There's two problems here:
1. The costs they're incurring supporting non-hue compatibility might very well increase. Lighting is something people think they universally understand -- if it doesn't work when they expect it to, they call. In the rest of the "lighting world", people expect the lamp won't have problems working with whatever switch they plug it into. I still hear people confused about CFLs that don't work with their dimmers.
2. The large enough part of the market that they control might be large when that market is limited to "extremely early adopters of this kind of brand-new technology". The market will grow in many stages over the next few years and this is way to early to start giving people another reason to avoid your products.
I'm usually an early adopter of technologies like this, however, I've held off because there were only a handful of implementations out there and the use cases for them was still a "that'd be neat, but not worth the money". I've recently become very interested and will be buying something in the next 6-months, though this is making me reconsider that now.
I know I'm a future customer of this segment of technology -- not just lighting but upgrading my existing home automation system -- and I'm starting with lighting. I'm not, yet, a customer because I haven't found a good mix of "good product", "good interoperability", yet. So they've gained lock-in of all of their early adopters -- a very short-term investment[1] -- and caused the second wave of (in this case, also) early adopters to look elsewhere. They're nowhere close to the market that even high-end, non-smart bulbs command and they won't see much of it this way. People complain today about having to buy special bulbs for their dimmer switches; this is a whole other level -- you can't even use the lamp you just bought. No way someone like my dad is going to be a customer.
I can see how this might improve sales in the very short-term. It also might not. It will increase support calls. It will hurt their ability to get future customers and might even delay the next wave of adoption as people wait for fear of buying something that will become obsolete. These are all the things I'm thinking about as a person who really wants this technology, has the money to buy it, but doesn't have the money to do it twice. And I can live without it for a bit longer.
[1] Short-term because the people who are buying these for the most part have a lot of disposable income or a very small deployment, so switching to something else isn't necessarily as much of a lock-in as they'd like. And this product in particular, the primary consumer is tech folks -- they'll hack together the interoperability eventually, defeating the purpose.