19
0 We Shall Produce Working Software Chad Vossen [email protected] 9/22/2015

Chad Vossen - We shall produce working software

Embed Size (px)

Citation preview

Page 1: Chad Vossen - We shall produce working software

0

We Shall Produce Working Software

Chad Vossen

[email protected]

9/22/2015

Page 2: Chad Vossen - We shall produce working software

1

About Me

Chad Vossen

- Worked in the Investments IT space for 14 years

(Securian, Travelers, Thrivent)

- 2 years on a Scrum Team

- Married, 3 girls (Charlotte 8, Harper 5, Everleigh 2)

Page 3: Chad Vossen - We shall produce working software

2

My Team – JDI

Page 4: Chad Vossen - We shall produce working software

3

• Flash Build

Focus on Product and

just enough

documentation

39% Success Rate*

• Waterfall Project

Focus comprehensive

documentation and

process

11% Success Rate*

Agile Value #2

Working Software Over Comprehensive Documentation

Financial Times 8/13/2015, Standish Group

On-time, On-budget, all features

Page 5: Chad Vossen - We shall produce working software

4

• End of Life (EOL) for Visual Studio 6.0 C++

• 28 C++ programs

• A web interface to run adhoc jobs

• 2.5 years

• Project Manager, IT Lead, Testing Center of Excellence

(TCOE), BAs, and developers (me)

The Old Way

The Waterfall Project

Page 6: Chad Vossen - We shall produce working software

5

• 5 Project Managers

• 4 IT leads

• 6 People Managers

• Using TCOE

• 20,000 hours

• 28 requirement

documents

• 28 Test case

documents

• 58 change requests

The Old Way

The Waterfall Project

Page 7: Chad Vossen - We shall produce working software

6

• 1 – 2 days to build something awesome, Scrum in a Bottle

• Use when project is stalled, needs a jump start,

technologists SMEs are spread thin

• Benefits

Jump start a project

Build something cool and useful in 2 days

• Roots are in the flash mob concept (Nordstrom’s

Sunglasses app video)

The New Way

The Flash Build (Scrum Day 2014)

Page 8: Chad Vossen - We shall produce working software

7

The New Way

The Flash Build:

First Attempt

• First attempt in October of 2014 building a semi-replacement of an existing

product

Focused on Minimum Viable Product (MVP), used a Story map to set focus

Followed a recommended 2 sprints/day for 2 days (2 planning/demo/retros)

Tried to get dedicated time from all the right people

Had to get stations setup and moved around to 2 different rooms

Page 9: Chad Vossen - We shall produce working software

8

The New Way

The 1st Flash Build Results

• Scheduled around 65 points to be completed and finished around 25 points

Average velocity around that time was ~40 points

• 2 sprints a day was too time consuming, not enough time to work

• Teams focused on what they were doing and didn’t collaborate with other

teams

• Needed more dedicated time from the right people

• Coordinated moving of equipment from room to room

Page 10: Chad Vossen - We shall produce working software

9

The New Way

The Flash Build: Second Attempt

• Creating 20 new house/reference code tables & corresponding UI screens

• Also, load, test and validate our work.

• Dedicated team for 2 days, including the key stakeholders

• Using a newly formed room built

• 2 sprints (1/day vs. 2/day in first attempt)

• Scheduled 47 points, not 65 points like in first attempt

Page 11: Chad Vossen - We shall produce working software

10

Collaboration Room The New Way

• From the learnings

of the first flash build

• New space, new

rooms

Page 12: Chad Vossen - We shall produce working software

11

Task Board The New Way

• Used the same task

for building, testing,

validating

Page 13: Chad Vossen - We shall produce working software

12

Retros The New Way

• Both days on the

same board, made

changes day 2

based on first day

ideas

Page 14: Chad Vossen - We shall produce working software

13

End of Day Goals The New Way

• Met day 2 goal, not

so much on day 1

Page 15: Chad Vossen - We shall produce working software

14

The New Way

The Flash Build: Second Attempt Result

• Created 20 new house/reference code tables & corresponding UI screens

• Loaded, tested and validated our work.

• Additional SOP documentation was created during the process

• Key stakeholders helped with building, loading, testing, and validation

• Instant feedback! The shortest feedback loop!

• 2 teams that collaborated throughout build

• 44 points in 2 days!

Page 16: Chad Vossen - We shall produce working software

15

The Result

Page 17: Chad Vossen - We shall produce working software

16

Compare

Comparison Waterfall vs. Flash Build

• Not Apples to Apples

Existing functionality vs. new functionality

New technology for most of team vs. somewhat proven technology.

• Key: Working software OVER comprehensive documentation

• Documents crafted: Requirements, Test plans, Change requests vs. Design

decision, pattern, stories

• Time: 20,000 hours vs. 136 hours

Page 18: Chad Vossen - We shall produce working software

17

Compare

Waterfall or Flash build?

Activity EOL Hrs EOL % Total EOL % FB FB Hrs FB % FB as EOL EOL w/ FB%

PM/Planning 4600 23% 27% 16 12% 37 753

Requirements 2300 11% 13% 6 4% 18 282

Analysis/Design 3300 17% 19% 8 6% 26 376

Development 4600 23% 27% 98 72% 37 4600

Testing 2200 11% 13% 8 6% 18 376

TOTALS 17000 136 6400

Page 19: Chad Vossen - We shall produce working software

18

Questions?

And thanks for attending!