What do Wal-Mart and Microsoft have in common?Besides being giants of their respective industries, they share a similar organizational structure?at least with respect to how each company approached integration, says Marco Iansiti, the David Sarnoff professor of business administration at Harvard Business School and author of Technology Integration: Making Critical Choices in a Dynamic World (Harvard Business School Press, 1998).While many companies outsourced their integration initiatives to Accenture, EDS, IBM or a host of other consultancies, both Microsoft and Wal-Mart created in-house teams (product managers at Microsoft; systems analysts at Wal-Mart) that focused on integration, not on one product or one process. Those teams are charged with knowing how all the technologies and processes that link their respective companies to their business partners function both inside and out. That structure allows Microsoft and Wal-Mart to be masters of their own fate and, Iansiti argues, gives them a distinct advantage over their competitors. (The first chapter of Iansiti\u2019s new book, Keystone: Operating and Technology Strategies in Business Ecosystems (Harvard Business School Press, January 2004), is tentatively titled "Why Wal-Mart and Microsoft Are Similar.")\nIansiti, who has a degree in physics, came to the Harvard Business School in 1989 after tiring of "being chained to a lab bench." The 41-year-old has spent more than 10 years studying the technology strategies of nearly 100 companies from all sectors, including retail, technology, manufacturing and the pharmaceutical industry. He has come to several conclusions about effective integration strategies and how proper integration translates directly into business advantage. He has seen companies handle integration well, and he has seen companies handle it poorly. The former, such as Wal-Mart and Microsoft, dominate their industries. The latter, such as Polaroid, end up in the headlines for all the wrong reasons. Features Editor Lafe Low sat down with him in his big Harvard office.\n\n \n\n\n\n\nCIO: What are the essential elements of an effective integration strategy?\n \n\nMarco Iansiti: What I\u2019m focusing on now are the external aspects [of integration]: How do you manage assets that are outside the company, and how do you integrate them? Microsoft has 40,000 business partners, and there are approximately 6 million people who build software on its platform. [Microsoft\u2019s] value is tied more to the integration with that ecosystem than it is with the company\u2019s internal resources.\nIt\u2019s as if you have these concentric circles. The first circle is the integration team, which could be a few product managers. The next, bigger circle is built around the core team of other resources inside the company. Then there is an even bigger circle around them that consists of developers, if you\u2019re Microsoft, or manufacturers or supply chain members, if you\u2019re Wal-Mart. The integration challenge expands from five people in a conference room to Wal-Mart\u2019s 50,000 suppliers.\nBoth [Wal-Mart and Microsoft] form a hub for dispersed networks of people. The companies\u2019 value is largely dependent on resources they do not own. Integration becomes not just integration of a small number of people inside the team or resources inside the company but the integration of a vast network of people or organizations, many of which are outside the company.\n\n \n\n\n\n\nHow do you integrate assets you don\u2019t own? What strategies are most effective?\n \n\nThe best strategy for the CIO is to keep technology couplings as loose as possible. That provides for greater flexibility. You can keep assembling best-of-breed solutions without committing to any particular external architecture.\nBecause of a combination of market and technology perspectives, the looser approach to integration is very powerful. There are a lot of technologies available, and vendors are trying very hard to sell them. At the same time, from a technology perspective, it\u2019s much easier to integrate without having to commit to any single vendor. For example, with IBM\u2019s On Demand initiative, it\u2019s very easy to try a bunch of new technologies without being committed to [IBM]. Just couple in to what IBM is offering, experiment with [the applications], then integrate them with your internal assets.\n\n \n\n\n\n\nHow can a company ensure that its integration strategy is on the right track, effectively translating strategy into action? And how can a CIO push this along?\n \n\nFrom the CIO\u2019s perspective, these should be fantastic times. Not in some ways, of course, but from a technology development perspective. CIOs have a huge number of external options, as well as an enormous amount of leverage for how to bring in those options, such as different technologies, different software applications and different consulting services.