Rory Cowan
An Invisible Infrastructure for the Software Industry
Rory Cowan, David McGrew, and Gunnar Mallor talk about Stream International and Stream Japan, and their views of Japan's software and networking markets
Rory J. Cowan, CEO ofStream International Inc. andan executive VP of RR Donnelley & Sons Company, holds BA and MBA degrees from Harvard University. President and CEO of CSA Press when that company was acquired by RR Donnelley & Sons in 1986, he thereafter held such positions with RR Donnelley as president of thedocumentation services group,executive VP of information technology, and executive VP and director of theinformation resources sector.David L. McGrew is President, AsiaOperations, of Stream International Inc.
Gunnar Mallor is General Manager,Business Development,of Stream Japan KK.
Stream International Inc. was formedin April 1995 through the merger ofCorporate Software Inc. and GlobalSoftware Services, a business unit ofRR Donnelley & Sons. Massachusetts-based Stream International, with 7,000employees in 17 countries, achieved FY1995 sales of $1.7 billion from suchservices as software manufacturing,production, and product salesand support.
Stream Japan KK, formed on January 30, 1996, with 130 employees, expects salesof ¥15 billion in its first year of operation.PC software reseller and service provider Corporate Software Limited KK, a jointventure between Fujitsu and the former Corporate Software Inc., will playa vital role in the developmentof Stream Japan's business.
interviewed by Wm. AuckermanFirst, could you give us some background on the reason for the merger that formed Stream International?
Rory Cowan: There traditionally have been two sides to the software industry: a knowledge base on one side, and all the transactions that have to be managed on the other side. We realized that the manufacturing and distribution processes were really becoming one. The economic model began to change about five years ago, when bundling became an important channel for software distribution.
Global Software Services had developed the capability to take code in at one location, burst it out to plants around the world, and provide bundling services for all the hardware manufacturers. Stream now provides that service to about 100 OEMs [original equipment manufacturers] around the world, for over 50 products. We execute bundling with localization, interpretation, and translation. We can take a product and get it into the bundling channel worldwide within six weeks. Imagine 100 OEMs, 10 languages, and you can see the logistical challenge in that.
That's one operation. Corporate Software Inc. was a traditional product reseller, but it became clear that the customers were looking for more value-added services in the reseller market. Technical support, license management, and integration services were really wrapped around this transaction of selling software. We have been quite aggressive in investing in these areas; today, in the US, Stream has over 50% of the outsourced help desk market. One out of two help desk calls goes to one of our locations around the United States.
So you put these two aspects together to form Stream?
Cowan: Yes. We prepare and translate printed manuals, replicate and distribute magneto-optical disks and CD-ROMs; we handle 8 million help desk calls per year to help people with software, and we run online help services as well. We also have license management capabilities and response centers. If you buy a PC, you can register it and then we can transfer you directly to a help desk to talk you through the setup process.
These are all services that we offer through Stream now. Think of us as surrounding an industry, and facilitating the transactions from a publisher to the end user -- in a corporate setting, or a SOHO [small office/home office] setting as well.
Who are your main competitors?
Cowan: Within each of the segments we have competitors, of course. There are competitors who handle bundling services, though not to the breadth that we do, and those that handle the replication. There are resellers, like Software Spectrum in the US. And in integration services, competitors run the gamut from EDS to a small 5-man shop on the street corner. But while we have a series of focused competitors, there isn't a single competitor that has our business model, or the same range of services.
What are your initial plans for the Japanese market?
Cowan: We have two organizations here in Japan: one is a joint venture with Fujitsu, Corporate Software Limited, and the other is Stream Japan. Stream Japan provides support and manufacturing services for hardware manufacturers. Meanwhile, the joint venture and Stream work very closely in a number of other areas, such as support and product resale.
Stream Japan will have revenues of over ¥15 billion this year. We're growing at double digit rates that are equivalent to the growth of the industry. I know there are dog fights for share in Japan at the moment. Since we supply most of the major PC manufacturers here, we're derivative of that market growth.
Anything else?
Cowan: Yes, we also manufacture retail packages for software publishers. We are the manufacturer and retail manager for a lot of software products in Asia. We have plants in Korea, Taiwan, and Singapore as well, so we've built an Asian network that is able to move software into the channels quickly.
Are the majority of your customers here Japanese or foreign?
Cowan: We deal with both. Given that the microcomputer revolution is mostly US driven, in many cases it's the local country operations in Japan of the major computer companies, OEMs, and ISVs [independent software vendors]. But increasingly we are seeing companies focus on a global procurement program, so we might deal with US manufacturers and act as their transaction manager and infrastructure in 3 or 4 countries simultaneously. So, it's a little bit of both: we deal with a company's global coordinator in the US, and with their local country manager here as well.
How will forming Stream Japan affect the way you do business here?
Cowan: I think in a couple of ways. As I mentioned, we've put these two businesses together. One of the nice things about operating in Asia, and Japan in particular, is that we've learned so much in other countries where the computing industry is far more established, and everyone's roles are clear: you know what a reseller is, you know what a distributor is, you know what an OEM is. In Japan, as the industry unfolds, there is much more opportunity to define a new value space here. So we thought that if we got into the business as Stream, rather than as a reseller or a manufacturer, or as a replicator or a technical support group, that coming in as one bundled infrastructure for the industry gives us far greater advantage here.
Last year seems to be the year that personal computing just went vertical in Japan. We're very excited that we've come here at the right time, when the industry has enough scale and enough demand for the infrastructure services that we can provide.
What do you foresee for the Japanese market over the next few years?
Cowan: I think it will go through phases similar to what happened in the US, although -- I hate to say this about the industry -- where the US did the waltz, Japan will do the jitterbug. The Japanese market will move from individual computing to groupware and group messaging, and directly to network-centric computing.
