15.04.2013 Blog 3 Comments on Samsung versus Intel continued

Samsung versus Intel continued

Posted by Marc Brien, VP Research, DOMICITY LTD.

Domicity consults on the strategies & operations of I.T. companies and market trends, and advises government on economic development and investment attraction opportunities.

Last week’s post presented information on which to base an analysis of the Samsung versus Intel struggle. This week Domicity examines these differences and some significant implications for the broader IT industry.

Samsung versus Intel — structural differences

Merchant market component sales account for just 35% of Samsung Electronics’ total revenue. This contrasts with Intel, which generates virtually all its revenue from merchant market component and related software sales.

Intel currently gets 93% of its revenue ($49.4-billion in 2012) from microprocessors and other logic devices. By comparison, Samsung Electronics generates an estimated 18% of its merchant market component revenue from logic devices — some $11-billion in 2012. It is Samsung Electronic’s market-leading memory chip and flat panel display businesses that push its total merchant market component revenue past Intel. (Another significant component business, producing lithium ion batteries, is run by a sister company to Samsung Electronics — Samsung SDI — which claims market leadership with FY2012 Lithium Ion battery revenue approaching $3-billion.)

Samsung has succeeded in developing a $30-billion business in flat panel displays for the merchant market, with a comparable volume supplied internally to Samsung divisions. This demonstrates how the company’s downstream product businesses help shape, and are in turn shaped by, its component operations. Strength in flat panel displays helps power Samsung’s competitiveness in televisions, computer monitors, mobile phones, tablets and notebook PCs. The company has further strengthened its position with the purchase of 3% of Sharp, which is a leader in production of very large — over 60-inch panels. Meanwhile, Intel has chosen not to turn its attention to the flat panel display market or to lithium ion batteries.

Samsung’s market-leading 35% share of the global memory semiconductor market helped deliver an estimated $19-billion in revenue in 2012. By comparison, Intel’s NAND flash memory business only generated about $1.5-billion.

There’s more profit in logic

Samsung is particularly intent on taking processor market share from Intel to boost both the revenue growth and profitability of its merchant market component business.

Although smaller, Intel’s logic-centered business has been able to closely track the revenue growth of Samsung’s merchant market sales over the past five years. Samsung needs to take significant Intel market share to change this dynamic. An even bigger attraction for Samsung is undoubtedly the much greater profitability of Intel’s business. Intel ’s 2012 operating margin was 27.4% (on operating income of $14.6-billion) compared to a less impressive 11.4% for Samsung’s merchant market component operations (on operating income of $6.8-billion).

Samsung vs Intel: Exynos versus Atom processorsSeveral key market dynamics are at the heart of the battle between Samsung and Intel for leadership in logic devices:

  • The rise of the ARM architecture — Low-power ARM processors from multiple suppliers are becoming increasingly pervasive, to the point where an ARM executive recently suggested Intel should abandon its x86 line and put its massive production capacity behind the ARM architecture. Despite much investment in its Atom low-energy architecture, Intel holds less than 1% of the market for mobile processors. ARM processors from Samsung, TI, Qualcomm, nVidia and others rule the smartphone and tablet market, as well as in hard drives, major swathes of consumer electronics, and other areas. And ARM is threatening to expand into PCs and servers. Intel will not abandon its Atom architecture and transfer its loyalties to ARM anytime soon. But Intel has to worry about Samsung’s use of ARM as a vehicle for growing its client-side and server-side processor businesses. Samsung has been boosting its US processor R&D investment and hiring engineers away from struggling AMD. And Samsung is gearing up for a deep dive into 64-bit ARM-based server processors, a likely prelude to a plunge into server, storage and networking systems for data center operators.
  • Production technology — Over the years, Intel has cited a consistent lead in semiconductor process technology as a critical weapon in its battle to maintain processor market leadership. The market is transitioning to FinFET, a new generation of transistor technology, designed to increase performance for very fine linewidths. Intel was first to commercialize FinFET-based technology in the third (Ivy Bridge) generation of its x86 processors. However, at the end of 2012, Samsung announced a 14-nm FinFET process for making advanced ARM processors for itself and ARM foundry customers, such as Qualcomm.
  • Comparative R&D budgets — R&D may be Intel’s greatest strength. The comparative income statement above highlights why Samsung will not have an easy time displacing Intel as the processor market share leader. Intel spends as much money on R&D to support its logic-centered semiconductor business as Samsung Electronics does for all nine of its component and downstream product businesses.
  • Apple — Samsung is the dominant producer of ARM-based application processors used in arch-rival Apple’s smartphones and tablets. This represents a key part of the estimated $8-billion in components that Apple purchased from Samsung Electronics in 2012. The increasingly intense competition between the two in smartphones and tablets is inducing Apple to wean itself off Samsung-made processors (as well as the large amount of semiconductor memory and flat panel displays supplied by Samsung to Cupertino). Samsung provides some 80% of Apple’s ARM-based logic and Apple is said to want to reduce this dependency to 20% by 2017. There was speculation that Apple would transfer most of its ARM foundry business to Intel. But instead, it appears that Apple has chosen TSMC as its Samsung replacement. If Apple’s ARM business had gone to Intel, it would have made it virtually impossible for Samsung to make up significant logic market share against Intel.

Expect the Samsung versus Intel battle to heat up

Intel is a superb competitor and has proven it knows how to hang onto market share, generating handsome profits in the process. On the other hand, Samsung poses the greatest-ever threat to Intel’s market position. This threat is compounded by the rise of ARM and the opening it provides to Samsung to take client-side and server-side processor market share.

In 2012, Intel saw total revenue shrink by 1.2%. Operating income declined 16.6%. Over the same period, Samsung’s component revenue grew 4.2% and operating income rose 18.2%. The decline in PC demand is negatively affecting both companies, but Samsung has been more successful at catching the transition to smartphones and tablets. Intel is moving to catch up with a new mobile strategy, introduced at the 2013 Mobile World Congress (MWC) in late February.

