Startup: "Dear Amazon (or any Fortune 500 company,) you pay $xx in transaction fees annually, install our system to cater to the $xx-sized bitcoin economy, and save about $xx in fees."
Even Square: "2.75%. [But] if you enter credit card numbers manually, Square costs 3.5% + 15ยข per transaction." https://squareup.com/pricing
Can that be beaten with say an android client: https://github.com/bitcoin-labs/bitcoin-mobile-android and a printed qr code?
Philosophizing more, in the larger scheme of things, perhaps Bitcoin can remove the notion of socialized losses from society which has devastated the economy.
But I wonder if there is a role for an affiliate that actually makes the purchases and sends them to people. That is they make the purchase on Amazon on their own account, and have it shipped where you want it shipped. They accept bitcoins from you, pay dollars to Amazon, and trade back bitcoins for money on the exchanges. Their profit margin comes because they are an Amazon affiliate, and put their affiliate code on all of their purchases, so that they get the affiliate fee.
Something like this would not really need much cooperation from Amazon. Heck, you could just have a browser plugin that recognizes Amazon, and gives the user the option to make a bitcoin purchase instead of using your credit card.
add:
the other short-term use case for bitcoin and say Amazon is giving refunds in bitcoin given the reach of bitcoins is low right now. Refunds are a push to market, as opposed to a pull. Typical cc chargebacks can be expensive too, and so refunding in bitcoin avoids it.
Dealing with bitcoin doesn't seem to me to be much more complicated than the network of payment processors people used to use to get money to/from Full Tilt. Sure, the potential market is going to be smaller. But even if it is an order of magnitude smaller, $140 million/year in revenue is nothing to sneeze at.
For instance on April 15 the US government shut down all of the online poker sites that were letting people play poker for money in the USA. If you look at the charges, they pretty much all have to do with wire fraud. (The operator of the site can also be charged, but analyses that I have seen suggest that that charge may be harder to make stick in the case of poker.) If someone is willing to not travel to the USA, and is willing to set up a poker site that does transactions in bitcoin, then all of those financial issues go away. (Or are subsumed in the inevitable ones that bitcoin will face, depending on your point of view.)
There are a lot of issues to getting a poker site to work well. (People will let you charge surprisingly high fees if in return you do a lot of policing to eliminate people running bots and who are engaged in various kinds of cheating.) However there are a lot of online poker players suffering serious withdrawal in the USA. The market is there. And the online poker market in the USA until a month and a half ago was well over an order of magnitude larger than the current bitcoin economy.
So there you have a realistic company which could be built, now, using bitcoin. You can generally count me as a bitcoin skeptic, but if someone can get traction with something like this, bitcoin may indeed have a nice future.