<SPECIAL FEATURE>
March 23, 1999

TMF Interview With
Network Appliance CEO
Dan Warmenhoven

Part 2

Dale: What are some of the other appliances? Sun boxes?

Warmenhoven: Sun's actually withdrawn theirs. They had one they called a Netra and it didn't perform very well in terms of revenue, so I think they've terminated that one. In turns out new entrants and new exits every day. Sun just basically withdrew the Netra and along comes a company called Procomm and tried to do an imitation of ours. So it's a very dynamic market. We've obviously proven there is a market there and that there is an advantage to an appliance approach, and we've had very good business success as a result and that draws imitators.

Dale: In the coming year, how do you gain more share in the $9 billion dollar market?

Warmenhoven: Part of it is expanding sales capacity. Let me back up. It starts actually in creating a preference in the customers' mind for an appliance. Make a simpler, more reliable, more cost-effective, and faster method. And I think you see if you look at the industry press and all the rest, a growing emphasis on appliances as preferred solutions. You've probably seen the Oracle Raw Iron database appliance, you hear Hewlett-Packard talking about appliances. So step one is getting the customers in the marketplace familiar and comfortable with the notion that appliances are a superior solution. And in our case in particular, build a sales capacity as fast as possible. And that's both hiring direct sales people as well as some of the OEM relationships such as Dell and Fujitsu and indirect channel partners.

And then continue to broaden the application and partner base. For instance we just added Sybase within the last month or two. And that opens up another segment to us. Microsoft just put some endorsements on Sequel server on their website in December -- you know, that opens up another segment to us. So we're working with Microsoft and certifying their applications on Exchange server kinds of things. We're working with other application developers such as SAP, for instance, to run on top of Oracle, so we're building up, if you will, the certification list so we increase the comfort level of the customers that this is a safe approach to handle their data.

Dale: How soon is it before you can be in a position where you're an infrastructure standard or where you can really shoot for the goal of being an infrastructure standard to the point where people say about NetApp what they might say about other companies: "You can't be fired for ordering Cisco, or you can't be fired for ordering IBM." How soon is it before you can get to that position?

Warmenhoven: It varies by sector. I think we're there in the semiconductor sector. And I think we're there in the ISP sector.

Dale: In talking with one of the techies, he said there's a difference in how your things perform on write applications versus read applications. Is the semiconductor field a write application-intensive task?

Warmenhoven: First of all, I guess I disagree with your technical guy or maybe your interpretation of what he said. This is a write-optimized file system, but it's not read disadvantaged. The more write intensive the application the better we do, or the bigger the difference is between us and the alternatives. The ISP applications of mail and news and so on have a tendency to be very write-intensive. And in the ECAD world, you know, for chip simulations and things like that, it's also fairly write intensive. So if anybody should have a high volume of writes, we're going to blow the doors off anything else you consider.

Dale: So you think that's your competitive advantage when certain customers have come from that so far?

Warmenhoven: No, you know, it goes along the following lines. In general, our response time is either reads or writes, I mean the whole next application mix generally runs about twice as fast as local disks, that's on an average mix. So it starts with raw performance and people get enamored with performance, but the thing they really like is the combination of the availability and the ease-of-use features. It's hard to explain how simple something could be to somebody because you have to describe the absence of things they've always taken for granted that they have to do. Booting up you PC, right, you just assume you have to do that. This one just operates itself. I think of it as a data refrigerator. You know, it's easy to install and once you install it you don't do anything to it for ten years.

Dale: You know, I think in general, many people have difficulty with the concept of opportunity cost.

Warmenhoven: In this case, trying it is actually what gets people enamored with it because it's so simple, and all of a sudden they find they have all kinds of free time. As you put it, there's an opportunity cost to maintaining what they have when they can be doing different things. And that's what makes them deploy it in large volumes. Availability and simplicity are the things that make for the widespread volume deployment.

Dale: And that's pretty much what you attack these four segments with in the coming years?

Warmenhoven: Generally we get in the door because people heard about our blazing performance, but it's really availability of data, not just availability of the systems, and the simplicity of the use.

Dale: Now, I'm wondering with the growth rates that I've seen analysts forecasting, you guys will have a lot more sales dollars that you can translate into operating spending and you're going to be generating more free cash. Where do you go in those categories on investments in coming years?

