The Oracle of Development: Allen Miner
Arguably the most successful foreign software company in Japan, in terms of sales growth, has been Oracle Japan. Working behind the scenes for over eight years to build Oracle Japan into a ·10.7 billion ($100 million) company has been Mr. Allen Miner ó a man little known outside his circle of acquaintances, but highly respected within it. Computing Japan interviewed Miner to probe his views on today's RDBMS market trends, and some of the politics behind the players.interviewed by Terrie Lloyd
Computing Japan: What is your position at Oracle Japan?.
Miner: I've had some interesting titles over the years. I finally figured out a couple of months ago how to simply characterize what I've done: it has basically been business development. I see my job as looking where we are right now, and contrasting it with whatever is the most important thing we have to do over the next year or two to take us to the next stage, then making it happen.
For example, back in 1987, I was involved in picking our distributor, and in deciding to have them do the localization instead of us doing it. That was a difficult decision. After it was made, I arranged to get the source code over here so that they could get started as soon as possible. I wound up working with that distributor for a couple of years to make sure things worked out properly.
More recently, I was involved with recruiting Sano-san, the president of Oracle Japan. Concurrent with that, I was in charge of establishing a strong local development organization to work with the OEMs and headquarters' development, so that we could provide better support and more integrated development. I have a dream that the local development group will someday build a product here in Japan that sells worldwide. The last couple of years have been spent getting the Oracle applications ready to roll out in Japan: the financial products and manufacturing products. We've just assigned responsibility for carrying this project forward to our Sales, Development, and Consulting divisions. I am now functioning largely as a coordinator for those teams.
Looking at the next step, we have some document automation initiatives that are starting in the United States. The multimedia stuff is interesting, and we've got some exciting development tools and PC software that need to be promoted here, but I haven't quite settled yet on what should be my next major project.
CJ: You seemed well connected and well informed of what is happening back in the States. Are you are a member of the board?.
Miner: I am a director on the board of Oracle Japan, but not the States. Actually, that is my other primary role. I spend a lot of time liasing with different senior managers from headquarters.
CJ: What is happening with Oracle's sales for this fiscal year?.
Miner: We are hoping for a 50% market share in the RDBMS market this year. Can we do it? Well, we went from 32% to 41% this last year; that is a 9% jump. So why not try for a similar increase this year? If we can do it, we'll go above 50%; that is our top-down goal for the financial year, which ends in May 1995.
CJ: Are you on track for the 50%?.
Miner: Until Dataquest does their annual survey for 1994, we won't really know for sure. However, we are growing over 65% so far this fiscal year, and Dataquest has been forecasting roughly 50% growth for the total market, so we are confident our share is increasing. But we don't know yet whether or not it is enough to push our share over 50%.
CJ: How much of Oracle's revenue worldwide comes from Japan?.
Miner: About 5% at present. When we got started back in 1987, we were doing about 0.7% through a couple of small distributors and some OEM business. We reached a milestone in the first quarter of this fiscal year, when we became number two (second only to the US) within Oracle worldwide in terms of our database business. But we're still behind in tools and consulting, after the UK, Germany, and France, for example.
CJ: It seems that Oracle has been focusing mainly on database marketing rather than tools and consulting..
Miner: We have focused on the licenses business and growing our share in the database market ó that's been the emphasis for the past four years. We also have been pushing through a high volume of business by leveraging multiple sales channels. We work closely with a lot of hardware manufacturers, systems integrators, and other business partners. Currently, almost 75% of our revenues come from sales with and through these channels.
CJ: So, thanks to the leveraging, your revenue per employee is high? .
Miner: We have much higher revenue figures per capita in Oracle Japan than Oracle does anywhere else in the world. There are about 13,000 people worldwide in Oracle, with $2 billion in sales. In Japan, we did $100 million worth of business with 380 employees this past fiscal year.
CJ: In an earlier interview this year, you commented that Oracle was selling a lot of 5-user licenses as customers bought them to experiment with. Are you seeing any progression from these customers to larger licenses?.
Miner: We're seeing lots of multi-site distributed systems going into production. NTT has just finished planning a new 600-site system, and this follows hard on the heels of a 700-site accounting system put in all their branch offices over the past couple of years. Mazda has purchased a couple thousand UNIX and PC Oracle servers. And several major banks are rolling out systems for all their branches. So far, we're seeing a lot of multi-site, small-scale distributed systems; but recently, we've begun preparing to go after high-end data center-type applications. We are seeing our software bid on symmetric multiprocessor cluster configuration machines from Sun, HP, NCR, NEC, and others more and more frequently.
