Discussion: New Opportunities for the Enterprise (ONLY KATETUTOR)

Lecture Text F. Warren McFarlan

The New Strategic Weapon: Information Technology

(edited for clarity) INTRODUCTION OK, can we get started? For those few of you in the room whom I've not had as a student, my name is Warren McFarlan. I was trying to think about what would be a good way of organizing today. Since this is my fortieth year on the faculty teaching Information Technology, I thought it would be useful to think about what has changed and what has stayed the same. There are some very interesting themes about stability and change. That's what I want to focus on today. What has stayed the same? First, the Luddites are still amongst us. A pivotal article written in 1966 in the Harvard Business Review by my mentor, who brought me onto the faculty, was called “The Myth of Real-Time Information.” It gave enormous comfort to the AMPs (Advanced Management Program attendees) in the 1960s who didn't understand one hell of a bit about what they were talking about. This gave them a reason not to. In the most recent issue of the Harvard Business Review, the first article is “IT Doesn’t Matter.” That is a very interesting article, and I will take you through it this afternoon. The words intellectual bankruptcy come quickly to mind. I have produced a four and a half page letter on some of the more obvious things. I'm meeting with the editor of the Review to see whether he wants to publish it or whether it goes to the New York Times. But it's going in one of those two places. This theme remains. What has changed? There is a whole lot more focus on how to think about IT as a strategic weapon, how to align this technology with the heart of corporate strategy, and how to push things back and forward. The example I use here is my newest book, which came out in March, called Connecting the Dots, written with one of my former MBA students, who is a senior partner on the consulting side of Deloitte. I'll talk about some of the contents of this work as we go on. We have a required course on IT in the first year. You can see different parts of Harvard have different views. The faculty and the required curriculum on one side and the soon-to- be-former editors of The Review on the other side. WHAT HAS STAYED THE SAME? The cost performance of technology The single most important point has been the continuous evolution of the cost performance of this technology. This slide, which I've used and updated over the last fifteen years, is the most powerful explanation for why this has been and will remain a turbulent area. My doctoral thesis was the largest, most expensive, and most useless simulation model ever built. It took a half an hour on an IBM 7094, the fastest machine of its time, to simulate six months of a garment factory. The 7094 took up a room ten times the size of this. It looked

Copyright 2003

The New Strategic Weapon: Information Technology

F. Warren McFarlan

like a $10 million machine. Hundreds of lights winked, tapes spun, disk arms went in and out. You could see and feel the value. By last year, all the power of that machine was in a seven-dollar chip. It has been the change in the cost performance that has enabled and enabled and enabled. “IT Doesn’t Matter” assumes that the game is over. We have zero evidence on that. All the consensus forecasts, in terms of material physicists, computer scientists, and so forth, suggest that last year's seven-dollar chip will be well under a dollar by 2007, and will come in at about one penny in 2017. Like all technologies, it will run its course. But we stand perhaps twenty years away from that. At the same time, of course, we have the cost performance of telecommunications moving even faster. Many of you noticed last week that our three regional telephone companies finally shook hands and agreed on a common standard so that we can deliver fiber into every home in the United States on a standardized basis. Copper, over the next thirty years, disappears completely, and it's all fiber, and it's all wireless. What that means is that you blow down a three-hour movie in a quarter of a second. Out of that three-dimensional arrays and a whole variety of things come to happen. Rapid technology change and vendor fragility The point behind this is simple. We are able to do things differently today than we were in the past, and tomorrow we can, if we choose, do things differently. Vendors come and go. When I started, it was IBM and the seven dwarfs. Then we went to the 1980s, and the heart of the minicomputer industry was here in Boston: Digital Equipment, Data General, Wang. We were going to, in fact, annihilate Silicon Valley. We were annihilated to the last body. We are now building ourselves on biotech and we hope that we've learned something from our earlier experience. Organizational resistance to change Point two: organizational resistance to change. In the fifty-fourth AMP, those days they put together a fat yearbook for the program, pictures of the faculty members, students, all that kind of stuff. Underneath my picture—in those days, the hair was a quite different color— was a quote that went roughly as follows: “The dominant issue in computer technology will be the ability to implement human behavior change.” There are very few things that I said in 1968 that I am willing to be quoted on in 2003. That is one of them. Everything this technology does challenges, and allows things to be done and organized differently, if we want it to. The irony of my life, trained first as a physicist, then as a computer scientist, is that most of my life has been spent as an applied organizational psychologist. Organizational change is as much an issue in the cases that I wrote in 2002 and 2003 as in the cases of 1966, 1967, and 1968. Stages theory Stages. My colleague Dick Nolan wrote a wonderful Harvard Business Review article in 1973 called “The Four Stages of Data Processing.” The initiative phase, the growth phase, the control phase, and then the maturity phase. It was a remarkable article. Dick then began to apply it to new technologies. We had stages five, six, and seven. Before I was able to shut him off in 1999, we were at stage thirty-seven. In fact, the notion of how new technologies as they come in require different handling at different points of their cycle, that has turned out to be a very robust finding that has lasted a long period of time. Thirty years after that article was written, it is as useful to today's chief information officer as it was at that time.

