Because O'Reilly paid what its value was. It seems like O'Reilly started regretting having Pearson as a partner and just wanted to buy it back so they could do what they really want.
Pearson once had a team called the "O'Reilly Killers" whose aim was to compete in highly lucrative fields and well send a hammer blow to O'Reilly's diminishing revenue per title (dropping since the dotcom bubble). It may have been more of a mindset that an actual ambition to send O'Reilly out of business but that was the team's name (the name was retired when they joined forced according to Tim).
In regard to the birth of SafariBooks, apparently the story goes that the CEO of Packt chatted to Tim O'Reilly about future of books and ebooks and suggested that an online library would be something developers would use in the future. Tim offered to split the cost and risk having thought about the product and how to bring it into life (it represented a significant risk if no one brought the service as development cos was high) but the CEO of Packt didn't like the cost and risk so Tim, forced to find other partners went to his arch-nemesis Pearson.
Pearson having made tons of money out of subscriptions to libraries and institutions instantly saw the value of the proposition and took the risk (at worst, they would be down what they could afford and O'Reilly would be at risk) and that is how apparently the two enemies became uneasy allies.
O'Reilly have not seen their revenue per title average return to pre dotcom burst levels (currently around 80-100K per book on average in revenue) but they have seen great returns on SafariBooks and they get to monetize rival content so this move signals to me that they see it as their future beyond "just" content development.
-- if you are wondering what Packt did, they made their own cheaper version of SafariBooks on their own which only featured their own books called PacktLib.
-- sources: Worked in tech publishing for a few years, had a few chats and was told of some chats as well as http://radar.oreilly.com/2013/09/how-i-failed.html which is a great read from an interesting man
I've found quite a few Packt books on SBO. I'm kind of surprised they are publishing titles there since they have a competing service.
You can even buy Packt ebooks from the O'Reilly site now which shows the extent that Packt benefits from selling its content through O'Reilly (and indicates how much it lost out on by not being the other half of SafariBooks)
This is a very interesting figure. With printing costs and operations, there won't be a lot of profit in there. I imagine each title has a staff editor, copy editor, marketing team, etc.
The numbers seem to imply that only $100k / $20 = 5000 copies are sold per title, which is an underestimate, I imagine, because O'Reilly books are generally pretty good---but I guess crash and all...
Assumptions: $30 list price so say $20 revenue from each sale of which $5 went for printing and another $3 for logistics, there will be like $12 of profit left. After the $1-2 royalty to the author, we've got $10 profit per sale. It's a 33%-margin business. My guess is with eBooks the margin can go up to 66%, which is much more reasonable. No wonder why everyone is saying eBooks are the future ;)