The document discusses using experiments and failure as a way to learn and improve. It provides examples of experiments that could be run by product teams and to improve company culture. The key messages are:
- Experiments allow teams to test assumptions in a minimal and low-risk way to learn without large failures.
- Both product features and cultural initiatives within companies can be tested through experiments.
- Learning is optimized when experiments have around a 50% chance of success; this allows teams to fail fast and learn quickly.
- Companies should aim to run more, smaller experiments over time to accelerate learning.
8. Can We Use Failure to
Learn More and Faster?
The only way to win is to learn
faster than anyone else.
- Eric Ries, The Lean Startup
9. We need a scientific approach
to define, run and learn from
experiments
10. Lean Startup is About Experimentation
ideas
productdata
measure
Eric Ries, The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation…, 2011
11. An experiment is a (scientific) method to
test whether a small next step can
bring us closer to our goal
Now
Next
Goal
Experiments
Image by Tobias Leonhardt
13. LEAN ADVANTAGE Ltd. www.leanadvantage.co.uk
Say hello at: hello@leanadvantage.co.uk
State your Assumption
Make it Minimal
Connect it to Your Goal
Have a Pass Condition
Think Creatively
EXPERIMENT CARD
14. LEAN ADVANTAGE Ltd. www.leanadvantage.co.uk
Say hello at: hello@leanadvantage.co.uk
State your Assumption
Make it Minimal
Connect it to Your Goal
Have a Pass Condition
Will bring us closer to <goal >
To verify that we will
<run this small test>
We will know we were right if
<Quantitative or qualitative pass
criteria>
We assume that <doing x>
Think Creatively
EXPERIMENT CARD
18. Goal: strengthen our culture by practicing
our values
Assumption: Gratitude (KUDO cards)
will help us live our values
Test: Run a weekly cake meeting
at our wall and distribute KUDOs.
Success if >5 KUDOS distributed
23. Learning is optimal when we have a 50/50 chance of succeeding.
MISTAKES EXPERIMENTS PRACTICES
LEARNING
SUCCESS
FAILURE
24. •Which ended up as a success?
•Which failed?
•What have you learned?
Share an Experiment
25. Don’t change things; run experiments.
Label your next new idea
an experiment, and let
everyone know that you
are just testing it out.
- David & Tom Kelley, Creative Confidence
26. We need to create
moments to evaluate
our experiments.
I suggested that maybe we
should have a big bell in the
office, so that we could ring it
whenever there was something
to celebrate.
source: #Workout, “Yay! Questions & Celebration Grids” http://m30.me/workout
27. Set an explicit constraint on
learning and experimentation
20% Time
Exploration Days
Learning Tribes
33. Test: Run a month-long A/B test
to measure impact of the new feature
We will know we were right if
new feature generate host sign-ups
and conversion to active hosts
of 30%
35. LEAN ADVANTAGE Ltd. www.leanadvantage.co.uk
Say hello at: hello@leanadvantage.co.uk
State your Assumption
Make it Minimal
Connect it to Your Goal
Have a Pass Condition
Will bring us closer to <goal >
To verify that we will
<run this small test>
We will know we were right if
<Quantitative or qualitative pass
criteria>
We assume that <doing x>
Think Creatively
EXPERIMENT CARD
36. Learning is optimal when we have a 50/50 chance of succeeding.
MISTAKES EXPERIMENTS PRACTICES
LEARNING
SUCCESS
FAILURE
37. LEAN ADVANTAGE Ltd. www.leanadvantage.co.uk
Say hello at: hello@leanadvantage.co.uk
State your Assumption
Make it Minimal
Connect it to Your Goal
Have a Pass Condition
Share Progress Visually
Document What We Have Learned
Run one at a time
Sequence based on risk/investment
Run in a timebox
Think Creatively
10 TIPS FOR GREAT MINIMUM VIABLE
EXPERIMENTS
38. Run more experiments, faster, and cheaper.
What you want to do as a company is
maximize the number of experiments
you can do per unit of time.
- Jeff Bezos, Harvard Business Review