That brings with it a host of support requirements, training requirements, and supply requirements. But it will be a very different model than in the US, because people will want their PCs to talk sooner. In the US, individuals purchased their PCs and had a year or two of the learning curve in private. Then they started going to online services, and finally went online at the office. All that is going to happen at once in Japan. There's going to be great opportunity, and great confusion, as the market tries to skip a beat.
It will be challenging for all participants in the industry, trying to satisfy the expectations of the new user in Japan. We're seeing increased sophistication and increased demand all at once here. Before, there were simpler, easier questions. Today, everybody wants to do it all right now, particularly with the corporate purchases.
Gunnar Mallor: It's going to be quite interesting as Japan tries to integrate both individual computing and client/server architecture, all at the same time. I wonder if the absorption rates of the corporations, of the individuals and the IT organizations, are going to be great enough to handle to it, because it is going to require a fundamentally different way of thinking about computing, communicating, and managing.
David McGrew: There's going to have to be a real change in the management practices of large companies. The way people think, the way they operate and work together is going to be greatly impacted. And I don't know if the management practices and thought processes in large Japanese corporations are going to be able to keep pace with that.
Cowan: That's what we've seen in the US. The glass house was preeminent in the '70s and early '80s, and then we went to computing anarchy in organizations. Now that we've gone back to network-centric and client/server architecture, the glass house is starting to reassert itself with standards. But, at the same time, you have to give a local option so that the individual is truly customized and has programs that reflect his needs.
There is a tension going on now, because if it's not centralized, it isn't effective for the organization, but if it's not decentralized, it isn't effective for the individual. The degree of customization and the degree of control in the Japanese organization will be the challenge -- to see if the right balance can be struck in the 18-month to 2-year transition period.
What kind of changes should the managers of Japanese corporations expect to have to implement?
Cowan: If you're going to make any organizational change, you have to change three things simultaneously: redesign the organization, add a new technology to make that new organization effective, and redesign the way you interact with the market. If you give a salesman a laptop computer, for example, he will communicate with his customers very differently. You eliminate a lot of the old organizational habits, because he's now online with customer service or manufacturing. So you have to redesign the organization, and you have to support the new technology.
Too often, large organizations just change one of the three because they're more functionally oriented; they're not looking at things horizontally across an organization. I think that will be the biggest challenge of introducing computing in large Japanese organizations: doing this fundamental process change with market interaction, with technology, and with organizational structure. And that's when productivity begins. Change just one, and as a CEO in 18 months you'll look back and say, "What happened? We spent all this money, and nothing improved."
McGrew: Another big difference is that the computing paradigm is fundamentally changed. It was very stovepipe. You went to Fujitsu and you got a total solution; if Fujitsu changed any one of the components -- the chip, the operating system, application, whatever -- they were responsible for making all of the other pieces fit together.
But when you start using desktop technologies more, you have a chip manufacturer, you have a box manufacturer -- you have 10 or 15 different companies interacting in that equation. So, who do you turn to? The paradigm has changed.
One thing to remember is that when you introduce a change in any one of those layers, it doesn't mean that anybody was communicating across the spectrum to make sure that everything remains compatible and continues to work. When you make a change, it really ripples through the layers and disrupts the operation.
Do you think networking and groupware can be successful in Japanese organizations?
Cowan: Groupware can do two things: it can either flatten the hierarchy, or it can give you more horizontal reach within the organization. Within the Japanese management process, which has far more consensus building horizontally, it may speed up decision-making if they are technologically enabled to do it.
Depending on the process and hierarchy development, though, groupware may be held back. We had a similar situation in the US some years ago: some supervisors resisted empowering the workforce because they feared they would lose control. They thought their job was to allocate work among employees rather than facilitate the completion of work. Those that made the transition to the latter model generally had a more productive environment. You'll probably have that same process in Japan over the next decade as the newer managers begin to facilitate teams rather than try to maintain their own stature.
That's the tension. I think horizontally, with all the new messaging techniques, and the Internet, groupware could actually find greater acceptance in Japan because there is more of a group thought process than in the US, where the flash of individual inspiration has been a traditional process. But I could argue it both ways.
Mallor: Initially, there are going to be some real challenges. The rate of adoption will likely be a little slower than we saw in the US, but I think groupware will take hold and be quite successful in Japan.
How is the software localization market changing?
McGrew: I don't know that it is changing, other than that the rate of entry is very high. We translate into most of the European languages, and we translate into the kanji-based languages; not just the written documentation, but we actually create the right code to help our customers take products to the market from the US to Japan. I don't see any market change, other than the time-to-market factor -- there's just more demand.
Cowan: We have a couple of software engineers located in Japan, but more to spec a project; most of the heavy coding, double-byte enabling translation work is done in China. Because of our relationships with major vendors in the US, as they contemplate their entry strategies for Japan or other parts of Asia, we can provide them with market entry services and manage the cost of entry. They can thus focus on selling and marketing while we handle a lot of the operational aspects: customer service and support manufacturing. Certainly translation and localization are big factors in that.
Why do localized versions of US software products continue to dominate the Japanese market?
Mallor: Maybe part of it is the process of developing software. Japanese hardware engineering is world class, without a doubt. But I think software development is a different pursuit and process. Some of it might be cultural, some of it might be thought process. I don't think too many organizations turn to Japan to find C++ programmers, for example.
Any final words about Stream?
Cowan: We've sort of been an invisible company; we don't put a lot of time into getting media coverage. Not a lot of people know about us, even though at our size and scale we touch almost every hardware manufacturer and all the major independent software vendors worldwide. You might say we're part of the infrastructure for an industry rather than a participant in the industry.