Copyright © 2003 - 2 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

Technical skill obsolescence This is a rotten field in which to grow old on accumulated wisdom. The technologies change. When we made our huge surge to the Internet at the Harvard Business School, from October 1, 1995 to March 1, 1996, we had an IT department of seventy-three people. How many of them do you think made the transition? Four. They were wonderful people that went on to the next phase of their career. Fortunately, the Y2K problem was all around and so they disappeared with dignity in that era. We then brought in a whole new set of language systems. One of the most interesting cases I wrote last fall was Pfizer. Pfizer is a very IT-intensive company. How many technology people do you think they have writing code at the pharmaceuticals division in Pfizer? Zero. It's all outsourced. They spend a huge amount of money in IT. They spend $1.4 billion. So we see that there's some commitment to the technology in a number like that. But the people at Pfizer Pharmaceutical understand the business, the human change processes. They bring in from a network of 200 different vendors the contractors, the packages, to work on different pieces of the problem so they don't have to get mixed up in worrying about how to keep those people up to speed on the new technologies. Different people do things differently in this area, and it's going to keep right on going. Legacy systems Systems developed in the past. They last far beyond any conceivable human belief. The first case study I wrote at the Harvard Business School in January of 1962 was on how to balance a line of punch card equipment. The core issue in that case was the quality control issue posed by dangling chads. Three years ago, if you told me there was any part of that part of my life that was still relevant, I'd have asked what kind of illegal substance you were puffing. Last November, I was meeting with the head of the Elections Commission in St. Louis, Missouri. We were chatting over a cup of coffee, and I said, “How's your conversion problem coming on punch card ballots?” He said, “What conversion? We've got no money for that. Ninety percent of the ballots in 2000 were punch card in the state of Missouri, and they're going to be 90 percent in 2004.” So when you tune in on election eve, 2004, and you're wondering where the next dangling chad problem is, you've gotten a clue today that those old legacy system run on in the heart of the Midwest. You're saying, “That's what I'd expect from the public sector. We from the private sector, we're better.” Down at the Aetna Insurance Company last November, how much Cobol code do you think runs Aetna that was written prior to 1970? Code thirty-three years old or older. One hundred million lines of Cobol or earlier languages, running along like Old Man River. Nobody knows how or why it runs. Legacy systems. They have been a problem throughout my whole career, and I think a lot of these systems will make it well into my retirement, which is not imminent. Efficiency, capacity, reliability, and security Caregroup Efficiency, capacity, reliability, and security have always been dominant issues, and they are front and center issues as we stand here in early June 2003. The case that I taught that made my students' hair stand up more than anything else this last semester was a company called Caregroup. Caregroup is the second largest group of hospitals in Boston. One-third of all faculty members of Harvard Medical School are part of Caregroup: Beth Israel Deaconess

