1.
Agile Project Management in a Waterfall World
Managing Sprints with Predictive Metrics
Kevin Thompson & cPrime Guest Presentation by John Carter
February 11, 2014
JOHN CARTER
TCGEN INC
jcarter@tcgen.com

2.
•
•
•
•
MANAGING SPRINTS WITH PREDICTIVE METRICS
BIOGRAPHY JOHN CARTER
John Carter is Principal of TCGen Inc. and currently
serves on the Board of Directors of Cirrus Logic
(CRUS).
He co-authored “Innovate Products Faster” with
Jeanne Bradford, a visual handbook on tools for
teams
Prior to TCGen, he consulted to leading Silicon Valley
technology companies. John was the architect of
Apple’s product development process (ANPP) in use
by all divisions.
Before starting PDC, John was Chief Engineer of
BOSE Corporation where he was the co-inventor of
Bose’s Noise Cancelling Headphones. He earned
his MS in electrical engineering from MIT.
2

3.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILITY CASE STUDIES - SUCCESS STORIES
• Shortening of internationalization of software from 11 months
down to 1 month
• Reducing time to release from every 8 months down to every 2
months
• Having a savings of 40-50% of product development cycle time
But all of these successes were in SW… how can we apply to
Physical Products/Systems?
Conducted research of case studies where Agile methods were
applied to Products/Systems with over a dozen participants
2/11/2014
3

4.
MANAGING SPRINTS WITH PREDICTIVE METRICS
cPrime & TCGen 2014 Agility Study
• Kevin Thompson and John Carter plan a study Agile/Scrum
can be applied outside software
• We are also exploring the best practices found in coordinating
multiple Agile projects
• We will use this Webinar TODAY to recruit qualified study
participants
• If you have had experience in these areas we would like to talk
to you!
• There is no charge for participation and you will receive an in
depth participant report
• Participation will require 1-2 hours of time
• We will follow up at the end of this talk and ask for participants
Are you interested?
2/11/2014
4

5.
MANAGING SPRINTS WITH PREDICTIVE METRICS
HOW DO WE APPLY BEST PRACTICES TO HARDWARE?
The challenge is how do we apply Agile Software Methods to
programs that…
• Have components with long lead times?
• Development partners that do things their own way?
• Long supply chains with components, sub-assemblies, and
final assemblies that need integration around the world?
• Medical products that require FDA compliance?
• Large software platforms that are developed using Waterfall
Methods?
• …by adopting a couple of changes – easy to say, hard to do
1. AGILE TEAM SOFT SKILLS
2. SHORT INTERVALS WITH FEEDBACK
2/11/2014
5

6.
Freeze the Specs and Don’t Look Back
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
HOW IS AGILE DIFFERENT FROM WATERFALL?
Develop Fast, Learn, Improve
6

7.
7
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
KEY AGILE & DEVELOPMENT TERMS

8.
1.
Business people and developers work together daily
2.
Projects require motivated individuals, support & trust
3.
Face-to-face conversation is most efficient
4.
Agile processes promote sustainable development
5.
Continuous attention to technical excellence
6.
Simplicity---is essential
7.
The best designs emerge from self-organizing teams
8.
At regular intervals, the team reflects
9.
Welcome changing requirements
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE MANIFESTO – HARD TO DO IN SYSTEMS?
Software Specific
10. Continuous delivery of valuable software
11. Deliver working software frequently
12. Working software is the measure of progress
75% of the Agile Manifesto CAN apply to development of any type
2/11/2014
http://agilemanifesto.org/
8

9.
MANAGING SPRINTS WITH PREDICTIVE METRICS
SCRUM METHODOLOGY – BEST PRACTICES
The most common implementation of the Agile Manifesto is Scrum
2/11/2014
9

10.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE METHODOLOGY – BEST PRACTICES
Question: What are the most impactful elements of
Agile/Scrum applied to SW?
The top Four Scrum practices can be applied to Systems too!
2/11/2014
10

