Chapter 2
The “Industrialization” of Technology
Kevin O’Marah lives and breathes logistics. A senior analyst at Gartner, the research firm, he has focused for years on the science and magic behind logistics. His team analyzes practices and challenges in a wide range of industries including aerospace, auto, consumer products, and life sciences.
Each year the supply chain team at Gartner publishes a list of companies they respect across industry lines. For the past four years (2008–2011), Apple has topped that list.1
The Apple—Physical and Digital—Supply Chain Mastery
O’Marah says,
Apple's leadership has been pervasive. Traditional supply chain disciplines like managing an extended network of contract manufacturers and component suppliers are fully in force, but beyond these areas Apple has led in at least two vital ways. First is in its huge leverage of the digital supply chain. By fostering the development of a secondary market in applications for its iPhone, the company has shown again (as with iTunes) that consumer product revenue growth with zero inventory is not only possible, but repeatable. The other area in which Apple's supply chain leadership is increasingly important is in the retail experience. As one of a handful of deeply vertically integrated brands, Apple's retail chain achieves almost unimaginable success in its stores.
While Apple gets plenty of kudos for the elegance of its products, peel the onion, and, as we will show in various chapters, it has built a retail store chain that is the envy of even long-time retailers. It has built an elaborate global network of suppliers and contract manufacturers that has confused the traditional accounting that economists use to measure global trade. In addition to the elaborate physical supply chain, it has had to integrate the digital supply chain as iPhones are activated via iTunes at customer homes and via carriers. As it rolls out its iCloud, it has built one of the biggest data centers in the world. It has built an ecosystem of apps and games around its products at a scale never seen before. Admirably, it built its supply chain in a much more volatile industry than that of consumer products or chemicals.
Of course, Apple has itself driven the high-tech industry volatility with its own pace of product introductions. Dell used to be held as a benchmark of efficiency with its “build-to-order” supply chain. It manufactured mostly what was ordered and even paid for in advance. Apple raised the bar by showcasing a new product, guesstimating likely demand, and tuning its supply chain day by day and hour by hour. It broke traditional rules of demand forecasting—because there was little historical data to forecast from for a version 1.0 iPod or iPhone or iPad. It balanced the risk of overproducing or increasing buffer inventory and taking write-offs versus underproducing and losing customers to the next competitive product just a few weeks away. It took that risk, time and again, and made the rest of the industry do the same. And the risks are not small when you are talking 3 million iPads in the first quarter of introduction.2 The Dell model has been “out-Delled.” As a Dell executive acknowledged, the old model worked well for 20 years but the “environment has changed.”3 Dell, no slouch, is reshaping itself to this new world where we are not building to order but to anticipated demand.
Retail wizardry, contract manufacturing, supply chain efficiencies—we used to look to companies like Walmart, Fedex, and Procter & Gamble for leadership in these practices. Now we are seeing well-run technology vendors like Apple become the “best practice” paragons.
Not Just Apple
Not everyone agrees with the Gartner supply chain rankings. Some think its methodology weighs financial performance too heavily. Others think it does not factor sustainability enough.
An Apple competitor thinks Gartner is being too generous to Apple and other technology vendors in its ranking. “I would not call them or us a textbook supply chain management leader. Having said that, the elite who run operations in high-tech are all darned good at execution and adjustments.” He says that as his team furiously reworks supplier scenarios impacted by the Japanese tsunami in early 2011. “Not to be unkind about the tragedy, but this is what many of us live for—the adrenalin rush from such disruptions.”
As we will see in the HP case study later in this chapter, volcanoes, explosions at supplier factories, labor issues at suppliers, and shipping sea lane delays are fairly common disruptions in the technology supply chain.
“Constant course correction” is how one former executive refers to the approach. “If the executive team decides to change direction, it's instantaneous,” this ex-Apple honcho says. “Everybody thinks it's a grand strategy. It's not.”4
Transitioning from “make-to-order” to “make-to-constant-course-correction” is tricky, to say the least. Zynga, the social gaming company famous for Farmville and popularity with Facebook users, launched CityVille in 2010. “IT staff watched participation on CityVille build slowly the first few days—at that pace, it might have taken two years to reach the first million participants. Then registration started to ramp up. Four months after launch, CityVille had brought 20 million new users to Zynga's daily computing workload.”5
In early 2011, a document circulated about Amazon.com that had language like:
The diversity of products demands that we employ modern regression techniques like trained random forests of decision trees to flexibly incorporate thousands of product attributes at rank time.