As the broader electronics/IT market transitions to low-power processors, we expect the race for logic sector dominance to go to Samsung within 10 years, unless Intel can make Atom a vastly more popular market solution than it is to date.

10.04.2013 Blog 6 Comments on Samsung versus Intel: the processor wars heat up

Samsung versus Intel: the processor wars heat up

Posted by Marc Brien, VP Research, DOMICITY LTD.

Domicity consults on the strategies & operations of I.T. companies and market trends, and advises government on economic development and investment attraction opportunities.

The Samsung versus Intel prizefight in processors is light on media coverage in comparison to the Samsung versus Apple battle covered in our previous post. Nevertheless, Samsung’s threat to supplant Intel is real and Domicity believes it could have a greater long-term impact on the broader electronics industry.

Samsung Electronics is looking to beat Intel in processors and in the extended logic chip sector. In addition to feeding Samsung’s appetite for new revenue and earnings, greater strength in processors will make it easier to grow global market share in the company’s broad range of downstream products.

Samsung versus Intel is a battle of titans

The overall scale of Samsung Electronics’ components operation stands head and shoulders above any global producer, including Intel.

Samsung’s components business not only generates merchant market revenue from other electronics producers at an annual run rate of approximately $60-billion, it books an additional $60-billion in internal sales to other Samsung operating units. (Note: currency conversions in this post are at the rate of 1150 Won = US$ 1.)

A cornerstone of Samsung’s strategy — borrowed from Japanese leaders — is to use strength in memory, logic, flat panel components, and lithium ion batteries (made by a sister company in the Samsung Group), to power success in all six downstream electronics product operations: mobile phones and tablets; PCs and printers; infrastructure equipment for cellphone carriers; digital cameras; TVs and related visual display products; refrigerators, washing machines and other “digital” household appliances. Says Dr. Stephen Woo, head of Samsung’s logic (System LSI) business, “We believe the right component DNA drives the discovery of what’s possible.”

In addition to driving competitiveness in current businesses, strength in components will be key to Samsung’s move into other market segments such as data center systems or medical electronics.

The sale of components into the merchant market is central to Samsung’s larger strategy. These merchant market sales generate economies of scale and force the company to keep on the leading edge of technology.


Comparison of Samsung and Intel 2012 Income Statements


Samsung versus Intel, Merchant Market Component Revenue

Samsung versus Intel, Merchant Market Component Operating Income

Samsung Component Revenue by Sector

Intel Revenue by Sector

The table and graphs above provide a basis for analyzing the business model differences that underpin the Samsung versus Intel struggle. Part two extends the analysis.

01.04.2013 Blog 4 Comments on Samsung versus Apple: Dueling business models

Samsung versus Apple: Dueling business models

Posted by Marc Brien, VP Research, DOMICITY LTD.

Domicity consults on the strategies & operations of I.T. companies and trends, and advises governments on economic development and investment attraction.

Samsung versus Apple in smartphones, tablets and other cloud access devices is a battle that is increasingly inflaming the imagination of the business press and other industry observers. It is also interesting to watch the changes that the struggle is driving in the Samsung and Apple business models.

The ascent of Samsung Electronics is something we have been analyzing for a long time. Domicity’s report, The Rise of Korea in the Electronics Market, was published back in 1988. Even then, it was clear that Samsung was a breed apart and would become a significant thorn in the side of the global IT and electronics establishment.

After defeating the Japanese leaders in both semiconductors and consumer electronics, an emboldened Samsung is now intent on biting into Apple’s share of the smartphone, tablet and PC markets. Samsung is clearly causing Cupertino problems, as evidenced by a rolling series of Apple lawsuits against the Korean behemoth.

The Apple business model

Apple and Samsung Electronics are vying for the same prize using significantly different business models. Apple’s is a classic virtual integration model, with a high proportion of the inputs for its products sourced from external suppliers. Samsung is currently the most vertically integrated supplier in the electronics market, particularly focused on building price-performance competitiveness through internal development and supply of key component inputs.

Take a look at the table below which compares the Apple and Samsung income statements for the four quarters of calendar 2012 and at the graph which compares each company’s cash from operations over the past five calendar years.

Apple has been able to keep R&D and Cost of Revenue expenses relatively low as a proportion of Revenue by outsourcing most aspects of the hardware piece — component production as well as assembly. Outsourcing hardware provision radically reduces the amount of money that Cupertino must lay out for R&D and capital expenditures.

Margins are kept high by focusing on a few sleekly designed premium-priced products. These are designed to effectively showcase its highly-featured, proprietary software architecture. Demand for Apple products is skilfully stimulated through leading-edge advertising and promotion and a sales strategy that has turned the company’s glossy retail outlets and reseller counters into one of the places to be.

One quibble with the Apple business model … it is high risk.  All of the company’s relative handful of eggs are in the consumer cloud and the cloud access device basket.

Apple’s level of risk is beginning to increase. Its outsized profits and growth are attracting fierce competition from: Google and the Android ecosystem, Microsoft and its Windows partners, and last but far from least, Samsung. The Korean giant has placed a large foot in both the Windows and Android/Chrome camps. It is also promoting its products as the lead mobile platform for the newer Tizen (promoted with Intel) and Ubuntu Linux operating system alternatives.


Samsung vs Apple Income Statement Comparison

Samsung vs Apple Cash from Operations

The Samsung business model

Apple has made focus a virtue. The company gets virtually all its revenue from cloud access devices, plus related peripherals and services.

The Samsung business model is more cautious. Following in the footsteps of a strategy pioneered by NEC, Sony and other Japanese producers, the Korean leader’s strategy features a high level of vertical integration through a strong semiconductor and components business.

The components sector feeds multiple downstream product markets.  Samsung Electronic’ s strength in flat panel displays, plus memory, logic, sensors, LED light sources, lithium ion batteries (through a sister company) drives competitiveness in a range of downstream product units: mobile phones, tablets, and PCs; laser printers and MFPs; infrastructure equipment for cellphone carriers; digital cameras; TVs and related visual display products; plus refrigerators, washing machines and other “digital” household appliances. Strength in components will also enable expansion into new fields such as medical electronics, or even data center hardware.

