The Wired headline in April of this year read, “Amazon Reveals Just How Huge the Cloud Is for Its Business.” The numbers for AWS were $4.6B for 2014, up 49% from the year before and on track to hit $6.23B by year’s end. The TechCrunch headline from October was “Amazon’s AWS Is Now A $7.3B Business As It Passes 1M Active Enterprise Customers.” Revenue at $7.3B, not $6.23B. A growth rate no longer of 49%, but 81%.
It is the velocity and trajectory of this business that has everyone in the industry spooked and valuations of the business formerly relegated to the “other” revenue category on financial statements accelerating. Even after seeing sales shrink for 14 consecutive quarters, after all, and amidst calls to rebrand the company from Big Blue to Medium Blue, each of IBM’s non-finance business units generated more revenue in 2014 than AWS projects to this year. Three out of the four were a multiple of the seven billion figure: GTS was ~$37B, Software $25B and GBS came in at ~$18B.
But the market and evaluators alike are less concerned, at least in the case of Amazon and IBM, with present day revenue figures than how they project to change over time, hence the euphoric AWS headlines and the quarterly pillorying IBM receives. What IBM is going through at present, in fact, suggests that Michael Dell’s original decision to take his firm private was a wise one.
Market disruption is a violent process, and surviving it can be almost as drawn out and painful as succumbing to it. As IBM knows, of course, having been one of the few companies to reinvent itself more than once. Expecting the same patience from investors, however, is a lot to ask, particularly in an age of activist shareholders carrying Damoclean swords.
If bullish perceptions of cloud native players, Amazon and otherwise, are driven by expectations of future returns driven by current models, however, it is perhaps worth taking a step back and evaluating IBM’s current models rather than current returns. The question is how should IBM, or companies in IBM’s position, respond to the macro-market factors currently disrupting its businesses.
From a high level, all of the incumbent systems players – from Cisco to Dell/EMC to HP to IBM to Oracle – need to recognize, among other market dynamics, the following:
- Between the ascendance of ODMs and the explosion of IaaS, the market for premium low end hardware is gone. What hardware growth there is will come from the cloud – just ask Amazon, Google or Microsoft.
- Traditional perpetual license software models are not gone, but in systemic decline. Customers instead are shifting to services-based models, with additional value adds from data (both collected and sourced).
- Open source and commodity services have offered customers some relief from lock-in, but it remains as closely tied to profit as Shapiro and Varian described in 1999. This implies that while it’s important to offer commodity entrypoints, higher-end proprietary services will be critical to both profit and retention.
- New market conditions require new partners.
Measured by this criteria, at least, IBM is making logical adjustments to its businesses.
- Low-end hardware businesses have been divested, and investments redirected to potential growth businesses such as Softlayer.
- An increasing emphasis within its software business is on services, e.g. Bluemix, acquisitions like Cloudant/Compose/etc, or the just announced Spark-as-a-Service.
- Proprietary or exclusive offerings such as Watson or the Twitter and Weather Company partnerships offer IBM the ability to upsell customers to higher margin, more difficult to replicate externally services.
- IBM’s partnership with Apple gives them a premium mobile hardware story, and Box CEO Aaron Levie was prominently on display at Insight.
From a directional standpoint, IBM appears to be responding to the systemic disruption across its footprint with a combination of internal innovation (Watson), open source (Cloud Foundry, Node, OpenStack) and inorganic acquisition (Bluebox, Cloudant, Softlayer, etc). Betting on cloud over traditional hardware, or SaaS rather than shrink-wrapped software may not seem aggressive to independent industry observers for whom the writing has been on the wall since halfway through the last decade, but the larger the business the more difficult it is to turn.
Much of IBM’s ability to reverse its recent downward financial trend, then, depends on its ability to execute in the emerging categories on which it has placed its bets. Some adjustments are clearly necessary. A heavy majority of the airtime at its Insight show this week, for example, has been devoted to its Watson product. While the artificial intelligence-like offering is intriguing and differentiated, however, as a business tool it’s a major marketing challenge. Positioning compute instances or databases offered as a service is a simple exercise. Explaining to audiences what “cognitive computing” means is non-trivial. Not least because unlike cloud, IBM is trying to push that rock up a hill by itself. Strangely, however, the company seems intent on leading with the most difficult to market product, rather than using more widely understood cloud or SaaS businesses as an on ramp and using Watson as a secondary differentiator. It would be as if AWS led with Machine Learning and mentioned, after the fact, that EC2 and RDS were available as well.
