THIERRY ROGE
That gleeful background noise, as Hewlett-Packard presented its quarterly results on Wednesday, was Larry Ellison giggling. Last autumn, the Oracle boss got into a nasty public back-and-forth with Mike Lynch, then head of Autonomy, about whether or not Lynch had shopped his company to Oracle, only to be told the price was too rich.
Hewlett-Packard ended up paying $10.3-billion for the company. This week, HP said that Autonomy stumbled badly last quarter, and that Lynch was being replaced as its leader.
HP chief executive Meg Whitman thinks the problem at Autonomy is not the product or the competition. It is lack of grown-up management. "This is a classic entrepreneurial company [with]scaling challenges …" Regardless, the overpayment debate looks to be over.
Revenues at HP's total software division rose $173-million in the most recent quarter versus the prior year, when Autonomy was not in the fold. Even on the unlikely assumption that all of that growth came from Autonomy, the annual revenue run rate at Autonomy is now about $700-million, well short of the $1-billion the company was producing before it was bought. Some part of the decline has to do with changes in the way Autonomy's licence revenues are booked, but the upshot remains. HP paid even more than the dizzying 10 times revenues that was originally thought. Worse, HP software's pre-tax profits expanded by only $14-million in the quarter. Something is going badly wrong here.
What, though, is a company in HP's situation supposed to do? Its biggest businesses are suffering either secular decline or margin compression. It needs to cut costs, and it is doing so. And it needs to take a bet or two on adding growth businesses. One can argue that $10.3-billion is simply too much to gamble. But the hard fact is that bets, by definition, sometimes go wrong.