I am sure there are a lot of management books out there that define SUCCESS, and I am sure they have their respective models to describe the probability of SUCCESS.
I was just thinking about modeling the probability from a startup perspective. Following are the assumptions this model makes -
1. It takes 'n' number of iterations before an idea is a HIT, where n is greater or equal to 1
2. There is a probablity of success associated with each iteration, which implies that there is also a probablity of failure associated with each iteration.. duh!!
Model 1: Using geometric distribution
This is the most conservative and simplistic approach to model success
p(HIT) = Probability of success per iteration
P(SUCCESS) = p(HIT) x (1 - p(HIT))^n
This is simple in the sense that each new iteration does not carry the benefits from the previous iteration
Model 2: Bernoulli's trials based model
This probably is more realistic of the two models
p(HITn) = Probability of success in the nth iteration = P(HITn-1) +/- deltaP
Similarly,
p(~HITn) = Probability of NO success in the nth iteration = P(~HITn-1) +/- (1-deltaP)
where deltaP = Probability of iteration (n-1) increasing/decreasing the probability of success in the nth iteration
P(SUCCESS) = p(HIT0) x p(HIT1) x ... x p(HITn-1) x p(HITn) = 1 - p(~HIT0) x p(~HIT1) x ... x p(~HITn-1) x p(~HITn)
where p(HITn) tends to 1 AND p(~HITn) tends to 0, We need to be deterministic if we have to be optimistic... ;-)
Any thoughts on this??? Do you think this model is right/wrong?? Do you have your own models??? Please do share!!!
I was just thinking about modeling the probability from a startup perspective. Following are the assumptions this model makes -
1. It takes 'n' number of iterations before an idea is a HIT, where n is greater or equal to 1
2. There is a probablity of success associated with each iteration, which implies that there is also a probablity of failure associated with each iteration.. duh!!
Model 1: Using geometric distribution
This is the most conservative and simplistic approach to model success
p(HIT) = Probability of success per iteration
P(SUCCESS) = p(HIT) x (1 - p(HIT))^n
This is simple in the sense that each new iteration does not carry the benefits from the previous iteration
Model 2: Bernoulli's trials based model
This probably is more realistic of the two models
p(HITn) = Probability of success in the nth iteration = P(HITn-1) +/- deltaP
Similarly,
p(~HITn) = Probability of NO success in the nth iteration = P(~HITn-1) +/- (1-deltaP)
where deltaP = Probability of iteration (n-1) increasing/decreasing the probability of success in the nth iteration
P(SUCCESS) = p(HIT0) x p(HIT1) x ... x p(HITn-1) x p(HITn) = 1 - p(~HIT0) x p(~HIT1) x ... x p(~HITn-1) x p(~HITn)
where p(HITn) tends to 1 AND p(~HITn) tends to 0, We need to be deterministic if we have to be optimistic... ;-)
Any thoughts on this??? Do you think this model is right/wrong?? Do you have your own models??? Please do share!!!