Copyright © 2003 - 3 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

Medical Center, Mount Auburn Hospital, and so forth. They are the most sophisticated IT medical institution east of the Mississippi River. November 15. As a result of an unbeknownst-to-them hidden interaction of pieces of Cisco software, the entire hospital network went down for seventy-two hours. Nothing, and I mean nothing, happened for seventy-two hours. The physician expert system, which when a doctor prescribes a drug instantly sees what other drugs you're taking and makes sure that there are no lethal interactions, was gone. The emergency room entry system didn’t work. This case study is one of the most unusual ones I've written. The opening sentence on page one says, “November 21, 2002, John Halamka, standing in front of the Board of Trustees at Caregroup, said, ‘The good news is nobody died.’” This system made the front page of CIO magazine, the lead story February 15. As a CIO in a nonprofit healing industry, John wanted people to know so that others could learn and it would not happen to them. Eight hours after the system went down, a 747 from San Jose airport took off with two and a half tons of Cisco equipment in it, and twenty Cisco people came. Over the three days, they built the thing from top to bottom. That kind of operational vulnerability, backup, and security, these issues have always been there. In l969, there were a dozen companies in my case studies that went bankrupt because they lost control of the back office. On Wall Street today, we have the same operational dependency. We worry a great deal about whether it is up all the time. Is there adequate capacity? Is there too much capacity? Union Chimique Belge For the closing case in my course this spring, the students spent two days studying a company called Union Chimique Belge, a 3 billion euro company, UCB in Brussels. Total IT management case. I interviewed the head of computer operations. They had, of course, outsourced their old mainframes to Hewlett Packard in downtown Brussels, but there was this huge, great big computer room. I was looking at it and in it were 350 servers. So it was just as full of blinking lights as it was twenty years ago. Three hundred and fifty servers. I turned to the young woman heading it and I asked her, “What percentage capacity of these servers are being used?” She said, “I have no idea. Why is that important?” I said, “Based on industry expertise, general knowledge, I bet you've got 300 too many servers in here.” She said, “It couldn't be true.” I wrote a very hostile letter following my review, and the actual number you really needed was fifty-five servers, so I was wrong by five. We used to have controls over capacity. We had things called hardware and software monitors, and things of that nature. In the world of server farms and so forth, fanned by the server salesmen, they convinced us that they weren’t important. But as things are being used an average of 2 1/2 percent of the time, there may be some money to be saved on the table. The strategic grid This schematic I'm going to talk you through has been the heart of what I've been worrying about for the last twenty years. It is as important in 2003 as when I first invented it back in 1980, while flying back from the West Coast. I was right over Denver at 35,000 feet. I had a glass of a stimulant in hand and was wondering why it is that IT seemed to be more of a concern to some management teams than to others. I drew out on a napkin a version of this grid that somehow summarized what I was seeing.

Copyright © 2003 - 4 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

That actually has turned out to be very helpful today. There are two aspects of this technology that deeply influence how you worry about it, manage it, and control it. The vertical axis is what I call the strategic dependence on existing operating systems. How important is it that your networks run 99.999999 percent of the time with sub-second reliability? For some people, like Caregroup, that is the bare minimum. We need that absolute, utter reliability. Pfizer. How long does it take from the time you make an invention until you get the product to the marketplace? Eight to ten years. So we have a four-hour breakdown tonight of the networking service. In the big picture, how big a deal is that? No big deal at all. People, because they can't get on their machines, might actually have to think. There could be an intellectual brain flow by no longer data diddling. In general, we've been climbing and climbing and climbing up the vertical axis. In the world of the Internet, there's much more need to run operationally all the time. With that comes backup, with that comes grid computing, better security, things of that nature, as we climb up there. The horizontal axis: the strategic impact of the applications development portfolio. How important are the new software services in terms of what the company has to do to survive? Let's just take a very practical example: Pfizer. How important do you think information technology is to research? They spend 12 percent of their research on IT. Twelve percent. It is the core with which they differentiate and innovate. You give me more cycles, you give me faster bandwidth, and I can build more complex models. I can handle vastly larger volumes of data in a meaningful way and tease out important kinds of things. For last spring's AMP, I taught them ten classes on IT, and after that I then asked them, “Where is your company on this grid?” Forty-five percent of the class said we're up here in the upper right. We need virtually bulletproof reliability, and our development is absolutely critical, in terms of either being able to differentiate or to catch up and stay in the game. Again, that number is not 5 percent, and it's not 95 percent. This stuff is fundamentally more important at particular points in time to some organizations than others. Running down here in the lower left it’s 15 percent. Development is nice; it's useful. If you don't get to it for a year or two, it's not the end of the world. We compete on other bases, and if the network's going to a state of technical repose, again, it is irritating but it is not the end of the world. And on the diagonals, you have 20 percent up here on the upper left. Terribly important that it runs accurately, but not seen as a major source of innovation. This way of thinking has been around for twenty years. The names of the applications change as we move forward, as the new technology allows us to do different things. In general, we're climbing, climbing up on the vertical axis. But understanding where you are in this is absolutely critical. Leadership Finally, there has always been a need for leadership. Only the name has changed. The DP Manager became the VP of MIS, became the chief information officer, became the chief technology officer, and God knows what the next one will be. You can do this in a complex way. I said Pfizer has a $1.4 billion budget. Do you think anybody's responsible for it? Nope. My case is called “Pfizer's IT (The Virtual CIO).” It is run by a committee of eight people. But Pfizer has committees in many parts of their operation. And like George Orwell, there are many members of the committee and some are more equal than others. And it works.