What was remarkable about the document was that it was not a discussion aimed at university math majors or even attendees at a technology conference. It was part of the annual shareholder letter Jeff Bezos, CEO of Amazon.com, sent to his investors. (The complete letter is shown in Chapter 20.)
The letter went on to include statements like:
Service-oriented architecture—or SOA—is the fundamental building abstraction for Amazon technologies. Thanks to a thoughtful and far-sighted team of engineers and architects, this approach was applied at Amazon long before SOA became a buzzword in the industry.6
The financials that followed highlighted that Amazon does not just have large distribution centers around the world. It was well on track to sell over $1 billion in 2011 in services to companies like Zynga above (Zynga uses Amazon's elastic cloud when it launches products and brings the processing in-house when a game hits a more predictable customer level) from massive data centers it has set up around the world. Amazon's e-reader Kindle and e-books are expected to generate over $5 billion in revenues in 2011.7
Amazon is a legitimate technology vendor feared and respected by its competitors. Indeed, its subsidiary Lab126, which designed the Kindle, is based in Cupertino, California, a stone's throw from Apple's headquarters. What distinguishes Amazon even more is its savvy with physical logistics, which was tested from the beginning by competition with Walmart and other retailers. As it has consistently used discounted shipping as a competitive advantage, it has been sophisticated in its design and location of distribution centers. It has innovated with “postal injections,” in which it uses its own trucks or independent carriers to drive truckloads of orders to local postal depots from Amazon warehouses. The procedure eliminates processing steps for the U.S. Postal Service.8
Google is known for its search and mobile technologies, but even more impressive are its investments and prowess in clean energy innovation. For its data center and other operations, Google buys renewable energy from various wind farms and has invested in large solar panel farms and in geothermal companies. It has won approval from the Federal Energy Regulatory Commission (FERC) to become an electric utility. At a data center event it hosted in Zurich, Google shared what it has learned from operating efficient data centers, “laying out the best practices that any IT manager can do on a reasonable budget, such as maintaining hot and cool aisles, managing airflow, measuring the energy consumption in the facility, using outside air whenever available, running the data center at a higher temperature like 80 degrees, and optimizing power distribution.”9 We further discuss Google's green initiatives in Chapter 17.
Even More Examples—eBay, Facebook, and Twitter
The PayPal unit of eBay, in business slightly over a decade, has revolutionized the online and mobile payment world for millions (particularly individuals and small businesses) in 190 countries. In 2010, it processed over $90 billion in payments.10 To do this, it has implemented technology to handle the intricacies of regulatory controls across countries, protection against fraud, tracking money laundering, and more.
As mobile payments gain traction, this technology is even more in the crosshairs of much more established banks, credit card companies, other technology vendors, and even giant telecom companies. Telecom companies hope to take advantage of next-generation mobile phones with near-field communications (NFC) so you can pay by tapping the phone on a specialized reader at the store.
eBay has also demonstrated remarkable growth in its mobile marketplace business. CEO John Donahue told analysts in January 2011, “We extended our leadership position in mobile commerce, nearly tripling our eBay mobile (general merchandise value or GMV) year over year to nearly $2 billion, with strong holiday shopping momentum. In 2011, we expect mobile GMV to double to $4 billion. Mobile is clearly becoming a new way people shop. Our mobile apps have been downloaded more than 30 million times in eight languages across 190 countries.”
In 2011, Facebook opened the doors to its large data center in Prineville, Oregon, which at full capacity will be 300,000 square feet. Facebook needs such large centers to handle its explosive growth of over 750 million active users and over 3 billion photos posted by users each month. What was particularly interesting was the design of the center:
It has spent two years developing everything that goes inside its data centers—from the servers to the battery cabinets to back up the servers—to be as green and cheap as possible. For example, Facebook uses fewer batteries thanks to its designs, and to illustrate how integrated the whole computer operation is, the house fans and the fans on the servers are coupled together. Motion-sensitive LED lighting is also used inside. The result is a data center with a power usage effectiveness ratio of 1.07. That compares to an EPA-defined industry best practice of 1.5, and 1.5 in Facebook's leased facilities.11
We will discuss this data center in more detail in Chapter 8.