The upshot is that instead of getting all its revenue from cloud access devices and related services, like Apple, cloud access devices and related markets only account for a bit more than half of Samsung’s total revenue.

Samsung overpowered its Japanese competitors over the past twenty years by outspending them in most areas of the value chain, in particular R&D, capital expenditures, marketing and promotion, sales channels. The company is now leaning on the same formula in its fight with Apple. Even though revenue levels for the two companies were comparable, in calendar 2012 Samsung spent:

  • $10.0-billion on R&D to Apple’s $3.6-billion
  • $29.5-billion on selling and administrative costs (SG&A), which included $11.3-billion for advertising, promotion and other marketing expenses; by comparison Apple’s SG&A spending was $10.3-billion for calendar 2012, with approximately $1.0-billion for advertising
  • $20.0-billion on capital expenditures versus $10.4-billion for Apple

While it is true that Samsung’s spending in these areas is spread across all of its main business lines, it is also true that a growing proportion of resources are being devoted to the universe of cloud access devices and related software, peripherals and services, where Apple plays. As well, some of the capital spending and R&D money that is invested in other business lines, notably the components businesses, strengthens Samsung’s competitiveness in cloud access devices. Investment in broadband mobile phone infrastructure helps Samsung learn which direction cell phone operators want to take wireless networks.

Samsung’s business model comes with a cost. High spending on R&D and factories for its components and a diversified range of downstream products provide long-term stability. But it also means Samsung, at the moment, cannot hope to achieve the same rate of company-wide profitability as Apple, with its tight focus on a narrow set of products. However, if Samsung’s formula works as well as it has in other markets, its heavy investment in R&D, capital expenditures, marketing, and sales channels will steadily sap Apple’s market share and much of its profitability, signs of which are beginning to appear.

Samsung versus Apple: both are having to tune their business models

Like other market players, Samsung is finding Apple’s level of growth and profitability irresistible, causing it to throw some caution to the wind. At the risk of losing a degree of diversification, Samsung is devoting more resources to the cloud access device market and tilting its overall business in that direction.

Samsung’s revenue from smartphones, tablets, PCs, and related peripherals — Apple’s addressable market — grew from roughly 40% of total company-wide revenue in its 2008 December quarter to more than 55% for the same quarter in 2012. Skewing its business model in favor of cloud access devices means that Samsung Electronics is spending even more R&D, capital expenditure and marketing money on related technology, and related supply and demand chain activities.

Competing with Apple is also driving Samsung to invest at accelerating levels in software and cloud-based value-added. One key goal of these efforts is to be able to pull sensor-generated data from all Samsung mobile and consumer electronics products — data that can be sold to various types of business partners. Another goal is to better integrate the universe of Samsung mobile and consumer electronics products, in the manner of Apple’s generally well-integrated product portfolio.

Apple still enjoys a lead over Samsung in cloud access device and related revenue. In calendar 2012, Apple’s revenue was $165.7-billion compared to approximately $90-billion for Samsung across the same universe of products.  But, it is the difference in operating income that really explains why Samsung appears so targeted on Apple’s market share. For calendar 2012, Apple’s operating income was some $55-billion while Samsung’s OI for the similar range of products was less than $19-billion. Apple’s extreme profitability will drive Samsung to continue its focal shift in favor of cloud access devices (and related components, peripherals and services) at the expense of its other downstream businesses.

As the Samsung versus Apple battle intensifies, Apple is also being forced to make some changes. Samsung’s immense strength in semiconductors, flat panel displays, and batteries provides the wherewithal to build superior hardware. It also turned Samsung into Apple’s most important components supplier. Apple reportedly purchased some $8-billion of Samsung components in 2012. Every time Apple sells an iPhone or other product, it puts money in Samsung’s pocket. The reverse is not true.

And the high degree of reliance on Samsung as a components supplier has given its Korean rival a window into Apple’s product development and production plans.

Apple is clearly unhappy with helping to fund Samsung’s assault on its market share and with providing insight into its product planning. Cupertino is countering by becoming more involved in the design of the components in its products (discussed in our next post) and by signing up alternative suppliers to replace Samsung components in its supply chain. This strategy is not without risk. Apple is swapping out parts from the strongest components supplier in the market, in favor of those from lesser suppliers that Samsung has bested. (Nokia is also reportedly working to drive Samsung components out of its supply chain.)

As well, Apple must continue to strengthen its software and cloud offerings in the face of competition from Samsung partners, Google and Microsoft. The doomsday scenario for Apple is if Samsung can grab a significant lead in hardware and if Google and/or Microsoft, develop superior software and cloud offerings. Apple is already experiencing great pressure from Samsung Android phones and tablets. Windows 8 on Samsung PCs, smartphones and tablets may yet become a problem.

In this battle of sharply differentiated business models, Apple has been setting the pace. But Samsung is on the move and continues to impress 25 years after Domicity’s original report on the Korean electronics industry. It does not take too much of a stretch to envision Samsung ultimately overtaking Apple in the race for cloud access device market share leadership. This would be accomplished by offering superior hardware platforms that effectively leverage the cloud software and services ecosystems of Google, Microsoft, and others.

13.09.2012 Blog Comments

Amazon and Google — which business model works best?

Posted by Marc Brien, VP Research, DOMICITY LTD.

Domicity consults on the strategies & operations of I.T. companies and trends, and advises governments on economic development and investment attraction.

Amazon and Google, who would you rather ownAs businesses, Amazon and Google have things in common. However, their differences determine which company has the more advantageous business model.

Both companies are entirely cloud-based, founder-led, high-profile brands which compete in consumer, SMB, and large enterprise markets. They are active in the cloud as well as in cloud-access devices. Their annual revenue is broadly comparable.

For cloud mavens, another interesting point of similarity is that both Amazon and Google are striving to build up significant businesses selling machine cycles from their cloud platforms — IaaS and PaaS — to replace or supplement the internal IT shops of enterprise customers.