Copyright © 2003 - 5 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

The way you make it happen can be more complicated than you think. But the need for direction and leadership, these are the themes that run out of the past. Applications Development Portfolio And finally, you have a portfolio of work. This thinking about the IT portfolio has been floating around since 1972. In the fall of 1972, I was up for promotion. In about five weeks, I was either going to be cut from the squad or be around to inflict myself on you for the next thirty years. I got a phone call from a senior associate dean and the committee wondering, “What have you been doing in your research for the last nine months? Would you on the weekend whip up a nice little report with your current research findings?” I was bitter. I came home and I told my wife, “I spent ten years at this place, and it all comes down to one lousy weekend.” My wife, however, is a rather warm, friendly, practical person. As she listened Friday night, Saturday morning, she said, “Warren, I'm going to take the kids off. We're going to go shopping and see a movie. Here are a couple of pads of paper. Why don't you see what you can make out of it?” With the fear of hanging concentrating my mind, about one o'clock I conceived the notions of company-relative technology, structuredness, and company- relative size, a language system for looking at the risk profile, and the contingency management of a portfolio that, thirty years later, is still a very important part of life. Those are all things that have been around for a long period of time. WHAT HAS CHANGED? Open standards and global interconnectivity First, open standards. For those of you who graduated before 1995, we didn't bother you much about the research project going on in the government, although it had been going there since 1969, starting with ARPA, which was in fact the predecessor to today's Internet. With that, the world changed. For those of you I had in 1994 and earlier, we talked about competitive systems. We talked about SABRE. We talked about American Hospital Supply. We talked about Otis Elevator, Frito Lay, but they were all closed, proprietary systems. If American wanted to control a travel agency, it was an American terminal in that travel agency. With the open systems, suddenly, with that protocol, you were able to interrelate anywhere with anyone. The simplest way to describe it is with the most useful application Harvard Business School has developed for me in the last year. It's called WebMail. Up until a year ago, for the last four years, anytime I went on a trip I had to carry the bloody laptop with me, and since September 11, worry about getting it through security and all that kind of mess. Just painful. The last trip I made—Hong Kong, Shanghai, Johannesburg, and Beijing—was to thirteen different places. Each time, either in the airport lounge or in a hotel business center—in one case a Harvard Business School office in Hong Kong—I just dialed on the Internet, typed in http://webmail.hbs.edu. On the screen appeared the picture of Baker Library with a little tagline for me to put in fmcfarlan. Underneath it, the tagline for me to put in my password. Hit the send button and two seconds later the last weeks of unanswered e-mail messages were in front of me. Simply answer them, hit the send button, and the response went out. If I didn't care about it anymore, I simply hit delete. And it went right around the world. There was no way that could have been done in the world before 1995. It changed. It changed absolutely everything as we came to this massive interconnectivity. Changing completely the way companies operate.

Copyright © 2003 - 6 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

The opening case that I used for my last AMP, and for my MBA, was a company called Li & Fung. Connie would know them, but not many. Li & Fung is a publicly traded, Hong Kong- based, family-founded company, 100 years old, $5.5 billion in sales this year, with 5,400 employees. Now, Hong Kong—when you think of the centers of global technical excellence, does Hong Kong make your list? Hell, no. Great place to get suits. Wonderful tours of the harbor. Trading. Money moving back and forth. Fung is Hong Kong based and they call themselves today a sourcing company, a trading company, where 69 percent of their sales are in the United States. Their customers are the 300 largest retailers in the United States. They have a network of more than 4,000 factories, scattered across China, India, Southeast Asia, North Africa, and the Caribbean. Sixty-three offices in thirty-seven countries. Three-quarters of their stuff in textile, soft goods, and the rest of it in hard goods. In 1995, if you had a quality control problem in a factory in Bangladesh, and you needed headquarters' opinion on it, the best you could do was to put some swatches of the offending fabric in a DHL package and send it on to Hong Kong. Forty-eight hours later, it would be there. You could then look at it, have a telephone conversation, and figure out what to do. By 1997, they had a completely browser-enabled Intranet that tied all sixty-three offices together around the globe. If there was a quality problem in Bangladesh, they simply took ten pictures of the fabric with a digital camera, put them on the Internet, and fifteen seconds later the pictures were in Hong Kong and you were able to have a discussion. A Hong Kong-based company doing this in 1997, a year and a half after Netscape got started. This is going on at world-class time schedules. That's my point. We think that in the United States, because we invented these major companies—IBM, Cisco, HP, Oracle, Microsoft, Sun, and so forth—we are also the leading experts in the applications of this technology. That is wrong. This stuff is going on almost simultaneously right around the globe. William and Victor didn't waste any time trying to renovate their IT department. They simply went to a local software vendor, outsourced the development of this stuff, and bang, it was up and running. By 2000, it was a complete extranet to their major customers. If the offending fabric was going onto goods for The Limited in Columbus, you simply hit the send button, on went the ten pictures to Columbus, and you could have a three-way discussion— time zones permitting—sometime in the next five minutes between Bangladesh, Hong Kong, and Columbus. Extended enterprise in an open systems world Open standards, global interconnectivity. At every step along the way, what we talk about today is the extended enterprise in an open systems world. Those ominous initials—BPO. business process outsourcing. The best case we have on that is Nortel. I knew what a company was in 1978. I knew what belonged inside and what belonged outside. Accounting departments are part of a company. It's been that way forever. What did Nortel do? They outsourced the entire thing to PricewaterhouseCoopers. They said they can do it better, the Coopers part of Price Waterhouse. They outsourced the entire back office of the HR department. They sent their data center, 2,300 people, to CSC. As in this networked society, they went back and took pieces of activity, peeled them off, and pushed them in other directions. All over the landscape, we see this stuff going on, creating efficiencies. It flows on and on.

