General statistics are more helpful when the target group is homogeneous. Intuitively speaking, when a group has similar resources, goals and problems, what happens to others is more applicable to the individual. Whereas if a group is composed of highly differentiated members, general statistics is likely a result of a mixture of many distinct models, which in this case would be far less useful for the individual.
In case of start-ups, it would seem to me that the founders' motivations, goals, environments and backgrounds all have high degree of variations. General statistics like those in the Project should thus indeed be used with caution.
I wasn't familiar with the Project prior to this, so will look into it some more. The analysis Jason puts forward here reminds me a lot about the business owners I meet in the real world - they want an answer to questions like "If I do X what will happen?" and "Should I do Y or Z?"
The good part about my approach to those clients (broadly, 'coaching' vs 'consulting' but that's way broad) is that I can answer a question with another question like "What are YOU trying to achieve?"
Startups that pivot several times get more funding. "Do you need to pivot? Do you want funding?"
Solo founders take longer to scale. "Do you know any potential good co-founders? Of the potential bad co-founders you know, which ones do you think will help you scale vs break your vision for the company?" etc etc
Oh, and yes, I'm that really annoying guy when you just want a damn answer. But life isn't that simple, and certainly isn't as universal as you or I would like.
On the one hand I agree with the criticisms on survivorship bias and the risks of hypothesis testing. There's also presumably a selection bias. And while Ron Berman's post on methodology [1] talks about how the goals of this stage of the project were to formulate a model that could then be verified, the results are certainly being reported as conclusions.
On the other hand, despite its limitations, it's very interesting data. So the title seems very much an overstatement. "Blindly Believing in Startup Genome Project's Data Considered Harmful" would be a better description -- although far less effective link bait :-)
[+] [-] asknemo|15 years ago|reply
In case of start-ups, it would seem to me that the founders' motivations, goals, environments and backgrounds all have high degree of variations. General statistics like those in the Project should thus indeed be used with caution.
[+] [-] JacobAldridge|15 years ago|reply
The good part about my approach to those clients (broadly, 'coaching' vs 'consulting' but that's way broad) is that I can answer a question with another question like "What are YOU trying to achieve?"
Startups that pivot several times get more funding. "Do you need to pivot? Do you want funding?"
Solo founders take longer to scale. "Do you know any potential good co-founders? Of the potential bad co-founders you know, which ones do you think will help you scale vs break your vision for the company?" etc etc
Oh, and yes, I'm that really annoying guy when you just want a damn answer. But life isn't that simple, and certainly isn't as universal as you or I would like.
[+] [-] nbashaw|15 years ago|reply
[+] [-] jdp23|15 years ago|reply
On the other hand, despite its limitations, it's very interesting data. So the title seems very much an overstatement. "Blindly Believing in Startup Genome Project's Data Considered Harmful" would be a better description -- although far less effective link bait :-)
[1] http://www.systemmalfunction.com/2011/05/deciphering-genome-...
[+] [-] smartbear|15 years ago|reply
(See nbashaw's comment)
[+] [-] kevinpet|15 years ago|reply
Fortune, not so much.