[rachelandrew]

There are also ways you can mitigate the risk to some extent.

If possible build something your existing business needs. That way if it doesn't take off as a product you still have something that meets a need you have.

Launch the smallest thing that solves a problem your prospective customers have. The smallest thing they will be willing to pay for, then iterate from there.

Our product was something we needed in our service business (a small CMS that didn't insert unwanted markup) and took us about 4 weekends to get to launch in terms of the product, and about the same for the infrastructure around it (documentation, sales site and so on). It was tiny compared to where it is now, but enough people wanted it to make it worth continuing development.