Copyright © 2003 - 7 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

HEB and P&G My MBA students worry a great deal about the six cases I’ve written, including last fall, on HEB. HEB, the number ten grocery chain in the United States, standing for H. E. Butt, a privately held grocery company in San Antonio, Texas. What is of interest in this? For my MBAs, I started by asking them, “Grocery, is this a high-tech industry or a low-tech?” They said, “Warren, give us a hard question. This is as low tech as it comes.” They had visions of long checkout counters, kids yelling, dirt, everything floating around on Saturday afternoon. I'm going to take you on a quick, bing-bang trip through this. In 1989 who's their number one competitor? Wal-Mart. Bentonville, Arkansas. Wal-Mart—those of you who read the recent Fortune magazine article on them—America's largest company—will recall that in the heart of the article they described themselves in an interesting way. “We are a technology company. We are a technology company that happens to be in the distribution business.” On November 1, 1989, they launched a new partnership with Procter & Gamble. The day Charles Butt, seventy-sixth AMP heard about that, he saw the hour of the death of HEB. He saw that it's possible that Wal-Mart could get cost efficiencies that they had not been able to get before they could price predatorily, and they'd be stuck. Two days afterwards, HEB in Cincinnati struck a deal with Procter & Gamble to be the second company doing what they just agreed to do with Wal-Mart. The HEB system came up one week after the Wal-Mart system, so for all purposes, they were coterminous. Very simply, in the grocery industry, every product has on it a standard bar code, a unique bar code for manufacturer and for product type. Designed in the late 1960s, by the late 1980s it was just a normal part of life. HEB put a scanner on the back of every one of their warehouses and depots, and as a case of merchandise left the depot to go to an HEB store, that information was sent directly back to Procter & Gamble. P&G got a huge amount of data that they hadn't gotten before. P&G was told that we want 97 percent in-stock rate, in the source, in the depots, and you decide when to ship what. They picked a single price, the lowest price the previous year for the whole year, and with that the decision rights as to when to ship what passed from HEB across to Procter & Gamble. It was an enormous success. For P&G, it meant they were suddenly sending nothing but full truckloads to HEB. For HEB, they got rid of a department of people and drove their costs down. This thing unfolded between 1990 and 1994. By 1994, they had sixty-two of their suppliers running this. They had gone from 11.4 inventory turns per year to 38.6. The warehouses were three-quarters empty, except that because company had doubled in the meantime, they were only half empty. If Wal-Mart had had that cost edge, minus that inventory and so forth, they could have absolutely eaten them alive. By 1995, you've got that in place. This is built on shipments from depots to stores. What does P&G really want to know? Who's buying? What's happening at the store checkout counter? And what do you have at every checkout counter? Scanners. Between 1995 and 1997, live scanner data was picked up, packaged, and sent to Procter & Gamble. If you can tell Procter & Gamble when their stuff is being sold, what antisocial concept do you come up with? We used to order from you. We'd receive the goods. We'd pay you. We'd put the stuff on the shelves. People would pay us. That was the old model. The new model is that we're going to give you the privilege of delivering your products onto our shelves, and when they're sold, once a week we'll send you cash. Inventory goes negative. Three and a half days, that one even got Charles smiling. All good news. More and more efficient.