While critics say it is mostly frivolous and there are numerous jokes about its frequent “fail whale” when the site goes down, it is remarkable the influence Twitter has had across the world. That is especially true when you consider it is just a five-year-old company with a staff of fewer than 500. Whether it be a catalog of the uprising in Egypt or the Japanese tsunami, or even an inadvertent live tweet during the Osama bin Laden raid, as Brian Solis has observed, “The new reality of a real-time world is that news no longer breaks, it Tweets.”12
What is fascinating is the simplicity behind the design of the service. As the MIT Technology Review wrote in 2007, “(Co-founder) Stone says the company completed a working version of the software in only two weeks using Ruby on Rails, a programming language and a set of prefabricated software modules widely employed by developers of the new raft of Web services known as Web 2.0. The hard part, he says, was ‘navigating the business aspects of the mobile industry. It took us months to get a short code and figure out how to play nice with all the major U.S. and international mobile carriers.’ ”13
While it may have started simply, Twitter has scaled immensely, as it described its new search architecture in 2010: “With over 1,000 TPS (Tweets/sec) and 12,000 QPS (queries/sec) = over 1 billion queries per day (!) we already put a very high load on our machines. As we want the new system to last for several years, the goal was to support at least an order of magnitude more load.”14
Conclusion
Apple, Amazon, Google, eBay, Facebook, Twitter, and Zynga are mostly focused on consumer-facing technology, but under the covers show the power of “industrialization of technology.” These are technology vendors emulating best practices from varied industries, even competing with them and taking them to new levels. They tend to focus on each other. Amazon, Google, and Apple are all vying for streaming music service.15 Amazon competes with eBay for online commerce and so on. In the process, they are raising the bar for established technology vendors like IBM and Microsoft and for corporate IT. In Chapter 1 we saw how technology “buyers” were becoming vendors. The vendors described in this chapter and HP with its complex global chain described in the following case study are behaving more like traditional technology buyers—setting new benchmarks of productivity and efficiency.
Case Study: HP—The Quest for a “10 Out of 10” Supply Chain
“It could be your darkest hour or your finest,” says Tony Prophet as he discusses the impact of the Iceland volcano explosion, the Japanese tsunami, and other disruptions on HP's supply chain. Prophet, a senior vice president, oversees hardware purchasing and logistics for HP.
It is a massive operation, the largest in the technology industry, with over $60 billion in components, warehouse, transportation, and other logistics costs. The HP machine churns out two personal computers a second, two printers a second, and a data center server every 15 seconds.16 It is also a constantly evolving operation with a changing mix of company-owned factories and contract manufacturing utilizing air/ship/rail logistics from/to most countries around the world. So those challenging “hours” Prophet talks about come with alarming regularity.
In August 2011, HP announced it was looking at strategic options around its PC business; as the market moves to tablets and mobile devices, and as the company sought to invest more in software markets. Under a new CEO, Meg Whitman, it reversed course a few weeks later, citing, among other things, the impressive supply chain performance of the hardware business we describe.
The Short-Term Acrobatics
One of those defining hours Prophet talks about was after the Eyjafjallajokull volcano explosion in Iceland in April 2010. That event ended up disrupting flights to northern European airports for weeks. HP flies many of its products from Shanghai, China, to those airports. On many days, HP is the biggest single buyer of outbound airfreight from Shanghai. Fully loaded 747-400 cargo planes with HP notebooks and other products often take off three times a day from that airport.
Prophet's team immediately decided to switch shipments destined for northern airports like Frankfurt, Germany, to southern airports like Barcelona, Spain, and Naples, Italy. The problem was they could not find sufficient commercial flights for those destinations from China. So they booked charter flights. Even as HP products were arriving in Europe, competitors found charter freight prices had gone up 50 percent or were not available anymore. HP's 72-hour lead turned out to be significant.
In March 2011, Prophet was woken up at 3:30 a.m. to be informed of the massive Japanese earthquake and tsunami.