Warmenhoven: Let me give you some background business model. In general, we've attempted to spend between 25% and 27% of revenues in sales and marketing, and that allows us to continue to increase our sales capacity -- meaning hiring a direct sales force and the building of partner programs and database specialists and things of that nature. So we're investing at a pretty high rate and just expanding sales and marketing capacity. So as our revenue grows, basically we're just going to plow that back in. We're not going to take it in to profits, we're going to take it into building more capacity. So specifically where that goes [is] a function of where we see the emerging opportunities.

For instance, we invested very heavily in Europe last year and through the first half of this year, and then had a big payback in this quarter, where Europe jumped from 19% to 33% of total revenues. That was a very premeditated kind of strategy to spend most of our incremental investment dollars by building European staff and channels. And now we're starting to do the same thing in Asia-Pacific. I really believe that the Pac Rim is going to come back, and while it's not big yet, I believe it's coming. So, does that answer your question?

Dale: Yeah, pretty much. I mean a lot of people look at -- I mean you're not traditionally a capital intensive industry. I mean your investments come mainly through the income statement and your investments are expensed right away, whereas if you're a car manufacturer you don't see it investments right away on the income statement.

Warmenhoven: And you see very little on our balance sheet as capital equipment. If you looked at our factory, you'd be amazed. It has no capital equipment and no people. Most everything's done outside.

Dale: That's just what I was asking, and you answered it. As you have more sales dollars coming, you have more leeway in what you want to spend on and choices that you want to make on how to spend those sales dollars.

Warmenhoven: That's right. And the same is true in R&D. I mean our business model in general has been to try to keep the gross margins at roughly 59%, and about 40% of revenues goes to expenses and roughly 19% to 20% goes to the bottom line. And the expense structure is typically about 26% of sales and marketing, 10% R&D, and 3-4% G&A , and that gets you 40% and that gives you about a 20% profit drop.

Dale: I noticed on the conference call you were talking about how different sales people concentrating on different categories will introduce a specialist to a lead that they've discovered in one of their accounts. Is that something you set up consciously [to get] your sales executives to share leads and share accounts?

Warmenhoven: Absolutely -- as we've expanded into different markets and different applications and different types of buyers. The area we're addressing is so broad that it's difficult to have applications expertise and even technical expertise, so we have in fact built up Windows specialists, database specialists, and NetCache specialists, and the account execs are encouraged to bring those people in to help make a sale. Rather than having to split the commission we actually said we'll pay the commissions to both.

Dale: Full commission on both?

Warmenhoven: Full commission on both.

Dale: I'm not a real veteran of looking into high tech companies, but from what I've noticed of a lot of companies in high growth phases, a salesperson starts calculating next year's commission and can get greedy.

Warmenhoven: He can get greedy. In fact in our case -- that's not to say our people are greedy, but we are in a very high leverage comp[ensation] plan. Basically the structure of the comp plan is on the base salary, that might be 30% of the goal. And if you don't sell anything you probably can't eat. You can't live on that kind of base salary. You've got to sell. And it's all done through the commissions. So yes, it's highly leveraged and a by-product on that issue is that you tend not to want to give deals away.

Dale: Yeah, that's just human nature. I guess I shouldn't say greedy...

Warmenhoven: Well, it's just pragmatic. I mean it's their income dollars you're talking about. And so basically what we've said is OK, bring in your partner, bring in the specialist to help win the deal and you'll have no penalty. You'll get full pay.

Dale: That's interesting. Can I get your thoughts on the quarter that just ended [February 17]? Your sales were up 72%, net income was up 69%, and EPS was up 55% quarter-over-quarter.

Warmenhoven: I was actually very pleased with the overall result. The growth in revenues was driven by Europe, which jumped from 19 to 33% [of sales] in one quarter. Europe was very, very strong, especially in the first half of the quarter. There was a lot of what I would call year-end buying from European companies and that turned into revenues through January. I didn't mention the multi-protocol segment earlier, but one of the big segments for us is the segment where we're providing service for both Windows and Unix machines from the same data.

Dale: I was wondering about that.

Warmenhoven: And that grew to almost 45% of the total mix. There was exceptional growth here. And Windows-only business grew from 3% to 8% of sales. Net cache grew from 5% to 8% of sales. So all of the areas -- and database grew somewhat from 7% to 8%, or 6% to 7% or something like that, I forget the exact number, but all the major markets we've been trying to move into have all shown positive results. So overall I'm very pleased that the results of the quarter are in alignment with what we'd hoped to have from a strategic objectives viewpoint.

Next: Part 3 of the Interview

</SPECIAL FEATURE>