2017/09/07: Utah Ops & Goldratt Consulting: Breakthrough Reults For Government and Business Part 3b

by Gene Kim on


2017/09/07: Utah Ops & Goldratt Consulting: Breakthrough Reults For Government and Business Part 3

Up: Alex Knight: Breathrough Results in Quality, Safety, Timeliness and Affordabilty of Healthcare

  • Knight: “US tops spending in healthcare per-capita GDP; some US states have 30% of economy on healthcare spending
  • Knight: “Healthcare has some of the most complex value streams around, with catastrophic outcomes when things go wrong
  • Knight: “Goals of healthcare: provide high quality care for all patients vs. be financially stable; more vs less front line staff
  • Knight: “Cutting healthcare costs often leads to reducing staff who are actually deliver healthcare to patients
  • Knight: “The big problem: as much as healthcare medical costs are rising, revenue is rising slower: 3% cost vs. 1% revenue growht
  • Knight: “Prices of everything going up (radiology, etc), and # of elderly going up, and # of life-lengthening treatments going up
  • “I work at Texas Health Resources; read Pride & Joy; started in Emergency Dept; reduced length of stay from 4.9d > 3.9d in 2 months
  • “I run rural hospital in Quebec; in 2015, reduced length of stay almost overnight, incr physical engagement, discharge planning
  • Knight: “In healthcare, we don’t talk about lead time, we call it ‘length of stay’; has massive impact on quality, patient outcomes, sense/purpose
  • Knight: “Healthcare: complex routings, massive use of shared resources (doctors), lots of stastistical fluctuation (patient arrival, job completion
  • Knight: “whenever I see 4-5 or more improvement projects, I know that they’re probably making problems worse
  • Knight: “One reason: everyone has their own reason to pick the next problem or patient to see; impossible to control outcomes
  • Knight: “Value stream: patient admission date; planned discharge date based solely on clinical needs; ultimate of patient-centered clinical care
  • Knight: “...then each day, look at which tasks are causing are the most disruption/delay to all patient objectives
  • Knight: “...then we can create a manufacturing-like master schedule; in healthcare, we create list of top tasks for healthcare providers
  • Knight: “It’s patient-centered, not a top-down government target mandate;
  • Knight: “in decades, never found a bottleneck; it’s always mis-synchronization, handoffs, sequencing
  • Knight: “We’ve acknowledged the impossibility of trying to measure capacity of complex healthcare orgs; IMHO, biggest waste of time ever
  • Knight: “...instead, we look only at where disruption occurs; balancing flow vs capacity;
  • Knight: “Simplifying question: do I have capacity when I need it? (Vs trying to schedule/control entire system)
  • Knight: “Question: what are the financial implications of healthcare improvements; what do we do when healthcare orgs make more profit?
  • variable expenses not variable (docs don’t renegotiate contracts)
  • Knight: “Improve protective capacity, increase throughout; goal: treat more patients, improve healthcare outcomes
  • Knight: “My experience: healthcare doesn’t have to be a drain on economy: it can be a [massive value creator]
  • Knight: “Our work is expanding to beyond hospital level to national level [in UK] to improve health; enough assets are there
  • Knight: “In Peru hospital, 400 patients in ER; 60% shouldn’t have been there; how can we create better system that directs flow
  • Knight: “Because patient flow working wrong, they all end up in the ER. These need to be addressed at larger level
  • Knight: “How can nations turn their healthcare systems that creates GDP, instead of depletes GDP;
  • Knight: “Look at healthcare tourism: higher quality, cheaper, safer, timeliness; shows the potential that healthcare can be national treasures
  • Knight: “Q: how much excess capacity do you think US hospitals have? A: My guess: acute care: 20%; mental health: 50% (more complex flows
  • Knight: “In many cases, we’ve shown that we can deliver better case, shorter stays w/120 beds that used to take 600 beds; goal: increase GDP
  • Tomorrow, Utah governor will kick off conference here. Cool!

