History (1995): Exclusive Interview With Ed Zschau, GM of IBM Storage Systems Division

Since taking control of the Storage Systems Division of IBM Corp. 2 years ago, Ed Zschau has rarely been seen speaking to the press.

The GM of SSD took time out from a recent European press tour to speak with us.

CDSN: Why isn’t SSD a completely separate business?
Zschau: Rather than operating like a separate business, like our competitors, we are really part of the IBM organization, and drawing on the resources of the server division, and the marketing force and the industry and applications people, in order to address some of the storage intensive applications. This was a big change in thinking. You may remember 2 years ago, at which time I joined IBM. I did not join IBM. I joined an entity called AdStar, and I was announced as the new COB and CEO of AdStar. AdStar was this storage operation. And the concept was that AdStar might eventually be spun off as a separate subsidiary and maybe even a separate company. And after I had been in the company about 60 days, I had a conversation with Lou Gerstner and I said: “Our customers can’t understand why we think storage is somehow separate from information technology and why we’re talking about making this a separate operation, and two why we are trying to promote a new name.” And he saw things exactly the same way and the rest is history. We are the IBM Storage Systems Division and the point is that rather than being some separate entity, our activities are very much in support of the other areas that are required in order to deliver a total solution to our customers.

When you were recruited by Lou Gerstner, COB and CEO of IBM, 2 years ago, to take charge of SSD, what did he ask you for? What were your objectives?
Actually, I was recruited before he came to IBM. I was contacted in the earlier part of 1993 and made the offer to take the assignment prior to the time tha Lou joined the company. And then he arrived and I made the decision to join. Originally, the assignment was described as taking the potential that had been created by AdStar, with its technology, particularly magnetic and storage to exploit that in new markets like the OEM business. We had ignored the OEM business and one of the hopes was that I would be able to build an OEM disk drive company and take the product line into new markets.

Was there was never any pressure to become number one in the storage industry?
It was never described in that way, but I am very ambitious, and so is Lou Gerstner. He wants us to be the leader in the storage business. And even though, measured by total revenue, we are the largest, there are markets in which we don’t have the market share and we want to be the leader in all parts of the business in which we participate.

When do you think that you will overtake Seagate and Quantum?
Our growth is going to be limited by our ability to expand capacity and increase production. We could this year – I’m just guessing here, but we could probably sell twice as many drives as we’re able to make. Very high demand, very rapid growth of the industry and we are constrained by how fast we can increase. The other competitors are probably facing a similar situation. So I’m not able to answer your question when will we overtake them, all I can say is that we’re growing at the maximum rate and because our margins can be higher in those product areas where we have the uniqueness, the 2.5-inch and the high capacity, that’s where we’re concentrating most of our development effort.

What changes have you made since arriving at IBM?
The major invisible change has been to reorganize the SSD which operated as one large entity with a broad spectrum of products in a variety of markets, into 3 segments that are focused on specific products and specific markets. About a year ago, we organized a portion of the division into what we call the Storage Products Company. That is an OEM disk drive business that is vertically integrated, has its own development of components as well as products, its manufacturing and then its marketing. We did not really have an OEM marketing organization in the division, now we have a very effective one with channels of distribution including commercial distribution. In addition to organizing around the OEM business, we have also organized around our subsystems business. This business concentrates on the subsystems, not only attachments to IBM processors, but non IBM processors. That is another new direction that we’re taking. And then we have our storage management software business as a third entity.

Do you intend in the future to have your own sales organization in each country independent of the local IBM subsidiaries?
No. The way it works is that we have an OEM sales organization that we manage ourselves and that works with sales organizations in various countries. Secondly, in the subsystems business, there are people who have been identified as storage specialists. And these people concentrate on the sales of the storage products. They are employees of the subsidiaries and that is not going to change. The cost of the sales force worldwide is calculated and then allocated to us. We have the internal marketing expenses of my people and then we have external marketing expenses and we calculate our profit based on the combination of those.

Have you ever lost an OEM contract because you carry the IBM brand name? For example an OEM contract with a competing computer manufacturer?
It’s possible. I think that in the past that was more usual, but today people recognize that we’re serious about the OEM business. And we have OEM contracts with many of the IBM competitors, particularly in the workstation and the PC business. It’s based on two things, one is a recognition that our products are leadership products and two, a recognition that we are serious about the OEM business, rather than it just being a hobby, if you will.

How many people are working at SSD?
These are rough numbers, but at the end of 1992, there were about 15,000 employees and at the end of 1994, there were about 8,800. So there was a significant reduction. Now it’s still about 8,800 employees.