That being said, marketing and positioning is a solvable problem if the strategic direction is correct. And 14 quarters of declining revenue or no – remember that as AWS itself demonstrates, revenue is a lagging indicator – IBM is in fact making changes to its strategic direction. The company just makes it harder than it needs to be to see that at times.
Whether they can execute on these new directions, however, is what will determine whether the company’s turnaround is successful.
Disclosure: Amazon, Cisco, Dell, HP, IBM, and Oracle are RedMonk customers. Google and Microsoft are not current customers.
October 27, 2015 at 3:31 pm
1. You never mentioned security which is one of the areas where IBM seeks to differentiate, and of course a key reason why lots of valuable things have not and will not go Cloud in the first place.
2. An analyst with an inquisitive mind should really dig a bit deeper at claims. How successful is the Cloud model for the provider in general and Amazon specifically? It is best not to stand slackjawed and take these things at face value because the company says so. That is especially the case when they make up their own metrics.
“OP: On the other hand, AMZN generated almost $10 billion in operating cash flow over the trailing 12 months, up 72% from the previous 12 month period and free cash flow was $5.4 billion up from $1.1 billion for the previous 12 months.
MF: Unfortunately, this is not true. You are using Amazon’s reported metrics which they call “operating cash flow” and “free cash flow”. But they made up those metrics and their names, and they don’t mean what the names suggest. Since they call “free cash flow” is really something else, let’s just “Metric W” to avoid confusion. So yes, they had $5.4bn in Metric W in the last 12 months, which was a big increase on the previous year’s Metric W.
But it’s not what anyone sane would consider to be free cash flow if they read the details. Amazon is egregiously, specifically and purposefully both structuring their business operations and their reporting to hide costs, in order to misleadingly flatter their Metric W (what they erroneously dub “free cash flow”), which is not coincidentally the metric they say they should be valued on. Let’s not forget that this is a late 1990s .com firm, after all : ) Pay no attention to the man behind the curtain, they’re saying, look at the flash and smoke. The things behind the curtain that you really have to pay attention to are things like owner earnings, profits, pretax profits, EBIT, the real free cash flow, and the real operating cash flow.
They’re oddly up front about their smoke and flash though. Note the line in the 10Q that says “Operating cash flows and free cash flows can be volatile and are sensitive to many factors, including changes in working capital, the timing and magnitude of capital expenditures, including our decision to finance property and equipment under capital leases and other finance lease arrangements, and our net income (loss)… The increase in operating cash flow for the trailing twelve months ended June 30, 2015, compared to the comparable prior year period, was primarily due to the increase in noncash charges to net income (loss), including depreciation, amortization, and stock-based compensation, and an increase in working capital.” This is where they tell you that Metric W (what they dub their own non-GAAP metric of “free cash flow”) is not actually free cash flow.
Metric W is very specifically designed to bamboozle naive investors. It does not bear any meaningful relationship to pretax earnings from operations, for example.
– First and foremost, Metric W does not include the cost of capitalized leases. Imagine you are a little firm that rents out tools. To a close approximation, your real free cash flow is your gross rental income, minus your shop rent, staff costs, and current period depreciation on the tools you own which have to be replaced every three years on average. Now, say you instead use a capitalized lease for all your tools (as Amazon has done for the vast bulk of its AWS gear). The Metric W does not include those capitalized lease costs…not the full amount, and not even the current portion. So, looking at Metric W and thinking it means free cash flow is like assuming that the free cash flow from your tool rental business is the revenue minus the staff and shop costs, period. In effect, it’s assuming the tools are free of any cost to you, ever. So, when you adjust for this wheeze and account for the current portion of the capitalized lease expenses, the Amazon TTM Metric W drops from $5.4bn to $637m. Figure from here at the Fool (I get the same figure from an article at FT.com)
– Second, Metric W includes increases in negative working capital. As Dell would once have told you, negative working capital is a great thing to have, and when it goes up that’s even better, but it’s a kind of debt. Cost-free debt, but debt nonetheless. Rising debt in a given accounting period is not in any way a component of earnings, real operating cash flow, or real free cash flow. I don’t happen to have the TTM figure for that, but quoting from an article on the subject it increased by $2.997bn in the 12 months to June. Figure from here.
