IDC:Providing Integrated Worldwide and Japanese Information

Philippe de Marcillac is Senior Vice President, Worldwide Research, for IDC. Previously in charge of the worldwide PC group of Dataquest, Marcillac has been involved in market research for over a decade. Although based in IDC's US office (Farmington, Massachusetts), he spends much of his time in Tokyo.

Interviewed by Wm. Auckerman

First of all, could you give us an overview of the IT research field?

Philippe de Marcillac: Well, there are quite a lot of companies who do something similar to what we do, in terms of providing information about computers -- everything from the press through to the research side. And on the research side, there are a number of relatively well-known companies, including IDC, which is probably the best for statistics, and Dataquest (now Gartner-Dataquest). Most other companies have a dual focus, typically providing information both to the user community and the vendor community.

IDC is focused almost exclusively on vendors, on the supply side. This is fairly unique, and it means our chart is quite different. Our clients are the IBMs, the NECs, the Hitachis: major computer vendors, but also the smaller ones, start ups who may be looking for information, as well as the venture capitalists, the financial houses, and software services companies. That's our focus.

Who are your major competitors in Japan, and how do you differentiate your services?

Marcillac: Like most of the regions we operate in, Japan has indigenous competition, which typically tend to be very small and very focused. We see our role as more globalist, providing integrated worldwide information and Japanese information. We have as areas of speciality semiconductors, systems, communications, hardware, networking -- and also personal computers, which is our flagship area.

Our major competitor, of course, is the Dataquest part of the Gartner Group. We differentiate ourselves by emphasizing the investment we're making here in end-user, demand-side research. Although we don't sell much to end users, our clients want to know what the end users think.

We also emphasize the links between Japan and our US operations, because a lot of our Japanese clients want to have information about the US. Our number one focus is on Japanese vendors, and providing them with world, Japanese, and US-based information.

Our second group of clients, of course, is the local subsidiaries of international companies, particularly the American companies. They are very focused on Japanese information, and like dealing with IDC because we also deal with their headquarters back home.

What is IDC Japan's ratio of general research versus commissioned research?

Marcillac: That's a bit hard to establish. In Japan, there are so few very large vendors. Sometimes we'll do a project for one or two of them; is it proprietary if we do it for one, but open if we do it for two? In the States, we commonly get 30 or 40 clients for a project.

We like a 70:30 revenue split. We really wouldn't like to have more than 30% or so of our revenue focused on one particular client, or unique sets of products, because such specific projects are very resource intensive.

Our focus is multiclient research, with 10 or 15 subscribers for a program. It's robust, and we learn from the subscribers about areas where we could improve or increase our coverage.

How do you typically gather your data?

Marcillac: It varies according to what type of information we're trying to get, and what region of the world we're trying to get it for. If we were told to come up with an estimate of locally assembled products throughout the Asia-Pacific region, we would do it in a certain way. And if we were told to do the same thing for Japan or the US, we'd do it in another way.

Japan is heavily focused on face-to-face interviews because of the nature of the society here. It is also a large, relatively homogeneous market with channels that are structured, and whose structures are well understood. We can gather a lot of information here by speaking to channels and large end users. And always, we crosscheck and validate our data with vendors -- because in a market like computers, which changes very fast, methodologies may occasionally go wrong.

What are your typical methodologies?

Marcillac: For the volume market, like PCs or peripherals, typically the methodology is to go and get information from the channels. We survey the market every quarter, because it's changing that fast, and we crosscheck with the vendors.

In the less fast-moving segments, such as the mainframe segment, we may have to speak to end users more, or we may just need to look for published sources in the press.

One discipline we do aim to achieve is to reconcile back to the stated manufacturers' revenues and their audited annual reports. We pay less attention to published vendor claims about how many of this or that have been shipped, because there can be some variation on how rigorous those numbers actually are. But once the books are audited, generally speaking we can feel this is a number that has some validity. And if the company reports a revenue remarkably different from that, we will want to understand why.

Just how trustworthy is the information announced by vendors?

Marcillac: In high-visibility segments like PCs, there's a very strong element in Japan of certain threshold market shares that companies like to have. If we relied solely on the vendors, we would be in a very difficult position, to be frank. There's no doubt that if you take vendors' claims and add them all up, this will come to more than seems likely to be the total market.

We do a lot of crosschecking, but there are times when that is quite difficult -- particularly if you're looking at, say, the PC market. Coming up with approximately correct numbers for the leaders is reasonably feasible. But estimating how much is being assembled by the very small or no-name players is quite difficult, even in a market like Japan.

