Or they do, but only in a small niche where a progressive manager made it a priority and hasn't yet earned consensus among his or her peers. The same things goes re: NIH at most large companies. It's not that the homegrown tool is better, but if it's the difference between spending $250k on dev and $100k/yr on maintenance and improvements versus $1.5m on licensing, $500k on implementation and change management and another $250k/yr ongoing... those things get a lot of scrutiny. Especially when they're not driving the profit. As a couple of examples, let's look at homegrown HR systems vs things like Workday. Workday is loads better and they are constantly adding new features, but it is always easy to excuse a homegrown HR system because of innumerable edge cases in localized employment law, or ties to tiny payroll systems in third world countries that you've already built but no one else has, or .... Netsuite is having the same adoption challenges vs Oracle & SAP, too. It's not that the big guys are objectively better, necessarily, but that they've been around long enough and have a huge enough consulting arm (and external partners) to have seen almost every possible business scenario globally. There's a lot of value there, and it often leads to short sighted and myopic decisions.

I don't actually know that this is true, but I suspect the vast majority of programmers work for large, non-software-product companies, and most of them spend their time primarily on KTLO work, not any kind of innovation or profit creation.