11.
Question: What are the most impactful elements of
Agile/Scrum applied to Products/Systems?
Three of the top “Agile” practices in Systems have little to do with Scrum
2/11/2014
11
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE METHODOLOGY –APPLIED TO HW/SYSTEMS

12.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE METHODOLOGY –APPLIED TO HW/SYSTEMS
So why is applying Agile/Scrum to HW/Systems so hard?
2/11/2014
12

13.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE METHODOLOGY –APPLIED TO HW/SYSTEMS
Because Agile/Scrum is hard
It requires higher process literacy
And greater cross functional teamwork skills
For more sophisticated teams; not for the faint of heart
The rest of the presentation takes you through the findings
from our study and provides you with some practical ideas for
implementation… …starting with the soft side and culture
2/11/2014
13

14.
MANAGING SPRINTS WITH PREDICTIVE METRICS
FIRST, APPLY THE FOLLOWING CULTURAL ASPECTS
Adopt the concept of
1. High performance teams
2. Self organized teams
3. Trust and empowerment
4. Customer owner & team interacting daily
5. Daily standup meetings
AND…
• Accept the fact that requirements may change
You might just say that this is just common sense!
2/11/2014
14

15.
1.
2.
3.
4.
5.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
THEN, TRANSLATE SCRUM TO WATERFALL
User Stories into Boundary Conditions
Burn-down charts into Deliverable Hit Rate
Sprint into HW intervals
Manage the project with Out of Bounds Process
Sprint Retrospectives into Event Timeline Retrospectives
15

16.
MANAGING SPRINTS WITH PREDICTIVE METRICS
HOW TO SYNCHRONIZE AGILE SW TO PRODUCTS/SYSTEMS
Side note
Target HW integrations to Sprint End Points
• If HW schedules move, target Sprint n+1 or Sprint n+2
Place SW engineers close to metal on HW team to shield SW
Use emulation/simulation extensively to buffer the HW/SW
interface
• PC’s to simulate mobile devices
• Custom Sprints to develop emulators
And when all else fails…use a very good Project Manager
2/11/2014
16

17.
MANAGING SPRINTS WITH PREDICTIVE METRICS
1. CREATING BOUNDARY CONDITIONS
• A program consists of product attributes and program
attributes
• Boundary conditions typically have both
• Create User Stories – Product Attributes
• Create budget and schedule – Program Attributes
• Select the top 3-7, define limits, and seek agreement with the
management team
As a <type of user> I want <some goal> so that <some reason>
THIS BECOMES YOUR BOUNDARY CONDITIONS… STAY INSIDE
THEM AND THE TEAM CAN KEEP MOVING FORWARD!
http://www.mountaingoatsoftware.com/topics/user-stories
17

18.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
1. EXAMPLE BOUNDARY CONDITIONS
18

19.
MANAGING SPRINTS WITH PREDICTIVE METRICS
1. EXAMPLE BOUNDARY BREAK
• Deliverable Hit Rate
too Slow!
• Key Engineer pulled!
• Three week delay!
2/11/2014
19

20.
MANAGING SPRINTS WITH PREDICTIVE METRICS
2. TRANSLATE BURNDOWNS INTO DELIVERABLE HIT RATE
Identify the key requirements that should be satisfied during
an interval
• Can be features implemented
• Can be tests validated
• Can be tasks performed
• Can be a customized metric of progress
This list of requirements can vary from interval to interval
• Front end is more definition loaded
• Middle is more task loaded
• Back end is more validation loaded
Create a target curve over the sprint interval
• Don’t get too stressed out over perfection
• Assume that the events can be distributed
evenly, unless you have clear knowledge otherwise
2/11/2014
20

21.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
2. TRANSLATE BURNDOWNS INTO DELIVERABLE HITRATE
21

22.
MANAGING SPRINTS WITH PREDICTIVE METRICS
2. EXAMPLE OF PCB LAYOUT PROGRESS
Number of Unrouted Nets
1400
1200
1000
800
600
400
200
0
1-Aug
2-Aug
3-Aug
4-Aug
5-Aug
6-Aug
7-Aug
8-Aug
9-Aug
10-Aug
• Aug 1 started tracking PCB routing progress to get an
idea of project velocity
• Aug 3 worried about progress, rate too slow
• Aug 4 increased # of engineers assigned to this task
Example how a Burn Down Chart can be applied to see the
progress in turning a schematic into a layout
2/11/2014
22

