Warren Buffett (who actually introduced the concept of "moat" into business analysis) says: "In business, I look for economic castles protected by unbreachable moats."
Many of the moats that protect businesses from their competitors are made of brand power, large scale and/or permanent cost advantages or intellectual property.
With IBM (NYSE:IBM) as one of the largest investments in Berkshire Hathaway's (NYSE:BRK.A) (NYSE:BRK.B) stock portfolio, we can be quite confident that Buffett saw a nice moat around Big Blue's business. However, during the 2013 Berkshire shareholder meeting Buffett said:
I don't understand the moat around IBM as much as around Coca-Cola (NYSE:KO). I have some understanding of it, but would have more conviction about the moat around Coke or Wrigley or Heinz than IBM, but I feel good enough about IBM that I put money in it, and nothing precludes Microsoft (NASDAQ:MSFT) and IBM both being successful. In fact, I hope both are. We have enough conviction about IBM's position. I like their financial position. Odds are good. I don't feel the same degree of conviction about that than BNSF railroad. I can't think of anything going wrong with BNSF. I can think of some things wrong with IBM.
Many investors interpreted this statement as an admission that Buffett was losing confidence. While I don't believe that a few bad quarters would change Buffett's ideas on IBM (remember that it took him several quarters just to build his position), it is certainly important to understand as much as possible about IBM's moat in order to be able to spot in time an eventual deterioration.
So what is IBM's moat made of?
In its 2013 Annual Report IBM describes its business as follows:
The company's business model is built to support two principal goals: helping enterprise clients to become more innovative, efficient and competitive through the application of business insight and IT solutions; and providing long-term value to shareholders.
The stated goal here is "value creation". Let's think about how exactly IBM creates value:
It does not sell a commodity product at a lower price than everybody else (allowing clients to save money), but it provides knowledge, intellectual property and/or technical capabilities superior to those of its competitors. In some cases, IBM basically promises its clients to stay ahead of the game - even if this service comes at a higher cost compared to its competitors. In other cases, IBM may provide a Total Cost of Ownership advantage, which might turn out to become a competitive advantage itself. (If a machine consumes less electricity, the machine owner can sell his product for less and will still pocket the same profit as his competitor whose machines have higher power consumption.)
While "value creation" is a nice promise to make, we know that there are many companies promising something similar: Oracle (NASDAQ:ORCL) and Accenture (NYSE:ACN) are only two of them. So, what is it that makes IBM different? What is it that virtually guarantees IBM will stay ahead of the game? Currently, mostly due to its struggling hardware segment, Big Blue seems to be lagging its competitors.
The following article summarizes nicely Wall Street's current concerns related to IBM's hardware business:
The computer industry has behaved more like […] steel businesses this year, with numerous layoffs and other job reductions. […]
IBM and independent analysts cite several reasons for the troubles, including an economic slowdown and sluggish capital spending by American companies.
But one of the biggest factors is that computermakers have worked their way out of sales. They have made machines so cheap and powerful that customers no longer need to shell out millions of dollars to get a job done. They can get by with spending a tenth or a hundredth as much.
There are some winners in this transformation […] such as Apple Computer Inc. […].
The losers are companies that get most of their profit from bigger, costlier computers, including IBM […].
All those companies have cut jobs this year, through layoffs or voluntary resignations and retirements. Their machines can't match personal computers for the combination of low price and high speed.
Personal computers won't take over entirely. Big mainframe computers remain irreplaceable for jobs such as controlling access to vast reservoirs of data and handling complex computing jobs that can't be split up among several smaller machines. Minicomputers, which lie between mainframes and personal computers, also have a valuable role in many corporations.
But more and more, personal computers are doing jobs that once were the exclusive reserve of bigger machines.
"The little guys are just creeping up everywhere and they're stealing their way into niches all over the place," said Doug Crook, a Dataquest Inc. analyst.