Do you want to continue to reduce the workforce?
It will be about constant, or perhaps grow. We are constrained in production, so as we expand our production, like in Singapore for example, we are hiring large numbers of people to expand production of disk drives. I would expect that the workforce has reached its low point and will grow, but as more production is taken over by Singapore, there will be less production in San Jose.

Is your division profitable?
Yes, in fact in the first quarter meeting, it was announced that we were one of the contributors.

Do you intend to enter the CD market?
I don’t expect that we will. There are other companies that are much better in that business.

Even rewritable CD?
We think that the rewritable CD will be made on the same base, and the people who are in the large volumes will have that base. One secret to high profitability is to focus on where you can be the leader and dominate and have a distinct advantage. And two, it is to have products that are a gen ahead. So you can get a high margin before they become commodities and then have the next product.

Don’t you want to be in the commodity product for HDD drives?
Commodity being the ones that everyone can make? No. We want to be the ones that nobody but we can make. And that’s why we focused on the 2.5-inch drive. That’s one where our areal density advantage is important, because we can fit more capacity in that small space. And in the high-capacity 3.5-inch drive, again we get an advantage from our areal density.

Do you believe in the IDC’s prediction that EMC will ship more terabytes on DASDs for mainframe computers than IBM this year?
No, I can’t say that. I’ve looked at those numbers and we would have to stop selling mainframe DASD sometime in the late summer for those numbers to be accurate. They have underestimated what we are going to ship this year. And that’s not surprising because we don’t break out numbers, so they don’t know the details of our shipments.

At a time it was said that “Anything from IBM can be bought.” Why do you refuse to sell MR heads?
I’ve changed the policy. At the time I arrived, there was that statement. But I’ve modified that. For example, when I arrived there was excess capacity in our MR head manufacturing and we were talking about selling MR heads to other people. Now, that’s one of our constraints. As we increased our success in the drive business, we consumed many more heads. And now we are consuming all the heads we can produce. We feel that we can get a better return on the head investment at the drive level, as opposed to the head level. So the situation has changed from a company with excess capacity looking for markets to a company that is capacity constrained and growing as fast as we can.

Which other components do you refuse to sell?
Actually, we’re not pursuing the sale of either disks or heads. The components that we do have for sale are the disk drives, but even there we’re not able to go after new customers as aggressively.

Do you buy heads or platters outside?
We buy a limited number of disks outside, some 2.5-inch glass disks from a Japanese company. But all of our heads are MR now, and we don’t buy those from outside because there are no reliable supply sources for MR technology at the areal densities that we’re using.

Are you going to continue to let PCM manufacturers copy your storage products (controllers, DASDs, tape drives) in the mainframe business?
In the whole corporation, the emphasis on patenting IP has increased, and the approach to licensing changed from a defensive approach that is attempting to get cross licenses for freedom of action, to licensing for royalties in order to get additional returns on the investment.

How many patents do you have in storage technology?
Roughly about 1,200.

Sometimes, it seems that you have the same strategy as Big Blue had some years ago: even if we have a better technology, we are not using it right now, waiting for the competition to approach your current technology. Is it true in the storage activity?
No, not at all. I think the MR technology is a good example of that, where we’re driving it as fast as we can and using it in all of our products.

Today, if you could put 1GB on one platter, would you build it?
We will build it if we can. But there is a difference in our philosophy that we will reuse parts. That is, us common building blocks to reduce risk, because product development cycles must be 1) more reliable and 2) they must be shorter. So, in the past when there was a new drive design, we would design a new case, new spindle motor, new channel. It was all different. Now we try to limit the number of changes. So we use the number of heads and disk technology in several drives to reduce the risk and compress them.

When will we see Giant MR?
It may be 3 or 4 years, not because Giant MR will take that long to perfect, but as we just announced the demonstration of 3Gb-per-square-inch recording. The most important aspect of that demonstration is that it used refinements of conventional MR, which indicates that there’s a lot more improvement, growth in technology from the conventional MR.

Are you losing your lead in technology? Toshiba is in front of you (1,35 vs. 1.2GB) in 2.5-inch drives, and Quantum and Seagate are not so far behind in 3.5-inch with 8 to 9GB (IBM at 10)?
We’ll see who’s first to the marketplace in volume production. As we have learned ourselves, it’s far easier to announce a product than to produce it.

This article is an abstract of news published on the former paper version of Computer Data Storage Newsletter on issue 90, published on July 1995.

Note: Zschau then was an investor in and board member at 4 tech-based companies, and the interim president of Sierra Nevada College. He and his wife, Jo, currently reside in Nantucket, MA, and Los Altos, CA. They have 1 son and 2 daughters. He is skilled in playing the ukulele and was known to some as the “singing congressman.“