Despite these similarities, Amazon and Google differ radically on at least one important score — how they attract the bulk of their revenue.

Google generates most of its money by selling advertising to companies that want access to the millions of eyeballs attracted to the company’s web presence through: search, browser, email, maps, YouTube, news consolidation, financial information, social networking and more.

Amazon generates most of its money by using its cloud platform to sell stuff while taking a cut of the proceeds, just like the old bricks-and-mortar retailers. The stuff that Amazon sells includes a growing array of merchandise shipped from its own warehouses. Along the way, the company began using its cloud-based store to sell the books and other products of outside parties, letting these business partners handle the fulfillment issues. The company reportedly generates better margins from these partner sales than from its own merchandise sales.

A proportional income statement analysis tells much

This compare and contrast is a preamble to the main question: Who has the better business model? Put another way: Which company would you rather find in your holiday stocking?

The following table and graph demonstrate the differences between the Amazon and Google business models.


Amazon and Google Income Statement Comparison

 Amazon and Google Cashflow Comparison

Without much fear of spoiling the ending … the answer is Google to the question which company has the better business model. On a dollar of revenue, Google currently generates 11.1 cents of net income and nearly 15 cents of cash flow. Amazon is scrambling for every dollar of profit it produces. On a dollar of revenue it is generating less than a penny of net income and roughly 4 cents of cash flow.

A dollar of invested capital  — total liabilities and shareholders’ equity minus current liabilities — put into either business currently gives back a return through net income of more than 15% for Google and less than 4% for Amazon.

Expect to see Amazon moving in the direction of Google’s business model. For example, Amazon could encourage product suppliers to buy more advertising on To give these advertisers more exposure, Amazon could put up content designed to have consumers spend non-shopping time on the site.  As a for instance, Amazon could induce publishers to put up free video posts by big name authors, or manufacturers to provide how-to information on use of their products.

These days nearly every business seems to want a cloud-based business model, but some models are more profitable than others.

01.06.2012 Blog 1 Comment on Getting HP back on track

Getting HP back on track

Posted by Marc Brien, VP Research, DOMICITY LTD.

Domicity consults on the strategies & operations of I.T. companies and trends, and advises governments on economic development and investment attraction.


HP going off track In what has become customary Hewlett-Packard fashion, after a soft quarter the company announced a major organizational shakeup … all in the hopes of squeezing out some more dollars to fatten next quarter’s earnings report and buy some respect from Wall Street.

As cloud computing buffets the industry, HP’s business model appears to be coming off the rails. Although I was not a fan of many of Mark Hurd’s moves when he was HP’s CEO, Hurd did have a strategy, consistently applied; even if that strategy was insufficient to cause problems for HP’s two most important competitors IBM and Apple.

All the sturm und drang after Mark Hurd’s departure in summer 2010 highlights the need for HP to come up with a new strategy that goes well beyond the short-term appeasement of Wall Street.  HP is a great economic institution with a storied history and its 350,000 employees deserve better.

Here are my gratuitous (in the meaning of “free”, not “unwarranted”) suggestions for some big moves HP could make to get back on track.


Pick an enemy — On the server or “cloud” side of the  business, HP is number two after IBM. On the client or “cloud access device” side, it is number two behind Apple. HP may be able to go toe-to-toe with one of these companies, but it is overstretched battling both.

For FY2011 ended Oct. 31, HP generated an unimpressive 9.7% operating margin in its combined PC and printer client-side businesses compared to Apple’s 31.2%. HP’s revenue for the combined businesses declined 1.8% for the year.  Having abandoned its WebOS gambit there is no obvious fix to HP’s profit problem. Even if Windows 8 is a success, most of the profit in HP’s client-side business will continue to go to Microsoft, Intel, and other companies in HP’s value chain.

In HP’s server-side businesses — servers, storage, networking, software, and services — the prognosis is significantly better. The combined FY2011 operating margin was 14.4% compared to IBM’s 19.6% (the two companies report operating margin somewhat differently). HP’s strength in server-side infrastructure and related services should enable the company to improve its position in the lower half of the cloud stack and to accelerate the process of layering XaaS services on top.

In my view, Mark Hurd’s successor Léo Apotheker had it right. HP should divest itself of the PC business and its printer sidekick. The channel, some customers, and analysts will howl, but they will get over it. IBM did well by selling off its PC business to China’s Lenovo and spinning out its printer operation as Lexmark.

Ignore the share price — During the Hewlett and Packard regime, share price was not top of mind. Steve Jobs did not obsess over Apple’s share price. Instead they focused on making exceptional products that customers craved and for which they were prepared to pay a premium. Profit margins were healthy. Now, after each quarter that does not meet with Wall Street’s approval, HP’s top management pulls the plant up to see why the roots aren’t growing.

Most revenue and profit for HP and other old-line server-side vendors still emanates from traditional I.T. businesses. These businesses are fast disappearing into the cloud. Over the next three to five years, HP must radically reinvent itself if the company is to come out the other side of the transition to cloud computing with a healthy organization and income statement. During this period, share price should take a back seat to maximizing the cloud opportunity.

Here are a few starting points for the HP reinvention:

  • sever any connection between compensation and share price for HP management
  • instead, tie non-salary, incentive-based compensation to operational and market-based goals aimed at improving HP’s position in the cloud
  • refocus on products by stocking the Board and the top of the management pyramid with engineers and scientists who have strong cloud credentials, displacing a few of the MBAs and private equity company partners
  • stop spending a majority of cash flow on repurchasing HP shares — $10.1-billion in fiscal 2011
  • use the money spent on share repurchases to finance a successful transition to a cloud-based business model … increase R&D, continue building out a leading-edge network of cloud hosting data centers, make targeted acquisitions, and retool HR policies to restore trust and enthusiasm by re-skilling existing personnel, providing generous buyouts for non-cloud savvy staff nearing retirement, and offering industry-leading salaries to attract new hires with strong cloud credentials.