[…] The danger to mainframes and minicomputers was obscured early in the decade by the overall growth in the industry. With growth slowing, the threat from personal computers is clearly delineated.
You may have noticed something strange in this excerpt. Well, it is from an article published in 1989. The death of mainframes was already seen as imminent. But the dinosaurs are still alive.
The next excerpt comes without a "surprise effect." I promise, it is from last year:
The death of the mainframe has been predicted many times over the years. But it has prevailed because it has been overhauled time and again. In the early 1990s, the personal computer revolution took off and IBM, wedded to its big-iron computers, was in deep trouble. To make the mainframe more competitive, its insides were retooled, using low-cost microprocessors as the computing engine.
Like any threatened species that survives, the mainframe evolved. It has been tweaked to master new programming languages, like Java, and new software operating systems, like Linux.
"The mainframe is the most flexible technology platform in computing," said Rodney C. Adkins, IBM's senior vice president for systems and technology.
That flexibility is a byproduct of investment. The new IBM mainframe, according to the company, represents $1 billion in research and development spending over three years.
IBM has also invested beyond its corporate walls. Nearly a decade ago, fearing that its mainframe business would wither if retiring mainframe engineers were not replaced, IBM went out to universities, advocating for mainframe courses and offering support. Today, more than 1,000 schools in 67 countries participate in IBM's academic initiative for mainframe education.
The sale of mainframe computers accounts for only about 4 percent of IBM's revenue these days. Yet the mainframe is a vital asset to IBM because of all the business that flows from it. When all the mainframe-related software, services and storage are included, mainframe technology delivers about 25 percent of IBM's revenue and more than 40 percent of its profits, estimates A. M. Sacconaghi, an analyst at Sanford C. Bernstein. (Source)
On April 7th, mainframe computers celebrated their 50th birthday, as on April 7th, 1964 the System/360 was first launched. In the fast-moving world of computing, they can be called dinosaurs, with the important difference that their extinction still hasn't happened and is unlikely to be witnessed anytime soon. Worldwide installed mainframes elaborate 1.1 million requests every second, while Google (NASDAQ:GOOG), (NASDAQ:GOOGL) processes "only" 59,421 requests per second. Almost all global credit card transactions are processed by Big Blue's "dinosaurs," and about 80% of global corporate data is stored on mainframes or is processed by these highly reliable machines with MTBF (Mean Time Between Failure) periods of over twenty years.
During the last earnings call, IBM's CFO stated that:
Looking at our results this quarter, System z revenue was down 40%, with MIPS down 19% year-to-year. We are in our sixth quarter of the product cycle, and gross profit margin was up year-to-year, consistent with this point in the cycle. System z is a core franchise which provides mission critical infrastructure for our customers and we continue to invest in the platform. It's worth noting that earlier this month we celebrated the mainframe's 50th year.
And during the Q&A session he reinforced the point:
[...] the mainframe revenue growth absolutely reflects where we are in the cycle. There is not a secular challenge in mainframe at all. So the mainframe growth rates are absolutely cyclical and when we […] get to announce the new mainframe that will absolutely turn.
So, putting the snippets together, we get to the following hypothesis:
If 40% of IBM's profits flow from mainframes, and mainframes are currently not challenged at all, as soon as the new mainframe cycle begins, IBM's revenues and profits should zoom.
Seems too easy to be true? It is far from impossible. A recent analysis by Gartner showed that 20% of IT managers asked had a totally wrong perception of mainframe's market position. They answered something like "I read the mainframe is dead in the airline magazine." Another huge portion of IT managers (18%) said mainframes weren't "cool enough." So, if IT managers get it wrong, why in the world should the street get it right?
But are mainframes really not challenged at all?