The New York Times wrote, “Japan is the world's third-largest economy, and a vital supplier of parts and equipment for major industries like computers, electronics, and automobiles. The worst of the damage was northeast of Tokyo, near the quake's epicenter, though Japan's manufacturing heartland is farther south. But greater problems will emerge if rolling electrical blackouts and transportation disruptions across the country continue for long.”17
Within two hours, HP had set up a crisis management team. Japanese component manufacturers are a key part of HP and competitors' products, so the HP team had to quickly assess the status of component inventory and its safety. Twenty-four hours later, the second risk-management phase began—to acquire additional buffer inventories of components in case the crisis caused a shortage or increase in prices. HP is the largest purchaser in the industry of many of the components like memory, so hedging against the disruption was critical.
“There are so many other examples of disruptions I could list,” continues Prophet.
There are explosions at supplier factories, labor issues at suppliers, and shipping sea lane delays. When you run a complex, global supply chain, there are thousands and thousands of links that can be disrupted and require rapid remediation.
Prophet chuckles when he talks about other spikes that happen on a regular basis. “We obviously work with our channel partners to plan for increased demand around back-to-school campaigns, Black Friday promotions, or regional holidays like the Chinese New Year. Those are significant events to plan for, but compared to the havoc from the volcano or the tsunami, those are so much easier to manage.”
The Long-Term Shifts
If those short-term maneuvers are dazzling to watch, even more impressive are the long-term shifts Prophet talks about.
One of our ongoing projects has been to streamline our supply chain, especially after major acquisitions like Compaq or Palm or 3Com.
The HP-Compaq merger, a particularly large one, afforded plenty of streamlining opportunities over the past several years. “In 2006, we had a combination of 70 (HP and Compaq) supply chain nodes worldwide, most company-owned and in relatively high-cost locations. Also, most were geared to making desktop PCs and the notebook market was taking off.”
As Prophet told a meeting of financial analysts in December 2010:18
Each of those nodes had inbound hubs, outbound hubs, unique IT connections, and drove a tremendous amount of overhead to support the complexity of these nodes.
Now it has been streamlined to
About 30 nodes and significantly fewer of them company-owned. Obviously biased more towards lower-cost locations, but not exclusively. We continue to operate a plant in Indianapolis. We continue to have a plant in Japan, so where it makes sense for proximity to the customers, to serve those customers with high velocity, we're there.
Prophet continues:
That's just the PC business. Now layer on our imaging, printing, enterprise server, network nodes. At one point we had more than 450 nodes across all our supply-chain networks. So just in Southern California, we had three warehouses within 30 miles of each other. So printing equipment, PCs, and notebooks would come on different shipments and go to different warehouses—and get shipped separately to the same distribution centers, as many were often going to the same customer. The warehouses, the containers, the trucks were not always fully utilized. So there were plenty of opportunities to consolidate real estate, shipping, and trucking and optimize all those costs.
Another opportunity was to move to standardized components across product lines. So, not to buy connectors all over the place. We have moved from about 60 percent standardized and centrally procured three years ago to over 85 percent.
That has brought another advantage. A few years ago, we bought many components and commodities on the spot market. We have shifted to long-term commitments to key suppliers. The suppliers can make longer-term capacity investments and we don't have to be as dependent on the vagaries of the volatile market.
So, there are times when HP overpays compared to the spot market. Can you imagine, though, being exposed to the spot market for critical Japanese components after the tsunami?
Finally, there are process and IT opportunities. As Prophet told that meeting of analysts, “Within the Personal Systems Group, the PC business, we have 65 ways of planning, You've got one for notebook and one for desktop, one for consumer notebook and one for commercial notebook, one in EMEA and one in Latin America, et cetera, et cetera.”
Across the whole business, HP had more than 1,000 processes and 300 IT applications. “So our objective is to cut the processes by an order of magnitude and cut the IT applications by more than a third. We're moving to common IT applications to drive those processes, and so that to our suppliers and to our customers we look like one company.”
“We've got a path, a strategy charted to build what we think will be a ‘10 out of 10’ supply chain, and we're about a year into this transformation.”
You cannot serialize these projects. They all have to march forward in parallel and they all have to adjust to the short-term hiccups and shifts in the market as the recent market move in many Western markets to tablets away from notebooks.
Another fascinating dimension of the HP supply chain is its global reach and its pioneering investment in Western China.
The Global Dimension
Prophet, in that presentation to analysts, highlighted some of HP's global reach.
First multinational to manufacture PCs in Russia.
Upgraded our operations in north central India.
