This talk was presented by Mike Hall at Southern Fried Agile 2017.

1. Welcome to Southern Fried Agile:
Experiment-Driven Development
1

2. Mike Hall
Senior Agile Coach/Trainer
@MikeH_agile
mike.hall@agilevelocity.com
www.agilevelocity.com
Accelerate Agility
Transformation | Training | Coaching
2

3. Braz Brandt
Agile Coach
@braz
braz.brandt@agilevelocity.com
www.agilevelocity.com
Accelerate Agility
Transformation | Training | Coaching
3

4. Predict AccelerateLearnAlign Adapt
Path to Agility
Status Quo Chaos &
Resistance Integration&
Practice
New Status Quo
The goal for the transformation
cannot be to do Agile. Understanding
and communicating the business
objectives that will be achieved with
the transformation is a critical first
step.
Through Agile training and
coaching, teams and
leadership are equipped
with new techniques and
an understanding of how
Agile works.
Ownership of processes
are transferred to an
empowered team and a
culture of continuous
improvement is put in
place.
Teams harden these newly
learned practices and
become more disciplined in
order to deliver working
product in a predictable and
iterative manner.
Once the teams become disciplined
and predictable, we can focus on team
and organizational improvements to
optimize across the full delivery cycle
and shorten time to market.
Agile will begin to permeate
throughout the organization and
executive leadership, enabling
empowered teams and adaptive
leadership to respond to ever-
changing market demands as they
have transformed to an
organization with true Agility.
www.AgileVelocity.com
info@agilevelocity.com
@agile_velocity
4

5. What if we found
ourselves building
something that nobody
wanted?
In that case, what did it
matter if we did it on
time and on budget?
5

6. Let’s celebrate our failures! ???
6
“A $200M market annually if we can hit the Oct 1 launch date.”

7. Experiment-Driven Development (EDD)
7

8. 8

9. Activity:
Huddle up in pairs or 3s.
Discuss what was different in
this approach vs. Traditional
or your version of Agile.
Consider: your release cycle
length, # features per
release, success criteria, etc.
9
2

10. Debrief - Potential Differences:
10
“Never work on
things that are
not valued by the
customer!”

11. Step 1: Start With a Product Hypothesis
11
Types of hypotheses:
• Customer
• Problem
• Solution
• Value
• Growth
Problem example:
“Potential customers have no easy way to compare different
versions of sunglasses.”
Value example:
“Allowing the customer to ‘compare’ themselves side-by-side in
different sunglasses will enhance their overall experience at our
sunglass station.”
Growth example:
“An in-store available iPad app will increase sunglass sales by 20%.”

12. Activity:
Write out a hypothesis
for a product you are
currently working on
(or recent one).
Identify type as
customer, problem,
solution, value, or
growth type.
12
2

13. Step 2: Identify Assumptions
13
Ask yourself “What assumptions are ‘baked in’ to the hypothesis?”
Identify and prioritize (based on potential amount of learning).
For example:
• All customers will be comfortable using an iPad.
• Side-by-Side is the best comparison technique.
• Most customers will have at least 3 choices of sunglasses that they like.
• Customers don’t like the current approach of trying on sunglasses and using the mirror.

14. Activity:
Write out 2 - 3 baked-
in assumptions for
your product
hypothesis.
Share with a new
friend from another
table you have not
met yet!
14
3

15. Step 3: Identify Experiments
15
Take your highest priority Assumption and define the smallest experiment
that will prove or disprove the assumption.
Examples:
• Assumption: All customers will be comfortable using an iPad.
Experiment: For every customer who looks over 50, show them the
demo and then ask them if they would rather try on glasses by
themselves or use the iPad. Measure the interest %.
• Assumption: Side-by-Side is the best comparison technique.
Experiment 1: Create a multi-layout technique showing up to 6
different pics in the main view
Experiment 2: Demo both approaches (side-by-side and multi-view) to
at least 40 customers. Measure preference %.

16. Experiment Backlogs
• Similar to Scrum product backlog
• But is learning-based prioritization
16
• List of all experiments 1..n
Work items, mocks, research, etc.
• Tagged with Assumption description

17. Nielsen DCAPI project - Learning Map
MVP 1
Assumption
Experiments
17

18. Build
Experiments Measure
Learn
Step 4: Run Experiments
EDD

19. Experiments - Validated
• Build out the MVP
• Measure progress based on validated learning from end user!
• Use modified storyboard showing Validated column
To Do In Work Done Validated
EXP-1
EXP-2
EXP-3
EXP-4
EXP-5
19

20. Experiment Test Iteration (ETI)
Scrum: fixed iteration length
ETI: variable iteration length
ETI 1
3 days
ETI 2
5 days
ETI 3
9 days
ETI 4
17 days
ETI 5
6 days
ETI 6
7 days
MVP1 MVP2 MVP3 MVP4 MVP5 MVP6
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
Sprint 1
Weeks
Sprint 2
Weeks
Sprint 3
Weeks
Sprint 4
Weeks
Sprint 5
Weeks
Sprint n
Weeks
20

21. ETI Review
• Dev team demos their working
software
• Team discusses validated
learnings with stakeholders
• Decision: Pivot/Persevere/Quit
• Plus:
- How is the team feeling about the
assumptions?
- Are there any assumptions not
identified previously?
21

22. SIMULATION
22

23. EDD Simulation: Ball Point Game
•Self-organize into Table teams of 5 - 6
•Take your Agile Velocity ball with you!
•Find some space - move to open area, corner, etc.
Hypothesis:
Our team can score 50 points or more in the
Ball Point game in a single round of 1 minute.
23

24. PlayRules
Ball Point GameObjective: Get as many ball points as possible in 1 minute
•A ball which is passed to each person and
returns to the starting person scores a
point
•Each ball must be touched by each person
•Each ball must spend time in the air
during each pass between people
•A ball cannot be passed in a totally circular
motion (left/right)
•Balls may be re-used
•A ball that drops is counted as a defect
and does not earn a point
•2 min to write down your assumptions
and prioritize them
•2 min to write down your experiment(s)
for the highest priority assumption
•Go announcement: 1 min to move balls
•Stop announcement: record your score
•1 min to adjust your experiments &
assumptions
•Repeat n times

25. Debrief
25
Observations?
Did any team reach 50 in a single round?
How did you use your validated learnings?
Did your Experiments Backlog change as you executed the iterations?
Would more planning time have really helped?
Did any team pivot to a new hypothesis (less balls or more balls)?

26. 26
WRAP

27. 27
EDD in a Nutshell
Product Hypothesis
Assumptions
Experiments
Build-Measure-Learn Loops
Validated Learning
Pivot or Persevere
The Right Product!

28. We value
• Validated learning over seemingly-reasonable
assumptions
• Data-driven decisions over plausible-sounding
arguments
• Building minimum learning products over additional
features
• The courage to build the right thing over something
that works
A Post-Agile Manifesto
28

29. Thank You! 29
mike.hall@agilevelocity.com braz.brandt@agilevelocity.com

30. Thank You! 30
mike.hall@agilevelocity.com, braz.brandt@agilevelocity.com
www.agilevelocity.com
Stop by the Agile Velocity booth if you
have further questions!
For our next Charlotte CSMs, use
code SFA2017 for 10% off Early Bird
& Regular Pricing.