Copyright © 2003 - 8 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

And then October 1, 1999. George Shaheen, Managing Partner of Anderson Consulting, now Accenture, earning $4.5 million, resigns, becomes CEO of a small Internet home grocery store. The Wall Street Journal's Charles Reese says, “What the hell's going on here?” On October 8 Goldman Sachs raises $1.2 billion for George Shaheen. One point two billion to take this on-line Internet concept and drive it right across the United States, so people don't have to come to grocery stores anymore. It'll be delivered directly to you. What do you think Charles thought about that? It twitched the family’s third-generation CEO. It could die on his watch. HEB really manages their money tightly. I mean, they are cheap. They know where a dime goes. They spent $20 million to develop heb.com, to start December 14, 2000, to be able to beat this thing. In case a quarter of the grocery shoppers went that way, they wouldn't be wiped out. They watched, and they watched. While they spent, they watched. And George Shaheen and his company, called Webvan, ran through $600 million and $1.2 billion and nothing was making any money. He ran through another $300 million. Nobody was making any money. I suddenly stopped getting the Goldman Sachs investors reports about what a great investment this was. It was a core holding on January 1. By July 1, they'd never heard of it. One week before heb.com went live, they pulled it. They said we don't need it; the world has changed. Not 25 percent of the grocery shoppers are going this way right now, but .25 percent of the grocery shoppers. Some day, but not now. Extranet—Cisco In terms of reaching outside, another pair of cases I taught my MBA students was on Cisco. Cisco, who went from 1993 sales of $500 million to $20 billion in 2002, who went from zero business-to-business sales across the Internet in 1996 to $18 billion in 2002—65 percent of whose products they never touched, as it's an Internet-facilitated shipment from the supplier directly to the Cisco customer. Eighty-two percent of the customer service is done across the Internet. A company with 18 million internal Web pages. That's the first case. The second case is one that I wrote last year. Cisco China. Cisco China is a $1 billion operation of Cisco. How do you think Cisco China runs? Just exactly like Cisco U.S. My Cisco China case has four exhibits at the end, all of them in Mandarin Chinese, all of them in my newest book, Seizing the Strategic IT Advantage in China, which you can't read. Connie can read it, but none of the rest of you can read it. One of my Chinese students looked at me and said, “Warren, I don't understand this.” She's got the Chinese case in one hand, and the English case in the other. It's the same words. It's exactly the same. It's just in Mandarin. This stuff is running on a global basis. The applications are taking place in a global world. For Cisco, the most important part of that case comes at the end, when we talk about the IT department being an outward-facing department, not inward. From the time I entered the field, IT people thought of the customers as the inside part of the firm completely until 1980, and mostly until 1996. Customers meant the users inside. Today it is an outward- facing activity. That's what the Cisco story is all about. Their CIO spends 25 percent of his time dealing with real cash-paying customers, out in the field. Cisco is a node, a node in an interconnected global network. Legacy systems: Cisco Underlying this, however, are the problems of the past. What is the Cisco problem today? I attended a speech last October by their new CIO. Everything I showed you in the last two