23.
MANAGING SPRINTS WITH PREDICTIVE METRICS
3. TRANSLATE SPRINT INTO HW INTERVALS
Divide the project into the smallest increment possible that
represents TRUE INTEGRATION POINTS or CLEARLY
DEFINABLE MILESTONES.
2/11/2014
23

24.
MANAGING SPRINTS WITH PREDICTIVE METRICS
3. TRANSLATE SPRINT INTO HW INTERVALS
Continuous learning, short intervals, measurable progress, autonomy
The secret to getting the benefits of Agile development is to
shed features and not slip the interval
2/11/2014
24

25.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
4. BOUNDARY CONDITION PROCESS
25

26.
MANAGING SPRINTS WITH PREDICTIVE METRICS
5. EVENT TIMELINES AND RETROSPECTIVES
Don’t use Post Mortem
Case Study
2/11/2014
26

27.
MANAGING SPRINTS WITH PREDICTIVE METRICS
RETROSPECTIVES
• Retrospectives should be carried out on all programs
• The retrospectives should follow a common process which has
the following attributes
• Fact based, and data driven
• Involve Cross-functional team members
• The retrospective process should be owned by the team
• The retrospective process should be used during every Interval
• The process consists of the following steps
1. Event time lines & Prioritization of the biggest events
2. Root Cause Analysis
3. Affinity Diagram to summarize results
2/11/2014
27

28.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE – BEST PRACTICES FOR HW/SYSTEMS
Study Participation
1. If you have had experience with the application of
Agile/Scrum/Iterative development out of software
environments… OR
2. If you have had experience with governance of multiple
Agile/Scrum programs we would like to talk to you!
Please respond positively to our poll…
“Are you interested in participating in an Agile Best
Practices Study?”
2/11/2014
28

29.
After the webinar…
• We will send directions to collect the PDU you will earn
from attending this webinar
• We will also send a links to the recorded webinar and
presentation slides once they are posted online
For more information, visit www.cprime.com

30.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE – BEST PRACTICES FOR HW/SYSTEMS
Thank You!
2/11/2014
30

31.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE – BEST PRACTICES FOR HW/SYSTEMS
Appendix
2/11/2014
31

32.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
ANOTHER APPROACH: WATER-SCRUM-FALL
32

33.
2/11/2014
MANAGING SPRINTS WITH PREDICTIVE METRICS
3. THE FIVE KEY ELEMENTS OF AGILITY
33

34.
MANAGING SPRINTS WITH PREDICTIVE METRICS
4. BOUNDARY CONDITION PROCESS
• What is the Tool?
 Used to realign teams when a project has gone out of bounds
 help course correct and realign to a new plan
• Which Business Problems Does the Tool Solve?
 Is an effective recovery vehicle when projects run into trouble
 Creates mechanism to quickly align project and management
teams
 Eliminates creating an exception-handling process each time a
deviation occurs
• Benefits
 Helps you realign projects within hours/days, not days/weeks
 Empowers the team to move forward with minimal guidance
 Minimizes team confusion by establishing a single agreedupon limits
 Engages the team because of the greater trust by
management
2/11/2014
34

35.
MANAGING SPRINTS WITH PREDICTIVE METRICS
AGILE – BEST PRACTICES FOR HW/SYSTEMS
Side note
1. User Stories needed to make the best tradeoffs
2. Plan to have multiple DVTs that involve the customer
3. Build multiple variants in DVTs (buttons, PCB layouts,
Firmware)
4. If lead time is 6 months and you need to go to market in 4,
ask for their buffer stock
5. Use reference designs & allocate BOM choices to supplier
6. The entire team must be 'Agile' for it to work
7. Rely on Postponement
2/11/2014
35