\nRight now, possibly more than any other time, CIOs are in the driver\u2019s seat. They have great leverage in terms of more technology choices and getting good deals. Obviously, CIOs are under a lot of pressure to cut costs. At the same time, it\u2019s a good time [to invest in technology] because you can get more for less. From a technology understanding perspective, it\u2019s much simpler than it was before because it\u2019s easier to link different technologies.The biggest challenge for CIOs is to prioritize what technologies to pull in. They need strong architectural expertise to figure out how the different choices fit because they are the ultimate integrators of all the external suppliers. There\u2019s a real opportunity to use this leverage right now, while the market is down, to push themselves in front of competitors that are a little too conservative. The CIO\u2019s role is primarily one of architecture and the process of integrating.For the retail study [done for Technology Integration], we looked at the technology strategies of several large retailers in the post-Internet daze. Walgreens potentially had the most integrated strategy. The company took a little longer but figured out how to best connect new technology with its existing assets.\nHowever, the number of organizations that went the other way and got slaughtered was huge (such as CVS, which bought and tacked on its online business externally). There is no external venture that was successful. Most?nearly 70 percent?were reintegrated and the rest were shut down. Integrators such as Walgreens have done well, while Hail Mary passers such as CVS have not.\n\n \n\n\n\n\nHow does implementing and maintaining an effective integration strategy translate into competitive advantage?\n \n\nIntegrating internal and external assets is more conservative [than buying or starting an external spinoff]. It\u2019s more efficient. In the retail study, we measured the difference in efficiency between integration players [like Walgreens] and the Hail Mary external spinouts [like CVS], and it was enormous. Differences of 3-to-1 were easy to come by in sales per employee or whatever productivity measure you\u2019d like to define.\nI\u2019m doing some work now on the anatomy of a transition. By the time people figure out how to adopt a technology, they also figure out how to leverage a lot of the old assets. The differentiation comes into the architecture. If everybody deploys SAP or everybody deploys Siebel, that\u2019s not going to be your competitive advantage. But the way you use it and the way you connect it to different systems?that keeps you unique.IBM is the ultimate example. IBM has reinvented itself many times. It was never a single idea but always this amorphous process. In some ways, what the company is doing now is the same as what it was doing in the 1960s. It\u2019s leveraging the Fortune 500 companies. That\u2019s still its basic model. IBM has gone from leasing mainframes to outsourcing and has done it very well.\nThe old saw, "Watch out CIO, technology is going to come to kill you, so you better move fast and do whatever you possibly can to react quickly" is gone. I\u2019m not even sure it was ever really true. I still haven\u2019t seen a company that\u2019s been slaughtered by technology. The companies that have done worse in the past 10 years have been the ones that overreacted.\n\n \n\n\n\n\nSo it doesn\u2019t necessarily pay to be the first mover in regard to new technologies?\n \n\nThere is no evidence whatsoever that first-mover advantage is good for you. Look at the numbers across industries. It\u2019s not first to move, but first to scale.\nYahoo wasn\u2019t the first Internet directory; Infoseek was the first widely used directory. The first auction site wasn\u2019t eBay; others tried to do it, but eBay got the architecture and business model right and scaled up first. Now that it is to scale, eBay closes 55 percent to 60 percent of its auctions. The number-two site [Yahoo] closes less than 10 percent.\n\n \n\n\n\n\nIs that part of what went wrong with Polaroid?\n \n\nWhat went wrong wasn\u2019t the technology. Polaroid has been dying a slow death since 1988. Companies get hurt for all sorts of reasons: poor management, inertia, they don\u2019t want to change the business model or the organization, they may have blind spots. Polaroid had a blind spot around marketing.\nIt made a bunch of mistakes, but you can\u2019t say that digital photography killed Polaroid. It had a lot of assets in digital photography. You could say it even overreacted to it. It was the company\u2019s inability to change to a business model that adapted Polaroid outside of its traditional products.\n\n \n\n\n\n\nHow has the relationship between technology and strategy changed during the past few years?