1) Venue relationships 2) Taking the fall on fees 3) Ability to scale
People are often surprised by #3. Having seen many startups try & fail (as well as large entities like LiveNation), ticketing is a hard problem.
In my time at Yahoo & Google, I believe they have the technology to build a ticketing company, but beyond on that, most -- startups to large entities such as LiveNation & CTS -- had big problems handling onsales for large U2-sized events. In fact, it always made me think "Good luck with that" when people would talk about the design of ticketing systems using relational DBs. :)
They could combine their fees into the ticket price, but they're afraid the artist will also demand 90% of the combined ticket price, putting significant risk on the table for them. Their motivation is minimal: People still buy tickets & nobody ever blames the venue or promoter. So from their viewpoint, why change?