Just a few big things needed to put HP back on the rails. Get started and let me know how it turns out.


21.03.2012 Blog Comments

The data center is the computer

Posted by Marc Brien, VP Research, Domicity Ltd.

Domicity consults on the strategies & operations of IT companies and trends, and advises governments on economic development and investment attraction related to the IT sector.


Stack of naan bread

Stack of naan

The cloud computing paradigm is placing an increased emphasis on the importance of the data center as a source of competitive advantage for cloud-based businesses and operational units. This importance is signaled by such catchphrases as “the data center is the computer” or “the data center is the supply chain”.

If the data center is the computer, a single engineered system for the equipment and the building in which it is housed will generate the highest value-added and the least cost per web click. In the search for maximum return on investment, data centers can be constructed using standardized designs or by heavily customizing the architecture of each facility to the micro-environment of its geographic location and applications.

The history of the server market provides a useful point of reference as to the overall direction the market will take. Over time, x64 servers have steadily increased their market share versus the mainframe and RISC/EPIC architectures. The same dynamic is starting to be seen in cloud data centers; only the applications with the very highest value-added will be appropriate for custom-built monolithic centers. For the mass market, more standardized commodity designs will predominate.

These are early days for “the data center is the computer“. The market is in a wild and woolly experimental free-for-all as it feels its way towards the most efficient models. A variety of creative alternative data center designs are popping up; here are three substantially different examples:

  • IBM’s big stack of naan approachIBM’s large data center for Tulip Telecom in Bangalore, India. Unlike a typical North American sprawling one-storey data center, the 100-megawatt behemoth developed by IBM for Tulip Telecom, maximizes scarce and expensive land by going vertical. Four 5-storey towers house 200,000 sq ft of raised data center floor space. They sit on three floors of shared space. This represents a good example of a fully customized data center design.
  • Google’s Finnish paper mill — Minimizing energy use is the basis for the design of Google’s new data center in Finland. In 2009, Google paid about $52-million for a paper mill in Hamina, Finland. In a twist of fate, the paper-making operation had been forced to close due to competition from Internet-based media. The site features a 450-meter underground tunnel that stretches into the Baltic Sea, which originally provided cold water to cool a steam generation plant at the mill. Google has repurposed the tunnel to cool the servers in the new data center. This semi-customized design has been uniquely adapted to the local micro-environment but undoubtedly contains Google’s famously commoditized equipment architecture.
  • HP Cloud Services’ factory-made metal boxes — HP Cloud Services has chosen HP EcoPOD shipping container data centers as the hardware platform for its new public cloud hosting business … today’s ultimate in data center standardization. The EcoPODs, which typically are populated with dense x64 blades, are production-line manufactured and can be delivered in as little as three months.

In the server market, the opportunity thanks to their virtues in running scale-out applications. However, any rumors of the death of the mainframe remain premature. Likewise, it will be a long time before big custom-designed monolithic data centers are supplanted by commodity prefabs.

One thing is clear though. In an era when the data center is the computer, standardization and commoditization will increase. The major public cloud hosters including Amazon, Facebook, Google, Microsoft, and Twitter are already driving this trend forward.


15.03.2012 Blog Comments

Amazon AWS — the power of learning curve pricing

Posted by Marc Brien, VP Research, Domicity Ltd.

Domicity consults on the strategies & operations of IT companies and trends, and advises governments on economic development and investment attraction related to the IT sector.


Amazon AWS cuts pricesAbout a week ago, Amazon AWS, the leading provider of commodity IaaS (Infrastructure-as-a-Service) cloud hosting, announced yet another in a long string of price cuts that extend back to its start of operations in 2002 — an exercise in learning curve pricing.

Amazon AWS’ learning curve pricing strategy involves cutting prices to grow its business at an accelerated rate, while at the same time making it increasingly difficult for direct competitors to enter the market or to thrive. (See Domicity’s recent press comment in ComputerWorld  about Amazon’s price reduction as well as our comment in NetworkWorld about Sony’s transfer of part of its business from Amazon AWS to Rackspace OpenStack.)

Under classic learning curve pricing strategy, as a market leader continuously expands its business volumes, it drives down costs through steady improvements to operating procedures and increased purchasing power with suppliers. Most of these cost savings fund further price cuts, rather than increased earnings.

Competitors in the same market are forced to respond to the leader’s price reductions. Otherwise, they stand to lose customers. If these competitors do not enjoy the market leader’s cost advantages, they must swallow losses and, under investor pressure, they will eventually depart the market or be forced to sell out to another competitor that needs to scale up.

The diagram below illustrates how a market leader, using learning curve pricing, creates a positive feedback loop of cost reductions, price cuts, and expanded business volumes to discourage the competition.


Amazon AWS and learning curve pricing

Learning curve pricing is made easier by a company’s ability to endure thin gross margins. Amazon, the parent of AWS, has built a value chain designed to eke out acceptable earnings per share from relatively low gross margins … just 22% in 2011. This compares to Microsoft at 78%, Google at 65%, and IBM at 47%. Hewlett-Packard, with a gross margin of 23% and Dell at 22% have more similar cost structures to Amazon.

An alternative strategy to challenging the market leader head-on is to develop a differentiated product line that will support a price premium. For example, HP is entering the public cloud hosting market and hopes to avoid direct competition with Amazon AWS by introducing a higher value-added offering. HP Cloud Services’ initial services are in beta, with the official roll-out said to be targeted for May.

Biri Singh, the SVP recruited from IBM in late 2010 to form and lead HP Cloud Services was quoted in the New York Times (Mar. 9, 2012) saying, “We’re not just building a cloud for infrastructure. Amazon has the lead there. We have to build a platform [PaaS] layer, with a lot of third-party services. … We won’t pull [Amazon’s] customers out by the horns, but we already have customers in beta who see us as a great alternative. We are not coming at this at 8 cents a virtual computing hour, going to 5 cents.”

