It took the government to intervene into a very long running and prosperous "free market" to bring an end to discrimination, and nobody was making a economic case for it.
Tim Cook was making the same assumption in the article when he said that workplace equality is good for business.
Also, the "barriers to entry" came from those state and local government regulations I referred to; asking the Federal government to provide a level playing field would have meant asking the Federal government to force state and local governments to remove those barriers to entry.
I am suggesting that when entering the market is already difficult (which is true of many American industries where discrimination might be a problem), a new firm being nondiscriminatory is unlikely to be a sufficient competitive advantage to displace existing firms.
It is possible for nondiscrimination to be beneficial for already existing firms but not an immediate, substantial, and sure-fire competitive advantage.
With a level playing field, I don't think entry into the market is as difficult as you claim for many areas of businesses. The discrimination that the civil rights movement protested against wasn't discrimination about who could build new semiconductor chip fabs: it was about who could ride at the front of the bus, or who could stay at which hotels. In a free market these would be easy businesses to get into; the main barriers to entry are government regulations.
The ENDA isn't about accommodation discrimination (buses, hotels, etc.), though, it's about employment discrimination. Discriminating against potential customers is obviously bad business, and the competitive benefit for not discriminating might be decisive. Even so, government action was effective in almost entirely eliminating that form of discrimination (to the extent that when it occurs, it's seen as a throwback and an outrage).
The barriers to entry for operating a bus service would be enormous even without government regulation. The initial capital outlay is substantial, and established firms would be likely to have lucrative exclusive contracts with businesses. Potential riders would take a long time to trust a new bus service, and it would have to operate at a considerable loss for quite some time to build up a reputation.
After government action had enabled it in the first place. Jim Crow laws didn't just prevent blacks from riding at the front of the bus: it prevented blacks from starting their own bus companies.
The barriers to entry for operating a bus service would be enormous even without government regulation.
If all the bus services in town discriminate against blacks by making them ride at the back of the bus, but there's no government regulation, a new bus service, even if it's just a single bus, that does not impose that restriction has an immediate customer base: all the blacks that don't want to ride at the back of the bus. (We're assuming that this is a substantial number, which seems historically accurate.) Similar remarks apply to any other product or service where discrimination exists. I find it extremely implausible that, in the absence of government regulation, there would not have been many, many entrepreneurs willing to take advantage of such opportunities, or people to fund them.
How would you know?
Purely competitive markets are a mathematical abstraction
So are government regulations that magically apply just the "right" correction for a "market failure".