Copyright © 2003 - 9 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

slides he confirmed. But he then went on to tell the group, “Our number one problem is the batch-oriented legacy systems of the past. Just because something comes up on a screen doesn't mean it's accurate up to the minute. If you update your inventory file once a day, by 10:00 at night it can be quite out of balance with what the reality is of your commitments.” And they're going back in a very hard, painful way, having to redo that kind of stuff. Clicks and bricks The first case I teach is about Charles Schwab, the most electronically enabled of our discount brokers, with 300 branches around the country. Now, sitting in my home outside Boston in these days of open systems, do you think I could open an account at Schwab from home? Absolutely. Dial in schwab.com, bing, bing, bing. Do you think, if I want, I can initiate from my PC at home an electronic transfer of a 401K from Fleet Bank to Schwab? Sure. Would I do it? Hell no. I need what I call a dog ceremony, lots of sniffing around down in a branch office. Do these people feel like my kind of people? Is this mafia north? Is this some form of a Chinese Tong going around? And if I like it, then I'll do business. To this day, 70 percent of all new accounts at Schwab are opened face-to-face by human beings dealing with human beings in an office. But 86 percent of their trades are done across the Internet. Clicks and bricks, drawing the line in different ways. Until three years ago, the lousiest Saturday of the year was the first Saturday in May. My wife would flip over the calendar and say “A-ha Warren—time for the annual shopping of suits at the Andover Shop.” That is eighteen holes of golf that will never be played. We walk up to the Andover Shop in Harvard Square. I see two fabrics in the window. I say, “If we can conclude this in fifteen minutes, we may yet get the back nine in before sundown.” Hell no, thirty-four more fabrics are brought up, including two from the basement, before, as a matter of sheer obstinacy, I insist on the original two. At that moment, Charlie, our friend the tailor, would arrive, greet my wife warmly by first name, say nothing to me, but simply slap a tape measure on my waist, look at a card in his hand, and say, “A-ha, a prosperous year indeed, I see.” My wife would say, “I told you so.” Non-vintage moment, as far as I was concerned. Three years ago, first of March, midnight at our house, everybody asleep, I dial onto the Internet to zenia.com, the online suit manufacturer. Click on into the fabric section. Look for fabrics almost identical to the ones I have. Some things I'm adventurous about, but not that. Hang the jacket on a hanger on the filing cabinet. Hunt through the styling patterns to find one that looks identical to the one I have. Then some requests for quantitative data. I gave them the real numbers, not the ones I merchandise around the house. Then there was a small credit card transaction. April 28, I come home and my wife says, “There's this big Federal Express package. What's this all about?” I said, “Good news, Karen, turn over the calendar: we have a 9:00 a.m. tee time on Saturday.” Different people have different value-added propositions, and resistance to change goes on in different ways. Literate users Twenty years ago, our users included in the AMP were illiterate in IT, and you knew it. Today, you're still illiterate, but you don't know it. Fresh from combat experience with desktop PCs, shopping on the Internet, word processing, e-mail, you're practitioners of the information age. But you don't know squat about what it takes to design a robust system that will run twenty-four hours, seven days a week, fifty-two weeks a year, handling all the odd-ball requests. There’s a wonderful case I taught my MBA students, a large project

Copyright © 2003 - 10 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

management case. About forty minutes into the class I turned to one of my students and said, “Valerie, how much do you think this is going to cost?” He said, “This is a complicated system. I don't think I can get it done over the weekend on my Macintosh. This feels more like a two- to three-week job, month at the outside.” The actual cost to the company: $16.8 million. Because you use it, doesn't mean you understand intuitively what it costs, what it takes to build something that runs reliably in a bulletproof, error-proof fashion, every step along the way. Planning: IT portfolio alignment In 1973, I talked about the applications portfolio. I talked about the risk of it. I talked about how different projects require different kinds of management processes. At the other end of my career, this new book is about how to do a better job of exercising and aligning your portfolio of IT against the real strategic needs of the company. The kinds of things we worry about here include chunking. Chunking Some of you come from organizations where there's great stability. Priorities don't change, and you can plan a long way into the future. Those are places where you can do two-, three-, and four-year projects. In other organizations there's huge volatility in budgets on a month-to-month and quarter-to-quarter basis. You start a three-year project, and a third of the way down priorities change—like the case I was working on this morning—and you have to write off everything spent, with no benefits. If you come from a volatile environment, you design it in little pieces so that if there's decommitment, you have something in place. Cultural overload Twenty individually justified good projects do not make a good portfolio. If eight of those projects all turn the same functional area inside out, the additive impact of it is simply undoable and you've got to go back and look at the portfolio and see where and how the totality is, in fact, being impacted. And that may mean half of what you want to do you can't do, or you've got to restructure it. Common threads The best way to describe this is with our friends at Pfizer. In the research department, there are twenty different proprietary databases that people can get into. To get into the proprietary database you have to have a unique password. That may mean that the typical person might have fifteen passwords. I'm only a professor. I'm not a line manager like you. You would have no trouble keeping track of fifteen passwords. It is beyond my ability. So how do you think I solve it? I have one password for the home security system, the vacation home security system, the bank account, the Fleet Asset Management account, the password to get into the Harvard e-mail system, etc., etc. Pfizer discovered the same thing. So what they do is you have one master password, you sign up, and they authorize you as to which systems you're allowed to get into or not get into, and every month they change that password. So you remember one. Going back and doing it smarter. The alignment of long-term and short-term needs If everything in your company hinges on meeting earnings for the next two quarters, and 90 percent of your IT portfolio is strategic projects whose benefits are not going to be in place for three to four years, there's something wrong, and you've got to go back and change it and put it back into place. This is what we're worrying about here, better ways of thinking about putting the plans together.