To a certain extent, Amazon AWS is also being forced to tap dance around the implications of its own IaaS learning curve pricing strategy — trying to leverage its lead in IaaS hosting to move up into higher value-added offerings.

In early 2011, Amazon AWS launched a PaaS service called Elastic Beanstalk. According to the company, developers only have to upload their application and Elastic Beanstalk “automatically handles the deployment details of capacity provisioning, load balancing, auto-scaling, and application health monitoring” on the Amazon AWS IaaS platform.

Farther off into the future Amazon AWS can be expected to move into the SaaS market.


21.02.2012 Blog Comments

Is IBM market value bolstered by too much debt?


Posted by Marc Brien, VP Research, Domicity Ltd.

Domicity consults on the strategies & operations of IT companies and trends, and advises governments on economic development and investment attraction related to the IT sector.


Chased by DebtIs IBM’s management falling into the same trap as many households? The siren song of low interest rates appears to be luring Armonk to take on too much debt and some competitors are on the same path.

Benjamin Graham-style value investors tend to shy away from companies that have less than 50% of their assets financed by investors through shareholders’ equity. The fear is that when interest rates spike upwards, or if a company hits choppy waters, investments in critical areas will have to be cut to cover increased debt charges.

As the graph below shows, IBM sits well below Ben Graham’s 50% margin of safety line. Just 17% of IBM’s assets were financed by shareholders’ equity (at Dec. 31, 2011). This compares to 21% for Dell, 25% for Accenture and 30% for Hewlett-Packard. In better shape, according to Graham, are Microsoft at 57% and Oracle at 58%.

IBM versus competitors - equity as percent of total assets

IBM has good cash flow, so why is it carrying so much debt and why are Dell, HP and others climbing into the same boat?

The answer … to pump up earnings per share. This, in turn, puts upwards pressure on IBM market value and share price (and also boosts the value of options and share packages for management).

In 2011, IBM generated about $20-billion in cash from operations and spent 121% of it on:

  • repurchase of shares plus dividends – 92%
  • capital expenditures – 20%
  • acquisitions (net of divestitures) – 9%

IBM made various financial moves including increasing its debt load to cover its cash flow deficit.

The company’s biggest 2011 outlay, totaling $18.5-billion, was spent on currying favor with investors through dividend payments and share repurchases. This compares to the $5.9-billion that IBM invested in expanding the business through capital expenditures and acquisitions.

Share repurchases have become very popular with IBM, HP and some others. They reduce the “per share” denominator of earnings per share (EPS) which puts upward pressure on EPS and share price.

In the decade between 2001 and 2011, IBM repurchased 557-million shares. As a result, the company’s 2011 EPS was 46% higher than would otherwise have been the case without these share repurchases.

Many investors are undoubtedly happier with an increase in IBM market value. However, more cautious investors would prefer to see the company increase its margin of safety by significantly decreasing the proportion of assets financed by debt.


07.02.2012 Blog Comments

IT job cuts — does the cloud have a silver lining?

Posted by Marc Brien, VP Research, Domicity Ltd.

In addition to published reports and consulting on the strategies & operations of IT companies and trends, Domicity advises governments on IT sector economic development and investment attraction.


The cloud is not just about IT job cutsIn our previous post, we talked about the likelihood that cloud-driven automation will cause major IT job cuts among internal IT shops and services suppliers.

Fans of long-departed Austrian economist Joseph Schumpeter will proclaim that there is nothing to worry about because the cloud will merely unleash a healthy spate of creative destruction. IT job cuts to those activities that prop up legacy architectures could be more than offset by new jobs from out of the cloud.

To a significant degree, Schumpeter was right. Many new cloud-based jobs are being created at IT companies which are:

  • undergoing rapid expansion using cloud-based business models (Google, Amazon, Twitter, Facebook etc.)
  • developing apps for mobile cloud access devices, in particular smartphones and tablets
  • implementing RFID and sensor-based applications for the “Internet of Things”
  • developing Big Data systems to analyze and manage the tsunami of structured and unstructured cloud-based data.

The problem with this rosy scenario is that the people losing employment from IT job cuts are often not the same as those taking the new jobs. Different skills may be required and the new jobs may be inconveniently far from the vanishing legacy jobs. The article in the Atlantic referred to in our previous post discusses how these problems have played out in the broader manufacturing sector throughout the US.

The loss of many good-paying manufacturing jobs has almost certainly been a significant contributor to the income crisis among a growing number of households. Growth in US median family income is falling well behind the growth of the overall economy. Lane Kenworthy at the University of Arizona developed the graph below which demonstrates this point.

 Manufacturing job cuts have affected family income

Domicity expects a major cloud-driven round of IT job cuts to put further downward pressure on median income.

The lack of effective transition programs to ease the plight of those laid off during the restructuring of the US manufacturing sector has been one of the contributors to the social and political unrest leading to the Tea Party and the Occupy movements. More social and political instability could be fueled if the impacts of cloud-induced IT job cuts are ignored and transition programs are not put in place, such as:

  • retraining and education courses in cloud-optimized programming languages
  • courses in cloud-centric business models
  • loans to business, particularly SMBs, to develop new opportunities that leverage the cloud
  • subsidies to offer older IT workers an early retirement.

Have enough lessons been learned from the restructuring of the manufacturing sector to ease the cloud’s potential for social disruption?


30.01.2012 Blog 1 Comment on IT job cuts — A hard rain’s a-gonna fall

IT job cuts — A hard rain’s a-gonna fall

Posted by Marc Brien, VP Research, Domicity Ltd.

In addition to published reports and consulting on the strategies & operations of IT companies and trends, Domicity advises governments on IT sector economic development and investment attraction.



IT job cuts - a hard rain's a-gonna fall from the cloud

I saw a highway of diamonds with nobody on it …
It’s a hard rain’s a-gonna fall
Bob Dylan


The Jan/Feb 2012 issue of the Atlantic magazine, has a fascinating article on the destruction of US manufacturing jobs.