\n \n\nThe biggest change is the evolution in couplings between technologies. What you see now is more loosely coupled technologies and the ability to exchange data without [requiring] a unifying architecture. We\u2019re starting to realize the power of plugging a lot of different things together, in contrast to having a single, monolithic architecture. We\u2019re getting better at managing technology diversity. You don\u2019t want to have a huge mess, but people are starting to learn how to manage a slight mess better and cheaper. Getting everything to work together is better than throwing everything out and starting from scratch. It\u2019s changing a lot of relationships. \nIf you look across industries, there\u2019s fragmentation. In the early days, one company owned most of the assets required to design and deliver a product. As industry evolves, there\u2019s more integration of assets from a variety of players. You see it in the car industry, the retail space, manufacturing or wherever. My students and I are spending a lot of time looking at that now in everything from PC manufacturing to semiconductors to pharmaceuticals to apparel.There\u2019s a [Hong Kong] company called Lee and Fung?it serves The Gap and The Limited?that could be considered a custom supply chain company. It gets an order and orchestrates a broad variety of external assets but guarantees delivery of the product at a specific time and place at a better cost than The Gap could get elsewhere. The company has approximately 8,000 employees, but it leverages about 1 million outside the company in its manufacturing base. It\u2019s like having a 1 million person company but without having to carry them on the payroll.\nThe interfaces between [technologies] have enabled industries to fragment, so the integration challenge becomes more broad. If you sit at the hub of these networks and manage the integration among the different pieces, you can enjoy tremendous leverage.\n\n \n\n\n\n\nAre there any reliable red flags raised when companies are going down the wrong integration path?\n \n\nA lot of companies get rid of their integrators [by outsourcing] because they don\u2019t understand their value. Architectural knowledge is foundation number one. I have good friends at Ford running global logistics. They\u2019ve outsourced so much of their process that they no longer have the expertise to understand how it all comes together. Leverage your resources as much as you want, but keep the integrators inside.\nNumber two: Don\u2019t think of technology integration as an administrative task. The challenge of managing the integration process runs simultaneously with the challenge of figuring out how to integrate. You\u2019ve got to understand how the technologies come together, which is not the same as putting a person with expertise in Technology A with a person with expertise in Technology B. We\u2019re talking about the cult of the architects. They understand the interactions among different system components and can figure out how to design the whole to best manage that interaction.The third piece is experimentation. We\u2019ve found people underinvesting in experimentation. Integration is hard. One way it blows up is when you finally pull everything together. Experimentation, however, is easier because many assets are external. Let\u2019s say you\u2019re thinking about whether to deploy CRM. You don\u2019t have to do this humongous internal project. You pick a target business to run a trial, get an On Demand version of CRM from IBM, and try it out for three months. If it goes well, you get a contract or bring it in internally. If it doesn\u2019t, you throw it out.\nIn every study we\u2019ve done, people who experimented more have always done better. \n\n \n\n\n\n\nDoes a company ever finish integrating? Can a CIO ever say, There, I\u2019m done?\n \n\nDo you ever hit the end of innovation? Do you know when it\u2019s done?\nWhat you want to do is build an engine that can do this systematically?scan the environment for technologies coming in, understand scenarios on the customer side, define the technology or product plan, and determine what projects to do. In companies that do this best, the process is like an engine. The method may have a time scale of a month or every quarter or every year where you resurface and select which [technology] you want to win the next phase. You pull it together, integrate and deploy, and come up for air again. It\u2019s a repeatable process. It\u2019s structuring the unstructurable.\n\n \n\n\n\n\nWayne Gretzky said he was successful because he didn\u2019t skate to where the puck was but where it was going to be. Where\u2019s the puck going to be for CIOs?\n \n\nThere\u2019s actually a Franz Klammer [Olympic gold medalist] analogy for that. When they clocked him skiing on any individual part [of the race course], he was not the fastest. The only part where he was fastest was coming out of the turns. He wasn\u2019t braking as much as the others. You can think of him as having a better view of the system, so overall, he was faster.\nIt\u2019s not so much understanding where the puck is going to be but having a view of the game that enables you to look outside and have the best understanding of how the game is played. It\u2019s always the system and how technology works within the system that makes the difference. Look at the global picture. Look outside and figure out where the business is going. That will tell you where the technology is going.