Something Weird This Way Comes…

Lies, damn lies and accounting…

Years ago there was an old joke about a CEO who needed to hire a new CFO. The interview consisted of a single question. Having narrowed the candidate pool down to three individuals the CEO decided to conduct the final interviews in person. So he asked the first candidate, a highly qualified individual, the following question…”What is 2 + 2?” To which the candidate replied, “Well, that is obvious, the answer is 4.” To which the CEO said, “Thank you for your time, we’ll be in touch.” The next candidate, a decorated MBA from an Ivy League school was given the same question. Obviously more crafty than the first, this candidate detected a hidden agenda and so couched her answer with a number of caveats but the answer was roughly the same “All other things notwithstanding”, she said, “the answer should be 4”. To which the CEO said, “Thank you for your time, we’ll be in touch.” The last and final candidate, a three year audit associate from a local accounting firm came in and took a seat across from the CEO who explained he had but one question, “What is 2 + 2?”. To which the aspiring accountant said “What do you want it to be?” The CEO looked up a tad incredulous and said, “When can you start?”.

I was reminded of this joke when Amazon coyly took the wraps off some of their Amazon Web Services (AWS) business unit numbers not too long ago. Low and behold, they indicated that the Amazon Web Services business enjoyed an operating margin of 16.9%. During the same period HP announced a consolidated operating margin of 7.1% while publically hemming and hawing over whether it was ramping up, tamping down or more than usually committed to its public cloud business. Back in May of 2014 HP made a major cloud announcement in the form of Helion, their public cloud platform. However, ever since then, HP has been slowly backing away from their declarations of cloud commitment and dominance. But one has to wonder that even in the face of a difficult internal transformation, if HP could more than double operating margins from a single line of business, why wouldn’t HP be all in and all over public cloud offerings?

Well, there’s got to be more than one reason.

For starters, the industry has been all over Amazon for the way that it accounts for its various businesses with a particular interest in the way in which it accounts for capital leases. As you can imagine public web service businesses consume enormous amounts of capital in the form of server farms and all the gear that goes with it. For years Amazon has been using capital leases to acquire and deploy technology. Under FAS 13 the cost of capital leases are not reflected as an operational rental expense but rather as a general corporate interest and obligation expense, and therefore need not be directly associated with any particular line of business but rather as an overall general corporate expense. Amazon has employed this tactic for years to help paint a positive cash flow story in the face of recurring net loss when it comes to profitability.

Amazon also indicated that it’s acquiring capital leases at a rate of around $4-5B per year. And one can only imagine that an enormous portion of those leases are related to the AWS business and that those leases are relatively short lived and quickly amortized or rapidly expensed. Analysts estimate that last year Amazon paid about $1.3B in capital lease payments but it is reasonable to assume that due to the AWS business these outlays will only grow as waves of additional capacity are added in advance of anticipated growth. And even though the growth of AWS has been impressive at 49% in 1Q2015, the growth in operating margins at 8% is not commensurate which would indicate that the business is not scaling efficiently. Or at least not yet. So even if AWS produced $272M in 1Q operating margin, if you added back in direct capital lease expenditures, the chances are you are looking at a boatload of red ink. Further, if market growth slows as offerings mature, and it usually does, or if capture rates decline as competition increases, as they usually do, profitability will be increasingly challenged, as it probably will.

Next, we have already seen the first and arguably only wave of cloud software vendor consolidation as emerging openstack vendors have all been bought or gone bust. Once considered one of the most vibrant sectors of enterprise software market this segment has seemly vanished and it happened all in the last six months, virtually, no pun intended, overnight. (See Here we go again…April 2013) Most of the acquisitions were by large players filling out their portfolios including HP buying Eucalyptus, Cisoco buying Metacloud and PistonCloud, IBM buying BlueBox, EMC buying CloudScaling and VirtuStream. For an entire software segment to disappear would suggest that the VC’s woke up sometime in 4Q2014 and decided there was no future in openstack and it caused a stampede for the exits. But one thing is certain, going forward, openstack will be less about open source and more about vendor visions or maybe that would be vendor apparitions.

The last item to consider is the state of independent cloud integration services, something of an oxymoron when you consider that all the big players already have service divisions that urgently need new revenue. So, that would make all independent cloud integration service providers dependent on those PaaS and Iaas players that lack service divisions which would narrow things down to Amazon, Google and a very narrow group of other minor players. A market so defined cannot sustain the number of players that already exist and VC’s don’t have the patience to stay with service plays as the big players with established service groups aren’t interested in buying them.

So, where does that leave us?

Client Server as a computing model came into vogue in the early ‘90’s and had a great 10 year run until the web came along and started the migration back to hosted service based models of computing. The cloud, that was conceived in the early 2000’s and first became commercially available in 2006, is now roughly seven years old, let’s see, in Client Server years that’s 70 years old. (See Forecast, partly cloudy…July 2010) So while market researchers insist that we are only approaching the knee of the curve when it comes to growth in public cloud computing that would imply that the market for public clouds have at least a 20 year life cycle. Judging by what is currently going on in the industry, that just doesn’t appear to be the case.

None of this is lost on HP or for that matter, Microsoft, Oracle, Cisco or IBM. However, when it comes to Amazon and Google, public cloud offerings could still be considered more of a cost dilution rather than a revenue contribution play. As such, declaring discrete investments in and profits from an emerging business opportunity is not much of a priority. That’s not really the case for established players who have legacy revenue streams at risk. (See Larry, you evil genius … October 2014)

Strategically speaking, it is one thing if you decide to burn your own boats, limiting strategic options can eliminate distractions and help concentrate efforts and investments, but quite another to have some idiot come along and burn them for you. We saw this with the first wave of open source plays, from operating systems to data bases to applications, for established players open source was a threat that had to be addressed. The public cloud is no different. We are certain to get another round of AWS numbers in about a month which will likely not shed any additional light on whether the economics of this are genuinely attractive or its just more lipstick applied to some barnyard animal.

Which brings us back to why suddenly HP has decided to become the Hamlet of the IT industry doubtfully dithering over whether its commitment to the public cloud is a foregone conclusion, debatable strategic option or merely some interloper hiding behind a curtain and worthy of being dispatched.

On the one hand, public clouds appear to be an inevitable feature of the computing industry. However, increasingly it seems that public clouds are only one of many options and not nearly as enticing as some have augured, more a footnote than a major phase in the computing landscape and now would be a good time for established players, both suppliers and buyers, to look beyond this option and begin evaluating what comes next.