The article makes the argument that, contrary to popular perception, the US remains a major manufacturer. Over the last 10 years, manufacturing has expanded by about one-third. However, the number of manufacturing jobs has declined by one-third.

Whipsawed between the twin forces of offshoring and automation, US manufacturing job losses between 1999 and 2009 totaled about 6-million. As the article’s author Adam Davidson points out, “About as many people work in manufacturing now as did at the end of the Depression, even though the American population is more than twice as large today”.

I fear there are strong parallels with the developing situation among IT workers — both in internal IT shops and among services suppliers. They too are being torn between the twin forces of offshoring and cloud-driven automation. Further major IT job cuts are on the horizon.

We’ve posted about why India and some other developing countries have become such popular offshoring locales for IT jobs. In a recent chat with a senior IT manager at a major bank, he bemoaned the fact that his operation is planning major IT job cuts in North America and will offshore the work to India.

Further down the line, the same senior IT manager saw cloud-based automation forcing further job cuts and was thankful his career was almost over and that he wasn’t “one of these kids coming into IT today.”

A real world example of IT job cuts

Here’s a real world example of the scale of IT job cuts that cloud computing has the potential to visit upon major IT shops.

Between 2005 and 2008, Hewlett-Packard undertook a thoroughgoing centralization and modernization of its internal IT platform. The goal was to put in place a leading-edge shared services/private cloud architecture for the entire company.

Much has been written about the positive outcomes of this modernization effort, led by former CIO Randy Mott: a halving of the total IT budget in 2005 terms; consolidation of some 85 significant data centres and 370 smaller sites into just three large new mirrored data centres; a 60% reduction in energy costs; a 40% reduction in total number of servers from 25,000 down to roughly 15,000 blades, yet a 250% increase in processing power; a 50% cut in networking costs, despite a tripling of total bandwidth; and a reduction in company-wide applications from approximately 5,800 down to 1,400.

What is not often mentioned about HP’s private cloud is that the modernization effort also targeted a 11,000 person reduction of IT jobs within the company.

As well, before HP’s private cloud project, about 50% of the company’s IT work was done by outside contractors. Now, 90% of the work is carried out internally. Cloud computing threatens IT job cuts among system integrators and outsourcers, not just internal IT staff.

In mid-2010, HP began a second major renovation effort, modernizing the additional IT infrastructure that came along with its 2008 acquisition of EDS. This targeted further IT job cuts of 9,000 positions, in addition to the 11,000 eliminated in the first round of cloudification.

A hard rain indeed.

Our next post looks for a silver lining in the darkening cloud of IT job cuts.


20.12.2011 Blog 3 Comments on A red flag for labour-intensive IT

A red flag for labour-intensive IT

Posted by Marc Brien, VP Research, Domicity Ltd.

This post is based on research for an upcoming Domicity CORProfile© report analyzing the strategies and operations of a global cloud leader.  CLICK HERE  for advance notice and 33% savings on early orders.


A Red FlagCloud computing’s Everything-as-a-Service (XaaS) paradigm has the potential to radically alter the face of the IT outsourcing industry, conferring market power on those companies who can deliver customer applications from global networks of technologically advanced data centres, at the expense of those whose business models are primarily driven by inexpensive labour.

The traditional “mess-for-less” IT outsourcing model is based on a supplier taking over a client’s increasingly dysfunctional muddle of incompatible multi-generational systems that span several acquisitions and mergers. The outsourcer promises to run the customer’s mess for less. How is it that leading IT outsourcers can deliver a customer’s applications for 20% less than the in-house IT group, while also generating handsome profits? The answer lies in economies of scale, technology, and particularly, in offshoring work to locales where labour costs are substantially lower.

In a previous post, Domicity highlighted the luxurious operating margins being generated by leading Indian outsourcers.  The seemingly bottomless well of low-cost, well-educated workers has also drawn many western IT outsourcers to build a significant presence in India.

Roughly one in every three IBM employees is now located in India.  IBM’s total employment in India by the end of 2011 is estimated at nearly 155,000, up from just 5,000 in 2002. Accenture has about one-third of its headcount in India and HP has a sizeable services contingent on the subcontinent along with several other “BestShore” service delivery centres in developing countries. Meanwhile a group of Indian companies  — Tata Consultancy Services, Infosys, Wipro, and HCL — have leveraged their country’s labour advantage to become a potent force in IT outsourcing.

Enterprise-class XaaS cloud-hosted IT outsourcing promises to eventually send mess-for-less outsourcing into decline, as internal IT shops find they can:

  • Outsource one application at a time, rather than having to commission a wholesale outsourcing of the entire IT group (software suppliers are turning their flagship products into applications that can be offered by a range of SaaS hosters)
  • Convert the fixed cost of running applications into variable costs, based on business volumes
  • Improve price/performance as cloud hosters with experience delivering a specific service — for instance SAP ERP or Microsoft SharePoint — compete for business
  • More easily switch from one IaaS or SaaS supplier to another than they could between traditional IT outsourcers
  • Simplify the design and integration of solutions carried out by internal IT staff; each application can run on standardized virtual machines, rather than on specifically tuned in-house hardware
  • Speed up service delivery through automation and reductions to the operational labour needed to deliver cloud-hosted services.

XaaS should constitute a red warning flag for all IT outsourcers whose business model continues to be based on using cheap labour to deliver mess-for-less outsourcing, integration and management services.

In-house IT staff should also beware. XaaS holds great potential to automate these jobs out of existence as well.


21.11.2011 Blog 1 Comment on Humans Matter

Humans Matter

Posted by Susan Sparrow, President Domicity Ltd.

This post is based on research for an upcoming Domicity CORProfile© report analyzing the strategies and operations of a global cloud leader.

Please click here for advance notice and 33% savings on early orders.


Machine cogs from Chaplin's Modern Times movieHewlett-Packard, once a beacon of progressive human resources policy, has lost its way. Regarded for years as a top place to work, the company has not been seen on Fortune’s annual ranking of “America’s 100 Best Companies to Work For” since 2000.