Significantly upgraded our operations in Brazil.
We believe we'll be the first multinational manufacturing PCs in Turkey. That'll allow us to effectively serve central Asia, Eastern Europe, and the Middle East, the Mediterranean region particularly.
Particularly impressive is the HP pioneering expansion into Western China.
China's west contains 70 percent of its land mass, 30 percent of its population, but just 20 percent of its total economic output. Per capita GDP in the mega-city of Chongqing, which has a population of 27 million, was USD 3,500 in 2009, versus USD 10,500 in Shanghai.19
It has been called the “New Wild West” and there have been multiple attempts to develop that part of the country that have been hampered by: “poor infrastructure, minute levels of outside investment, an ailing environment (especially in Chengdu, Sichuan Province, and Chongqing Municipality), and a weaker education system than the rest of China.”20
Prophet continues:
We have been in China over two decades, and like other Western companies we could see the spiraling real estate costs, employee absenteeism, stretched infrastructure in the developed East and South. Starting in 2007 we started to evaluate alternatives like Vietnam and Malaysia. But the more we looked at Western China, the more we also saw the opportunity for the domestic market. If you draw a circle out 750 miles from Chongqing, you are looking at about 300 million people—which by itself makes it one of the largest PC and other gadget markets in the world. So that is our first focus here in the west of the country—made in China, for China.
Indeed, the Chongqing plant has already produced for the domestic market what is being called a “rural” laptop—designed to handle intense heat and rain—and that model is also seeing demand in other regional countries.
Even more impressive is the infrastructure the Chinese are building around Chongqing. There is the high-speed freight rail line to the port city of Shenzhen.21 The route run by a contractor, Cosco, takes roughly 50 hours and emphasizes “five fixed” services—fixed stops, fixed trains, fixed timetables, fixed routes, and fixed rates. Shenzhen is well equipped to handle bulk shipping to most major destination markets.
Next, with encouragement from HP, the airport at Chongqing was extended by 400 meters in a rapid construction project to allow fully laden 747s (with merchandise and extra fuel) to fly nonstop to Europe.22
An even more exciting development is that of the next “Orient Express”—a rail service between Chongqing and Duisburg, Germany, covering nearly 7,000 miles in 13 days. That is 26 days quicker than the current rail-sea combination, and considerably cheaper than the air option. And it should get even quicker as China shares its growing high-speed rail experience with countries along the path—Kazakhstan, Russia, and Poland.
Summarizes Prophet, “We are going to see significant competitive advantage from our investment in Western China. Chongqing is expected to produce 20 percent of the world's laptops and we are proud to be one of the pioneers here.”
Prophet says he is even prouder of the leadership role HP has taken with sustainability initiatives.
Sustainability
Prophet provides a number of examples:
We took a leadership role in forming the EICC (Electronic Industry Citizenship Coalition), which promotes an industry code of conduct for global electronics supply chains to improve working and environmental conditions.23
We have been transparent in naming our top 100 suppliers and disclosing the carbon footprint of our extended supply chain.
We have set goals beyond those required for any jurisdiction, including reducing BFR/PVC24 and accountability for conflict minerals.25
We discuss sustainability in the tech sector further in Chapter 16.
“A Great Operating Engine”
Ray Lane, Chairman of HP (and a partner at the venture capital firm Kleiner Perkins), says about Prophet, “I first met Tony as a rising star in the supply chain practice of Booz Allen (now Booz and Co.) in the ’80s. He was part of the brilliant team of industrial engineers and other geniuses that my partner, Dan Lewis, put together in our Cleveland office, which was then a mecca for logistical leadership. That early disciplined grooming continues to show in the amazing results Tony has consistently delivered for the HP supply chain over the last several years.”
Prophet is more modest—he says he is proud to be part of HP, which he calls “a great operating engine.” As he told the analysts in the presentation discussed previously, “This wasn't a single individual driving this. We structurally changed things. We put rigorous processes in place to make these things happen. So it wasn't a personality or an individual driving it. It's a company driving it, 300,000 people strong.”
At the Consumer Electronics Show in Las Vegas in January 2011, the dominant device was a tablet and HP seemed like a distant runner-up to Apple. The same show a year later was dominated by Ultrabooks, and HP looked impressive again, and smart for not having spun off its PC division.