If it's true as I read here that only 20% of customers download games while the rest still buy the disk, that means only a small fraction of t2/firaxis customers will be downloading civ5 in the first place. Giving steam a monopoly, if that is in fact what is going to happen by default because of steam's business practices, means the civ people will even be missing out on some of those download sales because of customers preferring other download services or not liking steam. If steam has 70% of the download market, as I read here, then that will be 30% that may not buy civ5 because of the steam decision. So of the total market, mathematically that works out to 30% of the 20% which download games and around 6% of the total market that may not buy civ5 because of the steam decision. 6% may not seem like a lot, but for every potential 1 million sales, that's 60k of lost sales. At $60 a shot, that's a loss of $3,600,000.00. for setting up their download sales exclusively with steam. I wouldn't exactly call that chump change.
Another thing is that downloading games and using services like steam is still relatively new. Many people have little or no idea how these things work yet. The majority of those people who buy the disk and then go to sign up and use steam are going to be in for quite an education experience. Signing up appears easy enough, but there is a lot of things you need to know and learn if you are someone new to this sort of thing. And I would guess that a fair proportion of people will run into difficulties. From the civfanatics forum, here are a couple of links to posts describing problems people had using steam:
http://forums.civfanatics.com/showpost.php?p=9180669&postcount=31
http://forums.civfanatics.com/showpost.php?p=9177450&postcount=29
Now these people were probably more experienced using things like steam than the average person buying a disk at a store. Since one has to install steam and run it at least periodically, everyone buying the game will have to learn how to use the service. Someone brand new to this gaming concept, like quite a few will be, are going to have real headaches when these kinds of problems manifest themselves. They wont know what to do, and since they wouldn't be expecting such complications, they are going to be pissed off big time.
Back when civ4 came out, there were all sorts of compatibility problems and some programming bugs. It took them a while to fix that mess. Integrating a 3rd party's software into the civ5 programming will likely increase these usual new release compatibility/bug problems even more. This is much more complicated than just putting the game on steam to sell it like they did with civ4, it's mixing steam and firaxis programming together in the same product. That alone should make for some interesting problems down the line. I see a major train wreak approaching for t2/firaxis with civ5.