And that affects the market share rankings. If a leading player is at a certain threshold, just above 40% or 30%, for example, we may encounter pressures not to make the other numbers too big, because it might take them below that strategic threshold. Fundamentally, what we do in that case is go and do some research to establish how large we believe this other part of the segment is and, hence, what the true market share is.

We've had to start doing that in Japan more, because this has been an issue here. Other parts of the world are less fixated on market share numbers; there is more concern about which direction things are going in, but much less about these "magic thresholds."

Do you face any special difficulties in gathering information in Japan?

Marcillac: There are unique difficulties to doing business in Japan in any segment, and ours is no exception. Some of the difficulties are that things are just different here: a different fiscal year, and things not necessarily counted or categorized in the same way. And there are different ways of segmenting the market, different thresholds that are used.

The major difficulty we have in Japan is getting quality staff, actually. There isn't really much of an industry here for analysts, so we don't tend to get many people coming out from vendors into our type of company.

What do you feel were the biggest changes in the Japanese PC market for 1996?

Marcillac: The biggest change has been the lack of change: that the growth rate of the PC market has remained very high. This is tapering down, but we expect even the 4th quarter to grow substantially from last year. Japan, for the second year in a row, is going to be the fastest growing major PC market by a significant margin.

The big drop in D-RAM prices, and the effect this has had on the industry here, is probably overall the biggest change -- PC market growth was a non-change. The industry was clearly over-ambitious in terms of goals and investments, particularly on the component side. For the PC vendors, this has been a boon, by the way. For a company like Apple Computer, the cut in D-RAM prices must have helped their return to profitability tremendously.

How do you foresee Japan's business-use PC market developing?

Marcillac: The business-use market is developing in an interesting manner. If markets were completely equal, if water found its level, you'd expect the Japanese market to be somewhere between 50% and 70% of the US. But we find that all the markets segments are far away from this norm.

The home and corporate markets are in the 30% to 35% range. The home market is growing faster here than in the US, while the corporate market is growing somewhat. The large systems market, for example, is actually larger in Japan then in the US now. But as corporations spend more on mainframes they're going to spend less on PCs.

The interesting thing is the virtual nonexistence of SOHO buyers, of guys working from home or in very small companies buying PCs. Because if you believe that the major economic factors in the US will follow in Japan, then you'd expect these types of companies to be an important motor for growth for the next phase of Japan's economic development. So it is a concern that these markets are still under-penetrated.

So, to answer your question, the business markets are developing well in Japan, after many years of delay, though still very under-penetrated in small companies. Large companies are going to have to go through some kind of reengineering as they bring on more and more personal systems, as opposed to large systems. But they have a big legacy of large systems from the bubble economy days, so they have to work out how they're going to do that. That's going to be an interesting thing to watch in the next two or three years.

Will networking in Japanese corporations continue to lag?

Marcillac: In the past couple of years, particularly in 1996, there has been a very significant breakthrough. This is being helped very much by Microsoft Windows NT, which Japanese users have accepted in a big way as a major opportunity for them to catch up and become more productive and IT intensive. Japanese users are clearly preparing for NT in a big way.

We're very bullish about the next two years in terms of the shipments of servers and network clients. In terms of installed numbers, Japan will still be behind the US because proportionally there are fewer PCs to network. But the trend is very aggressively toward catch-up in that area.

What effect will the Internet and intranets have over the next year or two?

Marcillac: A big effect. There has been an exponential rise of interest in the Internet. The Internet and intranets will be part of the drive for Japanese companies to increase the effectiveness of their computer infrastructure.

How do you foresee Japan's overall information technology market growing over the next few years?

Marcillac: It's going to grow about 10% for the next three years, compounded. Which is a lot faster than it has done.

Finally, on a personal level, what has been your biggest surprise with regard to the Japanese PC market over the past couple of years?

Marcillac: I think the biggest surprise for me was the speed with which Fujitsu managed to take the market share they did. Japan is not typically a fast mover in terms of switching brands.

And what do you foresee for the coming year?

Marcillac: Taking an internationalist perspective, there's going to be the big question next year in corporations: We have all this PC growth, so what are we going to do with them? Companies are going to have to do something to integrate all of these systems that they've bought. We'll see the start of the professional services industry in the PC.

The challenge for the next year is going to be assimilating what has happened in the past two years. If not, things will grind down to a slower pace, which would not be desirable.

One thing we believe is that Japanese companies and users need to continue investing in technology, because they have to become more competitive in the world economy.