It remains to be seen just how much we can penetrate the data center mission critical area here; however, it is a strong segment for us in the United States and Europe, and I believe it is just a matter of time before mainframe-level downsizing takes off in Japan. My prediction is that 1995 will be the year we see dramatic growth in our high-end business, as we roll out the products and the infrastructure to support more intensive use.
CJ: I've heard you have a new multiprocessor product coming out soon?.
Miner: Actually, we have had very good multiprocessor support in Oracle since version 6.0 (1988) for online transaction processing. It has a very good locking architecture for supporting multiple users, and running parallel processors in shared memory. What we've introduced with Oracle7 release 7.1 is support for breaking down a single SQL statement ó say, a complex query ó into sub-units that can be processed in parallel. For example, you could have a large data table, the bits of which are being read in parallel. Or, you could run sorts in parallel while reading the data. You could do index creation, backup and recovery, mass bulk data loads ó all in parallel.
So, while the multiprocessor version of Oracle is not new, parallelization of SQL and of batch processing is. What we also have that is unique in the database industry is the ability to share a single physical database in non-shared memory cluster and massively parallel configurations with fully automatic support for OLTP update and query processing across multiple distinct nodes.
CJ: What is happening to the market overall ó is it still growing substantially?.
Miner: Yes, it is still growing. Dataquest was forecasting about 50% for this calendar year. I'm sure that it is growing very close to that pace.
CJ: Is this growth at the expense of mainframe installations?.
Miner: It is probably at the expense of mainframe growth. We are still seeing steady sales of mainframes and ongoing use of them, but the rate of expansion is slowing. From our own revenue mix in the US and Japan, we are seeing UNIX-related revenue increases of 40% to 50% year-over-year, 60% to 70% for PCs, and declining revenues for the proprietary operating systems such as VMS, MVS, and VM/CMS. For our market (relational database management) at least, there is a very definite shift to UNIX and PC platforms.
CJ: So apparently what we have is a lot of companies implementing RDBMS solutions on UNIX. What is going to happen when Windows NT 3.5 comes out?.
Miner: People asked that same question when NT came out a year-and-a-half ago, and also when OS/2 came out five years ago. I am sure that NT will have reasonable penetration, particularly in the PC LAN area. But it will not turn the industry upside-down overnight, or take over the data center in this century. UNIX and mainframes will not be its victims. The main reason why UNIX is not threatened at this point is that a lot of expertise has built up around TCP/IP networking and UNIX systems administration. Those functions are coming in Windows NT, but there is work on parallel systems tuning and systems administration tools that Microsoft and the industry need to implement. It will take time. I think that there is a lot of momentum and a very rich set of solutions built around UNIX, and this will allow UNIX to continue to grow. In any case, the nice thing about working for Oracle is that we aggressively support all these operating systems, so I don't have to worry about who will win, or when.
CJ: Can Microsoft compete effectively against companies like HP, Sun, and others for corporate systems? .
Miner: That is a key question. There is a big service and systems integration component involved when corporations choose to use a supplier for their core information systems. Microsoft's success with NT may depend a lot on how aggressively HP, DEC, IBM, and others push it or fight it. Microsoft clearly doesn't have the support infrastructure or the approach to selling and supporting systems that would make a Japanese customer feel comfortable using NT as their primary data center platform.
So we have to ask: Can Microsoft build a sales and support infrastructure of their own, or find a partner that allows them to move up? I think that what we're more likely to see over the next two to three years is a continuation of the two distinct open systems markets: the PC arena ó consisting of NetWare, OS/2, and NT ó and the high-end systems UNIX segment.
CJ: Does that make Novell smart for having bought USL?.
Miner: I don't know that UNIX can ever dominate the PC LAN area. Their primary product for PCs ó NetWare ó is still the core of their business.
CJ: I believe that Oracle's three biggest systems integrators are NTT Data, NRI (Nomura Research Institute), and Enicom. How would you characterize each of these companies?.
Miner: That's a difficult question. Because of the strategic relationship we have, Enicom, the Nippon Steel subsidiary, is by far our biggest non-hardware manufacturer reseller and systems integrator. NTT Data, Nomura Research, and Hitachi Information Systems all take turns being numbers two, three, and four from quarter to quarter, depending on what projects come in.
As far as characterizing these companies, Enicom has done the largest number and variety of Oracle projects. Given their breadth of experience, they seem to like a conventional, risk-averse approach. In contrast, I think NRI may be more innovative. While NRI doesn't have the same breadth or range of experience with Oracle, their projects always have some novel twist and interesting use of technology.
