I agree. Putting your entire stock online can take a lot of manpower, depending on who you are. (Best Buy? Probably fairly trivial. Grandma's and Grandpa's Little Shop of Flowers? Gotta hire more grandkids.)
Unless the time and money spent on the effort is going to pay off, and pay off quick, a store isn't going to bother.
Why just not parse/index all the products from stores websites and verify in-stock real-time when making a search via script or something (given that in-stock availability is indicated). Using this approach you can collect product/location data from more than a million stores I would guess. No?
Real-time inventory tracking for as large a company as Best Buy is "fairly trivial"?
For large companies, it is a multi-million dollar investment to build a warehouse management system capable of showing near real-time levels of inventory. Besides the capital costs, companies can have very specific workflows for managing inventory levels. Also things like drop-ship vendors and sourcing items from multiple sources can make it extremely difficult to calculate or predict inventory levels prior to a sale.
Unless the time and money spent on the effort is going to pay off, and pay off quick, a store isn't going to bother.