About one year go, Oracle's CEO, Larry Ellison, announced the first "mainframe" made by Oracle. You can read an unofficial IBM response here. From a technological standpoint, Oracle's server seems to be a standard 32 socket Unix server, and by calling it a "mainframe," Ellison probably wants his listeners to make the equation: "Oracle server = very reliable server." Hewlett-Packard (NYSE:HPQ) has done the same with Superdome, calling it "mainframe class."
Even if Ellison's attack on IBM's moat is doomed to fail (or maybe has never even been a real attack), how safe is this moat? What are its strengths?
While everybody knows about the mainframe platform's extremely high level of quality, many also believe it to be expensive and inflexible. However, most of the time, this is the opinion of people who don't actually use mainframes. Recent studies endorsed by completely independent parties show a very positive cost comparison for the mainframe against other platforms, mainly because it is not the platform that is inflexible, but the procedures built around it over time. The following graph clearly shows that hardware costs are only a small percentage of TCO:
And in fact, cost per MIPS has gone down over time:
So why are procedures so inflexible and costly? Couldn't it be that this is some kind of intrinsic problem of the mainframe structure?
The answer requires to delve deeper into the history of mainframes: In the first years of its existence, IBM had no choice but to invent its own standards. So businesses started to build their procedures on these standards. Over time, IBM's platform evolved, and nowadays it can be perfectly connected to other platforms. But it still supports its old standards, as customers don't want to rebuild all their old and reliable procedures on every system upgrade. Hence, while it is true that software maintenance on mainframes has a higher cost compared to other platforms, this is no intrinsic defect of mainframes. Actually, due to a built-in advantage no other platform can offer, which strongly ties customers to the hardware manufacturer, with mainframes, there actually is something to keep and to maintain. This in contrast to other platforms which would require writing new software for many procedures, and which, being totally untested, would have extremely high risks of failure. The costs and risks associated with such a move are well worth a few extra bucks for the hardware platform.
The stickiness is actually so strong that even other software providers try to capitalize on it, increasing their license fees for mainframe usage, knowing well that customers have no choice but to spend the money.
In addition, the mainframe platform probably is the most reliable piece of technology available worldwide:
- Systems can be added or removed as needed, while applications continue to process uninterrupted.
- Workloads can be cloned and distributed across the cluster, which can be spread up to a distance of 20 km cable length, while the different systems still act as one machine. This is perfect for true "continuously available" applications.
- Disaster recovery and high availability can be further extended through GDPS (Geographically Dispersed Parallel Sysplex) over much greater distances (easily 100 km and more).
- 100% availability with mean time between failure periods of over twenty years. (As monetary downtime costs for large corporations have been increasing strongly over the past years to millions of dollars per hour, it may be worth it to invest in proven reliability.)
- Over 1,000 schools in about 70 countries are part of the IBM Academic Initiative for System z, which means that all over the world, highly-qualified support is assured for a long time to come.
- Redundant internal components and engineering: IBM System z servers execute every instruction twice, in parallel, and compare the results ("lock-stepping").
- No threat of viruses and limited threat of hacking, because of the mainframe's architectural structure.
It is not possible to simulate 20 years of 100% availability in a test laboratory. It is not possible to run procedures written decades ago, which would be too costly to rewrite, on systems other than mainframes.
A proven track record of over half-a-century, reduced threats from hacking and no threats from viruses are certainly worth some extra money.
So there certainly is a market for IBM's System z. Presumably, it's made of very large multinationals or institutions that have as top priorities 100% availability and reliability. These clients tend to be and stay very sticky. Since its clients are very complex organizations, IBM can leverage its value proposition by offering them additional software and technology services (which it has been doing increasingly well over the past cycle).
All in all, I consider it far more likely than not that a new mainframe investment cycle will bring more of the same: increasing hardware revenues, additional services requests, overall increasing margins, growing cash flows, further share buybacks and more than decent EPS growth.
Trading at less than 11 times its operating EPS guidance for 2014 of $18, IBM remains clearly undervalued.
Disclosure: I am long IBM, BRK.B. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.