By David Gross
Read tidbits of the Goldman note on F5 (FFIV) this morning where they downgraded the stock to "sell", but very little of it made any sense.
On top line growth, the analyst said "85% of the company's growth comes from server refresh and share gains, rather than cloud build-outs".
First problem is that there is no such thing as "cloud build-out". Cloud is a buzzword that means different things to different people, and until someone like Rackspace (RAX) or Terremark (TMRK) actually starts getting a third of their revenue from a cloud services, I won't be able to say it with a straight face. The problem with using it like Goldman does is that it distorts what customers actually do with load balancers. F5 customers are generally thinking about issues like DNS Round Robin, content caching, and cookie-based switching, not "clouds". Even in the buzzword-crazy 90s, we didn't say cloud networking, we called it Frame Relay. At least call it shared computing and narrow down what you're actually referring to, and don't even begin to think there is such a thing as a "cloud" stock.
"Server refresh" is another meaningless catch phrase. Moreover, there has been little correlation historically between revenue growth in servers and revenue growth in load balancers. F5 had decelerating growth in 2007 before the recession, and when servers were flying off the shelves at IBM and Dell. Might make sense logically, but there is no data to support a link between server shipments and load balancer, excuse me, application delivery controller, shipments. Perhaps things would be clearer if vendors and analysts alike stopped using so many ridiculous product names and buzzwords that have little connection to actual product uses.
Regarding share gains, F5's biggest came between 2001 and 2005 when Foundry decided to emphasize Ethernet switches, and let its formerly market-leading ServerIron get crushed by F5's BIG-IP. Cisco (CSCO) has never been dominant in this sector, and I don't think they're looking at that Arrowpoint deal so fondly anymore. Still, unlike most other markets where they're not #1 or #2, Cisco won't admit defeat in layer 4-7 products, which extends to the TCP/WAN Acceleration market where their WAAS module has trailed Riverbed (RVBD) badly for over three years. In F5's case, share gain has been far less of a contributor to growth than it was years ago.
I can understand someone suggesting that F5 is overvalued at 7x its revenue run rate. But its growth has never been linear, has never been tied to "server refreshes", is not tied to any catchy buzzwords, and the company still owes its market leading position today to misguided decisions Foundry made nearly ten years ago.