Up: Yoav Ziv, Head of Amdocs Testing Services: TOC and DevOps

  • Ziv: Amdocs: massive IT org, b/c of merger/acquisition; apps 40+ yrs old to developed yesterday; lots of suppliers, integrators
  • Ziv: showing the true complexity of IT orgs, which may not be apparent to people in other professions
  • Ziv: “as head of QA/testing, ppl say DevOps doesn’t do testing, I worry about 3K family members who depend on doing QA work
  • Ziv: “my thesis is that QA/testing is the most important component for DevOps outcomes” (yes, for sure!)
  • Ziv: “Amdocs: market leader in customer experience solutions: $3.7B revenue, 25K employees, 300+ customers in 90 countries; 2500+ projects
  • Ziv: “Amdocs only has large customers; in US, Sprint is our smallest customer”
  • Ziv: “Our customers: not telco vs telco: it’s telco vertical being massively disrupted by Netflix; they and we were caught off-guard
  • Ziv: “2006: evaled Critical Chain Project Management vs Agile; CCPM too rigid for IT, no support for scope chg; no empowerment
  • Ziv: “CCPM: didn’t recognize how fluid software projects truly are;
  • Ziv: “OTOH, we found Agile to be too immature; didn’t address issues bigger than a team; didn’t address things outside of Dev
  • Ziv: “Given our choices, we chose CCPM vs. Agile; likely largest CCPM implementation in the world; saved 12% costs, but not sustained
  • Ziv: “We found that anyone trying to impose rules on devs never worked; ex: COBOL was perf language for biz; as # of devs incr, opt for themselves
  • Ziv: “We transitioned to Agile, but code never made it to production; biz never saw value of increased dev productivity
  • Ziv: “Although tech people disdain manufacturing, our problems were tackled in manufacturing a century ago” (Ha! So true!)
  • Ziv: “In manufacturing, WIP is easy to see;
  • Ziv: “Complex mfg value streams: ovens needed to cure paint so it doesn’t fall off, rust; ovens need to stay on/hot; 1000 doors at a time
  • Ziv: “Large batch sizes for heat-treat oven creates problems: waiting for large batch, then 1000 doors flooding downstream
  • Ziv: “We saw same probs w/large batches in IT, because of huge setup times for testing and deployment”
  • Ziv: “Our goal in DevOps is just like manufacturing: get us as close to single-piece flow, just like w/heat treat oven
  • (It’s so fun for me to hear people derive DevOps patterns from the principles in Phoenix Project. Ziv @ Amdocs talk is so great!)
  • Ziv: “For us, our enterprise application of DevOps at Amdocs is the creation of synchronized single-piece flow”
  • Ziv: “In our interpretation of Agile, it was perfectly acceptable to create untestable software; made impossible to test continuously
  • Ziv: “We created MTV: Minimum Testable Value; created test architects, shaping all dev work from beginning: Junit, TDD, etc
  • Ziv: “Massive synchronization effort across systems, generations of tech, vendors, integrators
  • Ziv: “Our test architects generate defects against design docs, invited to the first planning/design meetings; designing quality in
  • Ziv: “Next: automation, virtualization, microservices; architectural transformation
  • Ziv: I’m loving his application of lean value streams, TOC, improving flow; very nice non-parochial view of TOC!
  • Ziv: “My big epiphany: our biz stakeholders care about fast delivery of value, which req reorganizing work around QA/testing
  • Ziv: “Almost our entire org is technology oriented; testing requires huge amt of domain knowledge: telecom, finance, healthcare, retail
  • Ziv: “Digital transformation hit telco: WhatsApp, Netflix; now our domain expertise can help all our telco customers
  • Ziv: “The output from test architect actually generates all the work artifacts for Dev in Rally

Up: Charles Toups, VP Boeing Research and Development

  • So fun! Sitting next to Kurt Kinder from Alaska Airlines, whose presentation I loved!
  • Toups: “Boeing focuses on speed: then savings, market share, innovation will follow”
  • Toups: “Our goal is to innovate to create the next game-changing product: planes, satellites, defense, ... profit enables more innovation
  • Toups: “2004: recovering from dot-com crash: satellite orders went from 30 to 3; switched to CCPM to improve flow, profitability
  • Toups: “got promoted to head up defense arm from satellites; successor pulled plug on CCPM; my focus was on simpler version of TOC/CCPM
  • Toups: “One benefit of simpler version: didn’t need complex software systems, could do on paper; saved millions of $$ of SW costs
  • Toups: “2013: Promoted again to deliver better product development for Boeing: Operationalize Productivity & Initiative for lg programs
  • Toups: “787: 150 new nonstop markets; 95% common parts; 60% impr flow; leading tech driving cash flow and connecting cities
  • Toups: “Success factors: leadership champions; aggressive targets; standard work/remove complexity, focus on WIP
  • Toups: “Req of an aggressive target: team doesn’t initially believe it’s possible, transition to figure out how it can be possible
  • Toups: “Paradox of standardized work: we want things simple & widely used, but fully owned by teams” (standardization vs innovation)
  • Toups: “Key to throughput is reducing WIP: major program has 1000 engrs, each w/more work on their plate than they should
  • Toups: “Major 787 model comes out next year, doing great job returning cash to business” (love how biz leaders talk :)
  • Toups: “Another program: T-X trainer: firm concept to first flight in 36 months; in secret, internally funded
  • Toups: “Success factors: establish team culture/vision from Day 1; define knowledge gaps/solve toughest issues first; challenge norms
  • Toups: “Maximize learning throughput; focus on full kit (powerful way to lower WIP: don’t start work until u hv all parts)” (Cool!)
  • Toups: “Most quality are trailing metrics (error on drawing disc during assembly); full kit is best leading indicator of quality” (Huh!)
  • Toups: “36 months to first flight: 2 jets; an awesome achievement; would love to tell u more, but we’re still in competition”
  • Toups: “Next big challenge: Boeing Phantom Express: DARPA challenge: 10 rocket launches in 10 days; 1st launch 2020” (Cool!)
  • Toups: “10 launches in 10 days; first and biggest program of OPD; using TOC before program start; planning full kits points
  • Toups: “We’re thinking thru where our real constraint is: structural design? Engine design? Where are our risks? Full kits helps find constraints
  • Toups: “Our teams appreciate it, despite being held back from doing work; shows leadership support; we think 4-5x faster design/execution
  • Toups: “What’s striking is how much full kit can shorten wait times; no longer forced to start work too early; shave off all wait time
  • Toups: “Shaving off all wait time potential is huge: touch time still only 10-25% of all scheduled time”
  • it occurs to me that full kit principle is different for software; goal of arch is to minimize dependency; harder for physical goods
  • Toups: “We think OPD is likely not as powerful as CCPM, but its benefits make it more powerful for us

Up: Dr. Hirotoshi Uehara, VP Panasonic Automotive


  • Today I’ve been using a version of TweetSciber rewritten from TypeScript/React to ClojureScript/React/Re-Frame. So fun! So easy to change!