– Third, it excludes the costs of stock based compensation. They account for this as $1.755bn TTM, but the actual current market value of the shares issued in the last 12 months is $3.195bn. This is a compensation expense, just like salaries. Either way it comes from the wealth of shareholders. If they didn’t give people shares and options, they would have to pay them more in cash. But if they paid them in cash, it would reduce the reported value of Metric W, so they don’t.
– Fourth…there is no fourth. Metric W does not allow for the cost of growth capex, nor should it. That’s a perfectly reasonable thing to leave out when you want a metric of how well the firm is doing on what it’s doing now.
So, in a nutshell, Amazon as an operating business is burning through cash at a furious rate, and this is accelerating. It has nothing to do with “investing in growth”, it has everything to do with the fact that what they’re doing is having a complete lack of any pretax profit on the activities they are currently engaged in, when you count all of the current-period costs of those same activities. Things like AWS needing to own computers which depreciate, and having to pay people to work for you.”
Tech is great and all, but the cloud economics have yet to be proven
October 27, 2015 at 3:57 pm
In case you find the above interesting, just google for Amazon capitalized leases or read this
October 27, 2015 at 6:08 pm
Couple of thoughts:
1. I didn’t mention security because this is a post on general purpose rather than specialized infrastructure. There are many other IBM product categories left out of this piece; too many to list, in fact.
IBM is indeed attempting to differentiate on a security basis, but even outsized success in that effort would not be nearly sufficient to turnaround the financial fortunes of a $90B+ revenue business. IBM was a trusted supplier of basic infrastructure for the pre-cloud world; to be successful over the longer term, it needs to be for the post-cloud buyers as well.
2. Your cross-posted comment was in response to a specific metric – “free cash flow” – that is not mentioned in this piece. The only number included in the piece was revenue because this piece was about IBM’s response to changing market conditions, not the specific balance sheet implications of cloud businesses. The revenue number, in this case, appears highly unlikely to be “made up,” to use your term. The only reason that the numbers were included, in fact, was to help explain market perceptions of trajectory.
You have had strong opinions about valuations in the past, but as I have argued the market does not have to share them (https://twitter.com/sogrady/status/458626906779553792). The market is not obligated to be a rational actor.
3. If the motivation for the comment in general was to imply that cloud in general is permanently a losing proposition, I would submit that the entirety of market evidence to date suggests otherwise. Many cloud businesses, Amazon included, will operate near or at a loss in particular areas as a means of strategic investment – betting that the cost of acquiring a customer now will be lower than in future – but one way or another the impact of the growth of the cloud is difficult to miss.
Whether or not the businesses that emerge as winners in a post-cloud world will replicate the margins or relative rather than absolute revenue numbers of the entities that preceded them has yet to be determined. But the demand will be met one way or another, the slackness of my jaw notwithstanding.
October 27, 2015 at 6:17 pm
Fair enough on security. My point is only that that may be IBM’s “last castle”, their fins svcs and infrastructure customers won’t go cloud at this point and the reason is security, so is that good enough? I don’t know, but its something
I am sure that there are plenty of reasons that IBM should worry about Amazon, and you can make any number of bullish statements on Amazon, just saying that the pro Amazon arguments are tech or platform based and not backed by financials.
Absolutely, in the long run it could prove that the tech/platform differentiators are such that the financials come around, its just not the case today.
BTW, were you aware of the capitalized lease issue? I was not previously, its disquieting.
The reasons I mentioned FCF is because 1) its probably the most important metric 2) Amazon basically has no earnings so its either revenue (who cares?) or cash flow and 3) Amazon themselves want to be judged on cash flow (as they should) except they are playing some games here.
When I encounter new information like capitalizing leases I revisit my previously held opinions.