CJ: Many of our readers are interested in what opportunities exist in the Japanese market. Do you have any observations about what gaps are still open?.
Miner: The category that Japanese companies are most interested in buying foreign products is in the technology and office automation area, as opposed to the business applications area. If SAP and Oracle are successful here with their applications, it would be a very new event in Japan. DBS (Dunn and Bradstreet Software) managed to sell over 200 licenses, which was a great success for a mainframe application company in Japan; but in terms of share of their worldwide business, it's tiny.
Most mainframe applications were custom-made, but the move to open systems might create a chance for developers of foreign application packages to be successful in Japan. I think it's more likely, though, that the gap will be filled by Japanese software houses and hardware manufacturers porting their existing office computer applications, and that the business application market will be controlled by domestic suppliers, as it tends to be elsewhere in the world.
CJ: So, would you say that there is a strong "not invented here" attitude?.
Miner: No, I think it's just that there are cultural differences ó there are functional requirements that are unique to Japan. The closer you get to running a company's day-to-day operations, the more difficult it is to build a package that is attractive to them unless you are also based in that country. For a company that specializes in business packages, it is very difficult to break into the small-to-medium-size corporate market here. I think the only chance you really have as a foreign application vendor here is if you can go after the global top 500 corporate accounts. I think Oracle and SAP can.
Now, on the technology side, there are many opportunities. I see open systems management tools as coming on big: companies such as OpenVision, CA, Legent, Compuware, and several others. There are some good products in the USA and Europe, and we are seeing some of these companies come over now. Another area is for MS Windows-based development tools, such as Oracle CDE, Powerbuilder, Gupta SQL Windows, and Business Objects (a company from France). Windows-based client/server development tools will see a lot of growth over the next couple of years.
CJ: These companies that you see as being up and coming: are they represented in Japan?.
Miner: Most have distributors in Japan. But few of them have made the decision to go out by themselves yet.
CJ: Should applications companies seeking to enter the Japanese market be looking for alliances with Japanese companies that have an installed customer base but don't understand how to implement client/server technology?.
Miner: They would have to have something really novel in their technology for the Japanese not to understand it. If their strength is simply in the depth of understanding they have of business functionality ó say, managing a hospital ó there are companies in Japan who have already been doing that for a long time on traditional platforms. The companies that you might want to partner with already have the business functional know-how, and most would be capable of acquiring basic client/server technology know-how themselves.
CJ: Are there any examples of applications specialist companies who stand out for you as having been successful? .
Miner: In the client/server arena, probably just one so far: Avalon. It is an Oracle and Sybase VAR (value-added retailer) that does manufacturing products. They did it by picking the best Oracle distributor ó NEC ó and the best Sybase distributor ó Itochu ó then giving them the source code and saying, "Here's the manufacturing product. Translate it into Japanese; do whatever you want with it." That approach of picking the distributor well, and giving them access to whatever information they need to be successful on their own, has worked really well. Long-term, I'm sure SAP and Oracle will be much more successful though, because of the strength of our direct presence here and worldwide.
CJ: Can you comment about Nihon Gupta? They have just appointed a new president and started to beef up their presence here in Japan. Is there any action going on between Gupta and Oracle?.
Miner: We looked at possibly acquiring them in July; we did a thorough review of their products and business. During the course of this, they decided that they didn't want to be acquired; there were some poison pills put in place that made an acquisition unattractive. I don't think there will be any further moves on Oracle's side ó apart from beating them in the marketplace.
CJ: I believe that there are some ex-Oracle people working at Gupta?.
Miner: Yes, there is a lot of competitive recruiting and "cross-fertilization" in the database industry. I'm sure the folks at Gupta Japan are enjoying the entrepreneurial challenges of running their new company, but with Oracle, Sybase, Informix, Mitsubishi Trading (the original Japanese channel for Gupta) , and Microsoft as competitors, Gupta Japan is not a company I would want to have to run.
.CJ: What are Oracle's plans over the next 12 months?.
Miner: A combination of things. We'll continue to aggressively pursue market share growth on low-end UNIX and PC LAN environments. You'll see a really strong push from us to go after the mainframe replacement business ó getting open systems into Japanese data centers ó using our parallel server technology. You'll see a big push on the high-end server side, with us putting in advanced support and testing centers to facilitate that. There are several exciting products coming out on Windows and Macintosh for applications development, all of them coming out in Japanese at the same time they're available in English. There will be a big push in the client tools area. It's sure to be another exciting, dynamic year.
