IT Lean Six Sigma - PowerPoint PPT Presentation

1 / 40
About This Presentation
Title:

IT Lean Six Sigma

Description:

... Six Sigma ... In manufacturing, six sigma quality equals 3.4 defects per ... Lean, Six Sigma, and TOC are based on the following: People are problem solvers ... – PowerPoint PPT presentation

Number of Views:103
Avg rating:3.0/5.0
Slides: 41
Provided by: nni7
Category:
Tags: lean | sigma | six

less

Transcript and Presenter's Notes

Title: IT Lean Six Sigma


1
IT Lean / Six Sigma
  • SIM Utah
  • December 13, 2007

2
But First . . . Some Context
  • All I am, or ever hope to be, I owe to my
    association with SIM Utah.
  • Join Now!

3
Why Should We Care?
  • There is evidence it works (incredibly well) for
  • Manufacturing
  • Engineering
  • Product Development
  • Why not for IT?

4
Foundation Principles
  • Measure decisions and activities against
  • Does this create and deliver value?
  • Does this respect humans?
  • Does this respect the process?

5
Will This Create and Deliver Value?
  • Who defines value? The process customer.
  • One way to focus on value is to eliminate waste.
    What is waste? Anything that does not deliver
    value

6
Defining Value
  • What would a customer (external and internal) pay
    for?
  • Example When you order a pizza, what are you
    willing to pay for?
  • Pizza dough
  • Sauce
  • Toppings
  • Toppings dropped on floor
  • Energy for ovens
  • Energy for ovens left on overnight
  • Delivery driver with wrong directions

7
Forms of Waste
  • Rework (correcting mistakes)
  • Waiting (halts and stalls)
  • Over processing (making it better than it has to
    be)
  • Inventory (too much in-process and finished)
  • Motion (not near point of use)
  • Movement (moving things around)
  • Over production (too much too early)

8
How Does This Apply to IT?
  • Rework (pretty much everything I have done)
  • Waiting (approvals and workflow)
  • Over processing (64 of features and functions
    are rarely, if ever, used).
  • Inventory (40 shelf ware)
  • Motion (poor access to expert users)
  • Movement (Alistair Cockburn, People wont climb
    stairs to get an answer)
  • Over production (licensing)

9
What Is Lean?
  • The systematic elimination of waste.
  • What tools help us eliminate waste?
  • Value Stream Mapping
  • 5S
  • Measurables
  • One-piece flow
  • Visual controls
  • Standardized work
  • Problem solving (6 Sigma, TOC)

10
Value Stream Mapping
  • Get the involved people together to create a
    visual representation of the process in two
    forms
  • Current State
  • Future State (the one that has reduced the
    waste).
  • Why? It makes the waste visible (and gets
    everyone together)

11
Example
12
IT Purchasing Current State
Get Manager Approval
Get IT Manager Approval
Request Something
2 5 Days
Same Day
Order Stuff
Receive Stuff
Who is the customer? What creates value for the
customer?
13
Future State
Order Stuff
Get Manager Approval
Request Something
Same Day
Receive Stuff
14
5S
  • Sort
  • Set in order
  • Shine
  • Standardize
  • Sustain

15
5S Example
16
Sort
17
Set in Order / Shine
18
Standardize
19
Sustain
20
Sounds Easy (Wrong!)
  • Of the 5 Ss, which is the hardest to do?
  • How many times have you cleaned your garage? Why
    not just once?
  • We need up-front processes to sustain.
  • Architecture (designed for flow and flexibility).
  • Criteria to simplify in advance.

21
5S Applied to IT?
  • Sort what we use from what we have (COA with
    25,000 accounts).
  • Minimize exception handling
  • Who is going to use this feature / function?
  • What do they want to accomplish with this feature
    / function?
  • How often do they need to accomplish this task?
  • If they dont have this feature or function, how
    will they accomplish this task?
  • System stratification and treatment (A/B/C)

22
Measurables
  • Used to monitor and improve the process (not
    disguise performance or make us feel good).
  • Bad example On time delivery.
  • Good example Top 10 issues and trends.

