The steps are based on the Lean Analytics approach described by Alistair Croll and Ben Yoskovitz.
The BLOOM Lean Enterprise Experiment Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also references Javelin's Experiment Board. (http://javelin.com/experiment-board.html)
Download at: http://www.bloomamsterdam.com/wp-content/uploads/2014/05/BLOOM-Lean-Enterprise-Experiment-Canvas.pdf
Good Stuff Happens in 1:1 Meetings: Why you need them and how to do them well
BLOOM Lean Enterprise Experiment Canvas
1. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
Goal: to learn
More about BLOOM services at www.bloomamsterdam.com.
2. 2
Start-ups and large enterprises have different requirements for lean
experimentation
Start-up Enterprise Difference in approach
• Reduce most waste by limiting the time
spent on product nobody will use
• Reduce most waste by using roadmap
time only to solve issues that have
impact on the bottom line
• Experiments focus on feature
optimisation, less on new product
development
• Focus on customer development • Focus on customer satisfaction • Improve for existing customer base
first, not look for a new one
• Data-poor, decisions are often made
based on qualitative data.
• Data-rich, prioritise to find the most
important issues to improve
• Problems can be prioritised based on
data
• Still building a brand: it is good to move
fast and break stuff
• Reputation is one of the strongest
assets. Customers rely on quality of
service
• MVPs require a higher level of quality
• Everybody is in, small companies make
collective decisions
• Need top management buy in • Need top management buy in for
experimentation
• Start-ups are a small team with a
common goal
• Enterprises are comprised of many
departments
• Goal of department should be closely
aligned with goal of the company
3. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
4. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
5. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
6. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
7. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
8. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
9. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
10. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Did we reach our success
criterion?
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
11. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Did we reach our success
criterion?
No: Try a different
solution
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
12. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Did we reach our success
criterion?
Did we improve the one metric
that matters enough to cross
the line in the sand?
No: Try a different
solution
Enterprise Canvas is based on Javelin's Experiment Board, it can be distributed freely and derivative work can be created as long as it also reference's Javelin's Experiment Board.
This work is complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
More about BLOOM services at www.bloomamsterdam.com.
13. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
More about BLOOM services at www.bloomamsterdam.com. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. Template inspired by Javelin Experiment Board.
Complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Did we reach our success
criterion?
Did we improve the one metric
that matters enough to cross
the line in the sand?
No: Try a different
solution
No: Work on
solving a new
problem
14. Success
criterion
Get out of the building
Result &
decision
Learning
Date:
Responsible
person:
Project
name:
Step of cycle
Phrase the problem from your customers' perspective. Include the proof point.
☐
Is this really a problem? (validate with clients/data)
☐
Is this one of the most important problems? (estimate impact and apply Pareto principle)
Define the possible solution
☐ Is the solution easy to implement?
☐ Is there a balanced mix of iterative improvements (±80%) and major leaps (±20%)?
Design the test (“I will run experiment #X with Y% of customers in period Z”)
☐ Is this this the fastest way to test my solution with minimal amount of resources (time,
people, money)?
☐ Do actual users give feedback, implicitly or explicitly?
☐ Is the test time-boxed?
Experiments 1 2 3 4 5
Problem
Solution
Test method
The one metric that matters Control metric
Draw a line in the sand:
the one metric you want
to impact, with how
much, by when
If there is a control
metric you want to keep
an eye on, specify here
Current value of metric Current value of metric
Determine what success looks like (“During the test, I expect strong signal from X% of
customers”/visitors)
Describe the results of your tests. Based on the results, take a decision
Describe the lessons learned from the experiment
Did we change the one metric that matters significantly? Are control variables still ok?
If yes: pop champagne, and build the actual solution. If not: pivot/persevere.
LEAN ENTERPRISE EXPERIMENT CANVAS
More about BLOOM services at www.bloomamsterdam.com. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. Template inspired by Javelin Experiment Board.
Complementary to the Lean Analytics approach as described by Alistair Croll and Ben Yoskovitz.
Goal: to learn
% of visitors that answer the
poll “was this answer
useful?” with “YES”
Goal: 10%
4% of visitors reply to
poll with a positive
answer
Number of visitors/
month shouldn’t drop
1.5M / month
Search function does not
provide relevant results
Proofpoint: Low number of clicks
on search results
Add filters to improve
search results
Add filter for key question
categories
Run for 2 weeks with 10% of
visitors
Expect +30% click-through to
answer pages from search
results when visitors use the filter
Navigation structure is
unclear
Proofpoint: Large drop-off
before reaching any answer
Build a new search tool with
higher accuracy
Increase UX of the search
results
Did we reach our success
criterion?
Did we improve the one metric
that matters enough to cross
the line in the sand?
No: Try a different
solution
Yes: Roll out full version of the
solution to 100% of your customers
No: Work on
solving a new
problem