Any kind of antenna like that is creating an increase in the flat plate area of the aircraft and thus increasing the parasite drag on the airplane. Fortunately newer antenna designs, such as the KU/KA band satellite internet systems can be arranged in a manner which greatly reduces the frontal area of the antenna.
The ideal solution, which will inevitably come about on future airliners as these systems increase in popularity with the travelling public, is the creation of an integral compartment on the dorsal side of the fuselage aft of the rear pressure bulkhead and fitted with a dielectric which is flush with the aerodynamic envelope, thereby eliminating parasite and interference drag as well as affecting airflow over the vertical fin. However for earlier aircraft or those currently in service, this is not going to be possible for several reasons.
First, an installation like that would require major modifications to aircraft structure and pressure vessel of the fuselage and cost big, big bucks in order to get an STC or approval from the airline's engineering department. Current designs are arranged to as they are to be minimally invasive; all that's required for a typical installation is the attachment of a series of brackets and longerons to the cabin frames which can then be attached to brackets outside of the aerodynamic envelope with a minimum series of fasteners penetrating the fuselage. This arrangement is preferred by the airlines as their engineering departments and the FAA or other regulatory authority can sign off on the installs without requiring the OEM's involvement.
Secondly, many jets are not owned by the airlines themselves but are leased from financial brokerage firms and any kind of system which minimizes alterations to the airframe is preferable to major - and irreversible - alterations to the aircraft.
As for the economics of these systems, even if you did have an increase in fuel burn by 0.1% you could easily compensate for the loss if installed on a large enough aircraft. Let's assume on a 747-400 flight from LAX to Narita Intl in Tokyo, you will typically burn about 290,000 lbs or about 44,000 gal of JET-A. A 0.1% increase results in a mere 44 gal difference in fuel consumption at about $220.00 if you price the fuel at 5 USD/gal. For a long haul fight like that, lets say that 25% of the passengers use this service at a cost of 44.00 USD each. With a typ. 747 seating of 440 people, that's 100 passengers buying WiFi service totaling 4,400 USD in income for the carrier. It's well worth it for them!