Other IT companies in the Silicon Valley/Bay area have picked up the torch. For instance, in 2011 Google was #4, NetApp was #5, Cisco was #20, and Intuit placed #44. A couple of other companies were knocking on the top 50 door, notably Intel (#51) and (#52).

Our analysis shows that 30 IT/electronics companies made it to Fortune’s top 50 companies between 1998 and 2011.  Fifty-seven per cent were based in the Bay Area. (A few companies disappeared along the way into the belly of a bigger fish.)

Companies “from away” that cracked the top 50 in 2011 included SAS of North Carolina, judged by Fortune to be the number one company to work for in America for 2011 and 2010, and the only IT/electronics company to make the top 50 every year between 1998 and 2011 (Cisco was off for just one year). Qualcomm of San Diego clocked in at #33 for 2011 and Shared Technologies (now an Arrow Electronics subsidiary) at #43.

Yet, it was HP that pioneered the idea of an employee-centred corporate culture, not just within the Valley, or the US but throughout international business. The “HP Way” that Bill Hewlett and Dave Packard put in place included the introduction of such progressive HR practices and benefits as: profit sharing and employee share purchase plans; flex time; job sharing; telecommuting; open cubicles to support the free flow of ideas; promotion of women to senior management; support for employee volunteerism; a portion of time for engineers to work on their own ideas; development of a collegial environment where workers felt safe sharing their ideas with others without fear of having them stolen; and disciplined hiring that would prevent the need for layoffs.

HP’s employee-centred corporate culture was a fundamental part of its strategy for the first 60 years of its existence — from the late 1930s through to 1999. The notion was that the world’s best and brightest would flock to HP and generate a firestorm of creativity and enthusiasm that, if harnessed and channeled by a strong management regime, would result in revenue growth and high levels of profitability. A key element of the management regime that Bill and Dave instituted was the early adoption of Peter Drucker’s Management by Objectives.

All this worked. Over the years of the HP Way, the company grew from three employees — Dave and Lucile Packard and Bill Hewlett — with $5,369 in 1939 revenue to 124,600 employees and $47.1-billion in revenue for fiscal 1998.

By the late 1990s, as dot-com madness stalked the land, some believed that HP lacked the bloody-mindedness and marketing smarts to prevail. This led, in 1999, to the company hiring its first outsider and non-engineer as CEO. Carly Fiorina was a marketer with a background in networks (Lucent and AT&T).

When Fiorina took over as CEO, HP ranked #10 in Fortune’s 100 Best Companies to Work For. By the end of her first full year at the company, it had dropped to #43. After that HP disappeared from the list. When Fiorina left in 2005, HP picked another outsider as CEO, Mark Hurd.

At HP, both Fiorina and Hurd embraced the human resources principles now common throughout much of North America’s multinational sector. Workers, just another commodity, could be sourced globally from wherever they were least expensive and in best supply — sometimes referred to by the unfortunate term “labour arbitrage”. HP’s value chain was made lean and mean by waves of mass layoffs, typically in multiples of 10,000 workers at a time. Wages were slashed and benefits were stripped to help juice up the company’s earnings. At the time, employees mumbled darkly that the goal was to plump up the options-enriched compensation packages of those at the top of HP’s pyramid.

In mid-2010, a few months before Mark Hurd was let go, the results of an internal HP employee survey surfaced, revealing that two-thirds of the staff would work somewhere else for the same salary. Facing a serious morale challenge, HP’s new CEO Meg Whitman acknowledged, shortly after being appointed in Sept. 2011, that employee relations must improve.

As HP tries to maximize cloud computing opportunities in the face of tough competition from more employee-centred competitors, it could rediscover that humans matter.


02.11.2011 Blog 2 Comments on Apotheker or Whitman … HP’s challenge remains the same

Apotheker or Whitman … HP’s challenge remains the same

Posted by Marc Brien, VP Research, Domicity Ltd.


Sumo wrestlers

Boosting gross margins

Hewlett-Packard needs to generate fatter gross margins to fund the spending required to grab the advantage in cloud computing.  Under previous CEO Mark Hurd’s direction, HP boosted earnings by under-investing in core areas — notably wages, R&D, sales and marketing. Domicity believes this approach was unsustainable.

After several poor quarters, Hurd’s replacement Léo Apotheker moved to speed up the transition to a higher-margin business model by spinning out HP’s low-margin Personal Systems Group (PCs and other client devices). This caused such an uproar among HP’s customers, suppliers, channel partners, and investors that the Board dumped Apotheker.

Replacement CEO Meg Whitman reversed course, deciding to keep the PC business. But retaining the $41-billion a year Personal Systems Group returns things to square one. Whitman must still define a new path towards fatter gross margins to fund the necessary cloud-enabling investments.

It is worth reprising the table from the earlier post (shown below), with its 30-year proportional analysis of HP’s income statement. This analysis illustrates just how undernourished the company’s cost structure has become over the years. Note how R&D and SG&A (Selling, General & Administrative) expenses were reduced to the bare bones to enable thin gross margins driven by a commodity pricing strategy.


Click on image for a larger PDF version

30-year Proportional Income Statement Analysis of HP


Over the long run, HP will not be able to compete effectively for cloud opportunities if spending on R&D and SG&A remains a fraction of the spending by competitors like Apple, IBM, Cisco, and Oracle. Increases in gross margins must be found to finance increased spending in these areas.

Efforts to wring strong and profitable growth out of the PC and printer businesses will continue, but will require patience. Domicity believes that the lion’s share of HP’s focus over the medium term will be on boosting its market presence in server-side businesses including enterprise servers, storage, networking, software, and professional services.

In 2011, HP invested more than $10-billion in acquisitions on Autonomy and Vertica to establish a strong data analytics business. This is just a down payment on what the company will need to spend to close the gap on server-side market leader IBM.

Once the company’s balance sheet has been restored from the Autonomy purchase, we have already speculated that HP may find Accenture an irresistible acquisition target. Also look for HP to make acquisitions to build up its cloud hosting business, including as-a-service software offerings.