As to not relying on market prices, my opinion is shared by Ben Graham, Warren Buffett and plenty of others, feel free to disagree but I am happy to let Mr. Market serve me, not guide me ;-P
October 27, 2015 at 6:49 pm
Yes, the capitalized lease issue was raised by Charles Fitzgerald among others back in February (http://www.platformonomics.com/2015/02/aws-and-the-bonfire-of-amazons-vanities/).
As for pro-AMZN arguments not being backed by financials, what you mean is the underlying financials. You can argue about how Amazon is reporting its costs, but the fact is that in an era in which traditional legacy providers are reporting flat or negative revenue growth, Amazon’s revenue stream is growing. Fast.
Dismissing this metric as you do above (“who cares?”), I believe, ignores both the market context as well as the underlying dynamics of the cloud business which are quite distinct from traditional enterprise software markets – particularly in terms of retention.
IBM and every other traditional supplier, in other words, is worrying about Amazon not just because its managing to grow revenue where other suppliers have failed or that it is introducing directly competitive products at an accelerating rate, but because customers that go into Amazon are unlikely to come out.
October 27, 2015 at 7:03 pm
So cine you already knew about them, what do you think of capitalized leases as they are reported by Amazon? Does that change your view of the sustainability of their cash flows?
As to dismissing revene, absolutely. AS Ginni Rometty said revenue without profit its just empty calories. You and I can rent a garage this weekend and build a $1B of revenue if we want to sell $100 bills for $90.
You’re old enough to remember the dotcom days. I am not comparing Amazon to pets.com or something, but I am saying you cannot draw a straight line from quality financials today to even better financials tomorrow. You may see better or even great financials tomorrow, but they’re starting with low quality financials today.
October 27, 2015 at 7:25 pm
On the off chance, we’ve already exhausted interest in this topic here is my closing argument courtesy of Richard Armour
Shake and shake
The catsup bottle.
And then a lot’ll.
Right now, the financial metrics say that Amazon is engaged in shaking the ketchup bottle. Reasonable people can disgaree whether or not a lot’ll come later, but so far it hasn’t
On The Ever-Changing Cloud Race And/Or Face. Apprenda Marketwatch - Apprenda says:
October 28, 2015 at 11:09 am
[…] Crossing the Amazon: IBM in an Age of Disruption “…Market disruption is a violent process, and surviving it can be almost as drawn out and painful as succumbing to it. As IBM knows, of course, having been one of the few companies to reinvent itself more than once. Expecting the same patience from investors, however, is a lot to ask, particularly in an age of activist shareholders carrying Damoclean swords. If bullish perceptions of cloud native players, Amazon and otherwise, are driven by expectations of future returns driven by current models, however, it is perhaps worth taking a step back and evaluating IBM’s current models rather than current returns. The question is how should IBM, or companies in IBM’s position, respond to the macro-market factors currently disrupting its businesses…” Via Stephen O’Grady, RedMonk […]
IBM just paid billions for the Weather Company's digital data — here's why - bns24.com says:
October 28, 2015 at 1:20 pm
[…] upsell customers to higher margin, more difficult to replicate externally services,” writes Redmonk analyst Stephen O’Grady in a blog post written earlier this week, before this deal was […]
DawentsIT: IBM just paid billions for the Weather Company’s digital data — here’s why | DAWENTSIT says:
October 30, 2015 at 5:22 pm
[…] upsell customers to higher margin, more difficult to replicate externally services,” writes Redmonk analyst Stephen O’Grady in a blog post written earlier this week, before this deal was finalized.Brendan McDermid / Reuters IBM Chairwoman […]
Tom Brander says:
November 5, 2015 at 5:44 pm
It still amazes me that so many companies are prisoners of a mountain of legacy Cobol code running on IBM mainframes. Specifically Banks and Insurance. It is not that they don’t want to move to the cloud due to security. it is the code! And the fact that most of it is obtained from a limited number of third party (Fysys(big)/Jack Henry (Small/medium)) suppliers who have no interest in shifting platforms and architectures as long as they can just do maintenance and collect monopoly rents for obsolete software from the relatively clueless financial institutions who continue to regard the whole situation as sunk cost. really kind of funny, sad… but it does provide a big cushion to IBM mainframe business.