dan toma presentation on iwmc 2015
TRANSCRIPT
Everything we do can be viewed as: if activity X is performed, then result
Y is expected
key for experimental design
Everything we do can be viewed as: if activity X is performed, then result
Y is expected
key for experimental design
Defining a causal relationship between two phenomenons is
the key to the scientific approach
steps to experimental design
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
Design and run the experiment using the minimum set of features
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
Design and run the experiment using the minimum set of features
Outline the failure criteria
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
Design and run the experiment using the minimum set of features
Outline the failure criteria
Time box your experiment
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
Design and run the experiment using the minimum set of features
Outline the failure criteria
Time box your experiment
Collect, aggregate and interpret the actual data
get your A0 experiment board http://bit.danto.ma/expboard
steps to experimental design
Define what you want to test and the implication for your business model.
Specify the group the experiment is designed for
Design and run the experiment using the minimum set of features
Outline the failure criteria
Time box your experiment
Collect, aggregate and interpret the actual data
Take a decision based on the results
get your A0 experiment board http://bit.danto.ma/expboard
don’t need to test everything
Analog: successful predecessor companies that are worth mimicking in some way.
Reference: J. Mullins & R. Komisar
don’t need to test everything
Antilog: predecessor companies compared to which you explicitly choose to do things differently.
Reference: J. Mullins & R. Komisar
don’t need to test everything
Leap of faith: what you have no idea about.
Reference: J. Mullins & R. Komisar
characteristics of a good experiment
1. Reveals causal relationship between phenomenons.
Reference: C.M. Christensen & K.Popper
characteristics of a good experiment
1. Reveals causal relationship between phenomenons.
2, It’s fast enough so there’s still time for further iterations.
Reference: C.M. Christensen & K.Popper
characteristics of a good experiment
1. Reveals causal relationship between phenomenons.
2, It’s fast enough so there’s still time for further iterations.
3. It’s cheap enough so there are still money left for the next shot.
Reference: C.M. Christensen & K.Popper
characteristics of a good experiment
1. Reveals causal relationship between phenomenons.
2, It’s fast enough so there’s still time for further iterations.
3. It’s cheap enough so there are still money left for the next shot.
4. Didn’t cause irremediable damage to the vital assents of the business.
Reference: C.M. Christensen & K.Popper
characteristics of a good experiment
1. Reveals causal relationship between phenomenons.
2, It’s fast enough so there’s still time for further iterations.
3. It’s cheap enough so there are still money left for the next shot.
4. Didn’t cause irremediable damage to the vital assents of the business.
Reference: C.M. Christensen & K.Popper
5. Is falsifiable, since no number of positive outcomes can confirm a scientific theory, but a single counterexample is logically decisive: it shows the theory, to be false.
some times you win, some times you loose learn
More important than managing the rate of failure is managing the cost of failure.
some times you win, some times you loose learn
According to Thomas A. Edison, you don’t fail, you just find ways
in which it doesn’t work.
More important than managing the rate of failure is managing the cost of failure.
www.thecorporatestartupbook.com
@toma_danwww.danto.ma
@tendayivikiwww.tendayiviki.com
Let’s talk!