That’s a formula for disaster if you want to reap the benefits of a web company…
There is nothing special about web companies. You actually mean companies with a distinctive brand which have to be handled carefully, a good example might be how IBM handled the purchase and integration of Lotus which it kept as a separate division for over ten years and only recently dropped the brand. Oracle's design to fold Sun in quickly was equally as successful in disposing of loss-making sectors and improving the profitability of the rest and providing a coherent product line for customers.
For all its blurb Microsoft has hardly kept its mitts off Skype which is why the client's are increasingly welded to other Microsoft services and increasingly unwieldy as a result. Do we have any figures from Microsoft as to how well Skype is doing? Is it still operating at a loss?
What does seem to be different for web companies is the size of goodwill and potential write-offs. If Facebook can pay $1 bn for Instagram, why shouldn't Tumblr be worth the same notional figure? The justification is usually about the eyeballs for advertising but this is as much about shutting down the competition as anything else. This is less about market share than outright domination: IBM got out of the PC business when it realised the margins were terminally thin; Akamai bought Cotendo to shut them down. And if things don't work out, well the charges can be offset against any tax meaning that other taxpayers will actually take the hit.
Given that current monetary policy is making any of these purchases extremely cheap, we can expect more of them and that is probably one of the aims of QE and its ilk: raising valuations supposedly makes us all feel wealthier and we thus go out and spend more.