23
Measurables Guidelines
  • Measurables should
  • Align with strategy and objectives
  • Be few in number
  • Be mostly non-financial
  • Motivate the right behavior
  • Be simple
  • Measure process not people
  • Be visual
  • Show trends

24
One Piece Flow
  • Deliver what customers want when they want it.
  • The opposite of batch and queue processing.
  • Sounds like iterative development and
    implementation to me.
  • In IT, short projects, frequent iterations, 2-day
    task limit.

25
Visual Controls
  • We know what is happening by looking (not by
    running reports or asking someone).
  • Link to meaningful measurables.
  • For example, project status (the one page project
    manager) and number of medium and low severity
    issues.

26
Visual Controls
27
Standardized Work
Inconsistent
Inconsistent
Results
Process
Desired
Consistent
Results
Process
Standardized Work
28
Standardized Work
  • Use a consistent process as a springboard to
    improvements.
  • For example, project management, changes to
    production systems, testing, et cetera.
  • Leadership Standard Work audits the standard work
    while creating improvement opportunities.

29
Problem Solving - Six Sigma
  • A statistical measure of variation using standard
    deviations from a centerline.
  • In manufacturing, six sigma quality equals 3.4
    defects per million parts (in other words, not
    much variation).
  • Why should we care? Focus on variation to learn
    and improve (not blame).

30
Six Sigma Diagram
Standard
6 Sigma
3 Sigma
Process Variability
31
Examples of Learning From Variation
We scope a project to last 60 days. It takes 50
days to complete (a 10 day variation). What went
right? What can we learn from the variation? How
can we use this learning to improve our
processes? Our standard is SPAM filtering
performance of less than 60 seconds. We are now
at over 2 minutes. What caused the variance?
Firmware not current. How does our process need
to change in order to eliminate this variance
from ever occurring again (stop the bleeding and
find the knife).
32
Translating Data into Knowledge
  • In order to learn from variations, we need to
    establish a baseline or standard and then gather
    data to identify variations.
  • The baseline or standard can come from
  • Current performance.
  • Desired performance.
  • Benchmarking (industry, best practices, et cetera)

33
Problem Solving - Theory of Constraints
  • System throughput is defined by the capability of
    the constraint.
  • Increase throughput by improving the constraint.
  • Where is the constraint? Stuff piles up in front
    of the constraint and there is a drought behind
    the constraint.
  • What are our constraints?

34
Increasing Throughput
  • Focus on improving the constraint. In the
    meantime, what value is there in working faster
    than the constraint?
  • What do we do with non-constraint activities?
  • From Alistair Cockburn
  • Have them sit idle.
  • Have them help with the constraint.
  • Have them improve the input to the constraint.
  • Et cetera.

35
Testing / QA Example
Test / QA
How can we increase throughput? Hire more
resources. Give Test less to do (TDD).
Reallocate testing activities. Eliminate some
of the 64 that is rarely, if ever used.
36
Respect for Humans
  • Lean, Six Sigma, and TOC are based on the
    following
  • People are problem solvers (not problems).
  • Problems are an opportunity to learn.
  • We blame and fix processes not people.
  • Many minds are better than one.
  • In other words, a high trust, collaborative
    culture.

37
Respect For Learning
  • Focus on the process and the results will follow.
  • Remove fear (fear stifles innovation,
    participation, and collaboration).
  • Be patient.
  • Practice being leaders as teachers (rather than
    bosses).

38
Respect For The Process
  • Go, look, and understand (this is not MBWA).
  • How can we improve the process if we do not
    understand the process?
  • In IT
  • Days-in-the-life
  • Reduce email decisions and decrees

39
References
  • The Goal (Eli Goldratt)
  • The Toyota Way (Jeffrey Liker)
  • Lean Software Development (Mary Poppendieck)
  • Crystal Clear (Alistair Cockburn)
  • Real Numbers (Jean Cunningham and Orest Fiume)

40
Want More Stuff?
  • Training materials on the lean tools (5S,
    standardized work, measurables, problem solving,
    visual controls, et cetera)
  • nnick_at_headwaters.com
Write a Comment
User Comments (0)
About PowerShow.com