deliver value: lean kanban for portfolio prioritization

2

Click here to load reader

Upload: ram-srinivasan

Post on 12-May-2015

59 views

Category:

Business


1 download

DESCRIPTION

Hand-put of my talk at PMI Westchester PDD (http://www.pmiwestchester.org/chapter_event_details.asp#299) and at PMI NYC Symposium (http://pminyc.org/sites/default/files/Symposium2013-brochure-FINAL.pdf)

TRANSCRIPT

Page 1: Deliver Value: Lean Kanban  for Portfolio  Prioritization

Deliver Value: Lean Kanban for Portfolio Prioritization - by Ram Srinivasan

http://linkedin.com/in/ramvasan email: [email protected] blog: http://ramvasan.com

Deliver Value: Lean Kanban for Portfolio Prioritization by Ram Srinivasan is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Tips for managing queues

• Queues are the root cause of the majority of economic wastes in product development. In knowledge

work, product development inventory is physically and financially invisible, reduce inventory (waste).

• Capacity utilization increases queues exponentially, variability only increases it linearly. So, instead

of maximizing resource utilization, control queue size. Reduce variability by pooling variability.

• Manage queues by visualizing, limiting WIPs and monitoring using Cumulative Flow Diagrams.

• Model flow of projects: M/M/n queues (e.g. n agents with one queue) are better than M/M/1 (one

queue for each agent) queues.

• Reducing batch size (limit WIP) reduces cycle time and consequently reduces queues.

• Critical and centralized resources should have low level of capacity/utilization to reduce queues

• Source of queues in knowledge economy - Marketing, Analysis, CAD, Procurement, Prototyping,

Testing (most common and most risky) , Management Reviews, Specialists

Tips for managing flow by managing batch size

• Reducing batch size reduces variability in flow, accelerates feedback and reduces risk

• In knowledge work, large batch size reduces flow, reduces overall efficiency, and causes exponential

cost and schedule growth.

• Economics of batch size is U-Curve optimization problem (between transaction and holding cost).

Manage flow by reducing transaction cost for batch size (e.g. automated testing).

• Know the difference between productions batch (batch size of the process that changes the state of

the product) and transport batch (batch size that changes location of the product). Smaller transport

batches allow us to overlap activities, reduce cycle time and accelerate feedback.

• Dispersed teams tend to use large batch asynchronous communications. Proximity enables faster

synchronous communication, smaller batch sizes and faster feedback.

• Sequence first that which adds most value cheaply (or reduces risk cheaply).

• Reduce batch size before you elevate bottleneck (Theory of Constraints – Step 4).

• Source of large batches – Project Scope, Project Funding, Project Phases, Requirement Definition,

Project Planning, Testing, Capital Spending, Design Reviews, Market Research, Post-mortems.

Tips for controlling flow under uncertainty

• Limit WIP to sustain high throughput. When WIP is consistently high, purge low value projects.

• Use forecasts of expected flow times to make congestion visible (e.g. wait times for starting projects)

• Exploit pricing to reduce demand during congestion. Differentiate quality of service by work streams.

• Use cadence to enable small batch sizes and to make waiting times predictable (Product

Introduction Cadence, daily build-test cycles, operational review, Program level Stand-ups).

• Exploit scale of economics by synchronizing work from multiple projects.

• When delay costs are homogenous, do the shortest job first. When job durations are homogenous,

do the high Cost of Delay (CoD) job first. When duration and CoD are not homogenous, use WSJF.

• Make tasks and resources reciprocally visible at adjacent processes.

• For faster responses, pre-plan and invest in flexibility. Pull high-powered experts to emerging

bottlenecks. Develop people who are deep in one area and broad in many.

Useful Metrics to Track Queues: Design-In-Process (DIP) Inventory, Queue Size and Trends, Cost of Queues, Aging of Items

Batch Size: Batch Size, Trends in Batch Size, Transaction Cost per Batch, Trends in Transaction Cost

Feedback: Feedback speed, Decision Cycle Time, Aging of Problems

Flexibility: Breath of Skill Set, Number of multipurpose resources, Number of Processes with Alternate Route

Other: Processes Using Cadence, Trends in Cadence, Capacity Utilization Rate, Efficiency of Flow, DIP turns

Page 2: Deliver Value: Lean Kanban  for Portfolio  Prioritization

Deliver Value: Lean Kanban for Portfolio Prioritization by Ram Srinivasan is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Wastes Examples

Overproduction of Information

• Producing more than needed (by next process)

• Creating documents that were not requested

• Redundant tasks, unneeded tasks

• Over-dissemination, i.e. sending information to too many people (e.g.

excessive emails, big teams)

• Information churn

• Reinventing the wheel, lack of reuse

Waiting

• People waiting for information or decision

• Information or decisions waiting for people to act

• Large queues

• Long approval processes

• Unnecessary serial effort

Unnecessary Movement of

Information

• Hand-offs (to another person or functional silo)

• Extensive information distribution

• Dis-jointed facilities, geographical distribution of work, non-colocation

Over-Processing of

Information

• Refinement beyond what is needed

• Point based designs made too early, causing iterations

• Use of excessive complex processes

• Lack of transparency in process

Inventory of Information

• Excessive time intervals between sequential steps

• Poor 5 S (Sorting, Straightening, Systematic Cleaning, Standardizing

and Sustaining) practices

• Keeping more information than needed

Unnecessary Movement of

People

• Context switching during task execution

• People having to move to gain or access information

• Manual interventions to compensate for lack of process

Rework/Defects

• The “Re”s - Rework, Rewrite, Redo, Reprogram, Retest

• Unstable requirements

• Uncoordinated complex tasks

• Incomplete, ambiguous or inaccurate information

• Manual inspections to catch defects

Loss of tactical

Knowledge/Skills

• Lack of multi-learning, functional silos, specializations, working to job

titles,

• Knowledge and information scatter

• “Key man” dependencies (e.g. permissions, access)

• Creating dependencies through “knowledge hoarding”

• Personnel turn-over (attrition, getting moved to different projects,

etc.)

In knowledge work, most of the times, these wastes are translated as invisible queues and

are usually manifested as Delays (Longer Cycle Time)