leffingwell Day 3: Part 1

by Gene Kim on

#@deanleffingwell

2017/09/14: Dean Leffingwell Day 3: Part 1

  • @deanleffingwell

  • @intelligentbu: RT @Rgrrrl: #DataOps seems like a great description of the convergence of DevOps and #datascience @ma_upson #EARLConf2017

  • @RealGeneKim: @kaimarkaru @randyshoup @Deanleffingwell Pervasive themes: generative culture, servant leadership, leader as develo… https://t.co/S1Fecs0LBT

  • @RealGeneKim: @kaimarkaru @randyshoup This is funny and pithy. But after spending a week w/ @Deanleffingwell, this certainly doe… https://t.co/TtADk3bReF

  • @ScottPrugh: @RealGeneKim @PrettyAgile @Deanleffingwell See Team of Teams.

  • @kaimarkaru: @RealGeneKim @randyshoup @Deanleffingwell Regarding training, principles taught are often not the same as principle… https://t.co/MctYoveQSm

  • @kaimarkaru: @RealGeneKim @randyshoup @Deanleffingwell The context I was referring to was the objectives as frequently expressed… https://t.co/vhH7yoBamF

  • OH: “software in intelligence agencies: SDLC in concrete, fighting a very dynamic enemy; I hope they can transform quickly”

  • “33% Nope; 33% I’m here to learn;

  • Prugh: trained 2000 people on SAFe for Leaders; and certified them: “it’s so important that we know that our people know”

  • on levels: strange concept; teams are everywhere, but concerns are different

  • our goal at portfolio level: move discussion from teams, projects -> value streams

  • if we don’t have a budgeting charge code for it, then we hobble teams (set of teams trying to support the warfighter)

  • Geoffrey Moore: Escape Velocity: vision/strategy; core and context

  • in times of innovation, 2 week sprints were too high; went to 1 week sprints; in most orgs, 2 weeks is balance between transaction vs holding cost

  • one team coming up on release said, “we’re not learning enough; sprints changed from 2 weeks to 4 days, so that we can enough learning cycles”

  • strange: in times of time crunch, very paradoxical: more time for planning/retrospectives

Lean Portfolio Management

  • “irony: PMO didn’t learn waterfall from a conference & imposed it on us; they learned it from us [dev community]”
  • “Lean Program Guidance” << not management
  • connect portfolio to enterprise strategy
  • John Deere: vehicle guidance, data collection, data center... lower level than construction equipment vs. farm equipment
  • strategic themes vs. enterprise articles
  • taxonomy is basis of science
    • user story: easy; technical spike -> enabler
  • strategic theme: 3.. not “don’t forget to make great combines”
    • here’s how this year is different than last year
    • Cigna: user personalization
    • appeal to younger demographic
    • reduce warehousing costs by 50%
    • customer intimacy / improve customer experience: (drive down call center)
    • long term investment vehicles: customer education: how to invest
  • SAFe: key strategic theme: configurability (b/c people were freaking out at size of framework), DevOps and continuous delivery, implementation roadmap
    • drove courseware and framework value streams

implement lean budgeting

  • problem: cost accounting, cost center budget, driving agility through functional org,
  • wrote joint white paper, and created derivative
  • getting funding: test group
    • project comes up: here’s my people, here’s the number;
    • architects, security
    • project: temporary assignment for temporary people
    • vs Agile: long term assignment for persistent people
    • utilitization based budget: 160 testers
    • “during budgeting cycle: if I can’t identify more projects, I’m going to lose my people”
    • it’s even hard for success: customer wants more of what we built! But we can’t do it until next year!
    • people are already assigned to another project; everyone upset
  • countermeasure: organize and fund value streams
  • problem: “funding agile teams”
  • 5 year business plans: projects for 2018: creating work statements
    • create adapters; plan of intent; 6 months with high fidelity
    • migration: long pole in tent
    • necessary for creating an agile net-Erie’s
    • can’t work on it because I don’t have a charge code
  • epic: enterprise initiatives
    • portfolio epic: cut across value streams
    • program epics implemented by ART
    • business epic: customer facing
    • enabler epic:
  • first: SAFe plus Lean Startup
  • not Gartner word soup
  • Lean UX, Lean Budgeting, Lean Startup
    • cable CIO: $35MM; What is hypothesis: do 5K boxes in one region: $5MM
    • old way: “iron triangle, waterfall sucks, you big dummy, I’m agile”
  • WSJF: MVP: very tight relationship

    • pivot
  • transforming to an agile company

The Portfolio

  • Kanban appeared in v3 to throttle work from portfolio
  • Kanban is a state machine
  • essential safe: do you have to have scrum? No. SPC? No. Agile teams? Yes.
  • too big framework: too big to understand; need entry point

PMO

  • 150 person ART: who can get a PO for $10K? Dunno. We want to PI planning at hotel? Dunno. We need more 7 more license for QC? Dunno.
  • PMO: 300 people: it’s because it was so hard to get anything done; only person who can span teams
  • Most important community practice: RTE: wouldn’t be great if they could get reporting across those?

Estimation

  • .@Deanleffingwell: “Toyota: there is always a constant sense of danger” (which helps create relentless pressure to improve)

  • 5-10 stories; 7 people; 2 week sprints

  • US, Australia, UK: $1500 per story

  • not the case for 13 person in Mumbai

  • it allows high level planning

  • law of large numbers

  • do you want teams estimating or coding? CODING!

  • JBOSS: 5000 points: $5MM

  • Don’t bother feature teams: Architects, PO, RTE: they usually know

Value streams

Implementation Roadmap (Invar)

Aha moments

  • An amazing video on adult learning: learning how to ride the bicycle required 8 months: https://ed.ted.com/featured/bf2mRAfC @deanleffingwell

  • a highly opinionated framework, based on sound principles

  • Jim Fowler: acquisition of Nokia led to working with Rami: now 3 years later, we’er Changing the funding model at Microsoft

    • we can report out on the expenditure, but no longer have conversation on budgets for effort; on teams; executive sponsor: GM in supply chain;
    • biggest detractor: business exec; now stood up in meeting when someone wanted to increase PI from 8 weeks to 12 weeks: said, “no way; you burn $15MM/month. what you’re proposing is that I’ll burn $27MM before I get a chance to get feedback”