The history of scientific developments has been in the total embracing of failure. Literally developing a hypothesis and then running multiple experiments to test against it, starting with the notion that it may take forever or never at all. Learning, documenting and ultimately opening up to peer review (open source) for others to contribute. What's interesting is whenever an experiment is deemed ‘successful’ it usually goes to the academic community around the world to also test independently to determine its truth.
I see startups through this prism. In order to filter through to something close to your original hypothesis (Internet TV in Africa) you need to wade through failed experiments. A LOT of failed experiments. More than normal people with sound logic would be willing to accept. I learned that during my undergraduate degree in Chemistry. A lot of what we consider the foundations of science came to be via very fortuitous routes.