I would need more data on how many businesses failed based on ownership of the chair.
If you told me that chair had been used by n successive failed businesses, there is a value of n where I would reject the null hypothesis that the chair has no association with the business failing.
I don't think n=1 is sufficiently large.