Copyright © 2003 - 11 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

First-mover, fast-follower Understand who you are. You may choose like Schwab did in their industry to be first, if you're large with many other assets like brand and financial pockets. Or you can wait like Merrill Lynch did to see which way the wind moved, and then be able to move with force and power. But again I come back to why “IT Doesn't Matter,” is wrong. In this world, the standard has gone way up. I wrote a pair of articles in the mid-1980s in the Harvard Business Review that turned out to be actually reasonably influential, talking about IT as a competitive weapon. One of my colleagues at the University of Pennsylvania a year and a half later came back and wrote a very insightful article criticizing my articles. He pointed out that most of my examples were first-movers in the industry. For those who came behind it was a strategic necessity for them to say either I'm going to stay in the industry and do it, or I'm not. Different people can play this out in different ways. If you’re not in the game, however, you may simply die. And it's in this world that all of these complicated infrastructure issues really matter. Outsourcing Outsourcing is here. It's a very, very complicated kind of thing. In Boston, for example, John Hancock, the oldest IT shop in the city, outsourced two weeks ago to IBM. Shut down an operation built in 1958, and off it goes into the IBM grid. And the software development goes on in a different way, as we continue to do an industrial restructuring, always with a massively higher uptime. TRANSFORMING TECHNOLOGIES What group of colleagues on our faculty do I find most useful? They are the group that stands the furthest away from me intellectually. They are our business historians who, in studying other past societally transforming technologies, have come upon some very interesting insights. The best of all the books is not one I've written, it's this one by Al Chandler, now in his mid-eighties, our preeminent business historian: A Nation Transformed by Information. How Information has Shaped the United States from Colonial Times to the Present. Al starts the information war in 1700. What was the problem the United States had in 1700? A deeply religious nation. Our ministers were frustrated. They only got twelve hours of airtime on Sunday. That was not enough to motivate the faithful. They needed to have assignments for them the other six days of the week. The only way they could do that was a massive national program of literacy. So literacy became a dominant U.S. priority. By 1750, we were the most literate nation in the world by twenty percentage points. Now, with this newfound literacy, do you think we spent our evenings at the fireside reading the Bible? Few outliers did. But the hotheads in Savannah said, “We don't like the English.” They created a pamphlet-touring industry. The sent pamphlets up to the hotheads in Boston and said, “See what you can do.” And we got excited, and we dumped tea in the water, and wrote our pamphlets and sent them back down to Savannah. And in no time at all, with the law of unintended consequences, literacy allowed a complete communication across the thirteen colonies, and the English went back home in 1783. When they left us in 1783, there were sixty-three Royal Post Offices in the United States. By 1830, there were 15,000. The post office and the postal system were the Internet of the first three decades of the nineteenth century. And then came the first of the great transforming technologies: the first line of rail was laid in 1831. Our network was built out

Copyright © 2003 - 12 -

The New Strategic Weapon: Information Technology

F. Warren McFarlan

Copyright © 2003 - 13 -

by 1910. Most of these technologies take four generations, eighty years to play out. And when was the biggest impact of the railroad? It was after the Civil War. Almost always it is in the second forty years. The first forty years you apply the technology to the world as you know it, and then you began rejuggling, transforming, and reshaping. And you can go right on with that with automobiles, electricity, telecommunication, and aerospace. IT: most people believe that the computer era began about 1955. That was the year IBM came out with the 701, 2, and 3, and Univac came out with the Univac 1. We're now some forty-seven years into this, and it is ironic that forty years into it, in 1995, there was this deep convulsive shock of the Internet in which how we applied it was turned inside out. You go back to my cost chart, and I look at my younger colleagues with some degree of envy. Because it is clear that my career has been spent in the least interesting part of the information age. My generation will be known for having created the Y2K problem, and happily the resolution thereof. Just when it's getting really exciting, maybe another decade, and I'm out of the game. This thing will run out of steam somewhere in 2025, 2035. But that's why I don't like the article, “IT Doesn't Matter.” It is our ability to wrap our hands around much larger databases in much more sophisticated ways, in faster transmission networks, that allows us to continue to change, not forever, but for us in this room, for most of the rest of our professional lifetimes. Let me simply say that it is an absolute pleasure. I will tell all of you that you haven't changed a single bit in terms of the alertness when I had you as students. I will make no comments, however, about hair color or amount thereof, on the assumption you won't make a similar comment about me. It's great to have you back.