Watts got his sights filled with that view at the recent HP World show. He spent his time on the management roundtable there saying the right things and listening, hearing the discomfort of the 3000 faithful -- for what he said was the first time. "I've learned a lot," was an important statement to emerge. I spoke with him immediately after that session to find out how what he learned might help the HP's 3000's future. I was most interested in what he had to say about putting technology at the center of an information structure.
I would say it would be very high on my priority list. I heard that loud and clear. Frankly, I don't think it would take a massive amount of money. It would just take enough ammunition to have have our champions carry the message for us. If they're staring at a black hole, it's very hard for them to build a credible story that HP's serious about the market. I don't think it's going to require a blot-out-the sun major advertising campaign. I've very encouraged that we can find some creative ways to achieve that.It would seem like including MPE in the operating environments listed in HP's advertising isn't even a cost item, just a policy change.
The HP 3000 division has really impressed me. I hope people see there's a team of people there who are really dedicated to customers. When I first walked into CSY after I got this job, I was looking forward to my day. I thought I'd get a review of a division that feels like the poor stepchild and they don't get the attention, they're not front and center or growing as fast as our Unix business. I thought I'd hear a lot of grousing and complaints and see a lot of morale issues. It was an absolute delight. They are enthusisatic about their jobs and the customers. Yes, they're frustrated and they'd like to do more, but so would the Unix team. There is more energy and attention on the idea of taking care of customers whatever their future in CSY. I think it's Harry's leadership, and before him Olivier [Helleboid] and before him, Glenn [Osaka]. Those are three outstanding managers, and I'm delighted to have them on my team. You'll see that leveraged into other parts of CSO, to take care of customers that way.
Clearly, we've got to get to market with that. We sometimes say enterprise computing, which we think covers [the HP 3000] -- which it would do if there was a perception of it being an equal partner in the product line. If you've got a perception that it's not included, then by not explicitly including it people think there's just more evidence that it's being excluded. We have to give it a kick start, and then a lot of other things will work.
HP recently exited the disk drive business, the first wholesale computer product exit I can recall in the 12 years I've covered HP. Why does HP think the commodity systems markets like HP-UX, and especially Windows NT, will be any different than disk drives? What has HP learned from its disk drive failure?
There has been a lot of parallel learning. We have learned in both disks and the platform area that we cannot afford to run our business strategy on technological excellence along. We also have to recognize the economic models, business models, the channel and the fact that there are different ways of attacking the information systems needs of our customers.
Where we can, we want to leverage the capabilites that are available to us in the industry at large, which tends to be in the commodity space. Buying drives from Seagate and adding lots of value in the way those drives are managed and how the backup and other storage management software is developed seems to us to be a lot better way to bring real value to that equation.
Clearly, our more balanced approach to Unix and NT, now that they're positioned in our enterprise offering, reflects a recognition of what NT brings to the party. That's an economic model that's not overwhelming, but is very powerful: the volumes of software available, distribution channels, numbers of units being shipped on both the hardware platform and the operating system. These bring a certain economic equation that a customer would be very foolish to exclude. They may not choose to make it their only platform, but it clearly has a place.
This was a pretty painful move, particularly when you're dealing with a large organization. It's the right thing to do. The laws of economics and the nature of the disk mechanism manufacturing business are such that the offshore countries have produced dramatic gains in manufacturing costs -- because of very heavy investment by our competitors in building the infrastructure and skill levels in those countries. We can't compete with those costs in Boise, Idaho.
It's also a recognition of the fact that the disk mechanisms are no longer the key differentiator in what makes a good storage system. As long as the industry can provide adequate capacity, size, performance and speed, then the opportunity for us to differentiate is much more in what we wrap around that disk -- building the server capabilities into and around the physical storage medium.
This was an exception. I think it's a reflection of a clear, heads-up recognition that we were behind the power curve on investments necessary to build a very competitive line. There are a lot better things we can do in the storage area to leveage our strengths.
The Storage Solution Division moved from the Mass Storage Group to my group. That's where we think the opportunity is: to make storage, whatever the medium is -- Seagate or Conner drives, tape or CDs. We want to make that storage system a very integral part of our server offering.
Given the lifespan of computer products compared to others, what's the business case for putting more resources into lines with higher costs and lower margins? Has HP boarded a commodity treadmill by dipping so deep into Unix?
No. The two interact differently than what you've implied. What generally happens is as you move more to industry standard "commodity" platforms, your costs go down because you're leveaging other people's work. You pay Microsoft and Intel a lot less to buy an operating system or an X86 chip than it would to hire 500 engineers to invent a Windows operating system or run a chip through a fabrication facility. But because everybody can play in the same game, your margins go down, too. The more differentiation you have, margins and costs go up in proportion. That's the challenge -- to make sure the value added in the Unix or in the MPE space is perceived as sufficient value by the customer so we can recover the incremental costs we've incurred in adding that value.
The real challenge as we look to running our business in the future is the way that value is measured by the customer. It changes continuously. If you're in a more differentiated Unix world, you establish value relative to people who play by very much the same rules. What we're seeing with the increasing functionality of the commodity platforms like Intel boxes running Microsoft software is that customers have a third reference point. It's not just you and your competitors; there's another level that constantly rising which is what I can get at a significantly lower price point because it's mass market, high volume, low margin. Our challenge is making sure we add value relative to the costs we're having to charge -- relative to both the industry standard model and then win against the IBMs and DECs.
What's that third reference point?
It's the Wintel model, and their ability to deliver an increasing level of capability. It's not everything everybody needs. We're not going to pull up our tent and say that's all we need to sell people. We don't believe that. But it is rising. You get more features in Windows 95 than 3.1, and more in NT 4.0 than 3.5. It's a rising tide that sets the benchmark.
So you don't see your organization as getting into a commodity space with NT or HP-UX?
Our product development will focus on how to position HP-UX and MPE as the logical extension beyond that offering. That's when a customer says that NT is not going to do the job, and they understand it's cheap but it's not robust, mission-critical, and doesn't have the performance and they need a faster database. I put my product hat on when the customer says thanks for offering that [NT] solution. In the past our sales reps wouldn't even talk about it.
In CSO we have the charter of delivering the enterprise solutions to customers. We also have the product responsiblity for developing all of our Unix, MPE and other differentiated platforms. A piece of CSO has to deal with a more commodity-oriented platform, namely the NT box.
I have to careful in using the word commodity. To some people that means soybeans, and no one can tell the difference. I can assure I can describe the differences between our NT platform and Compaqs. The question is whether those differences are as great as the differences between an HP-UX machine and a Solaris machine.
CSO has a role to engage in that business when it makes sense that it's the solution to our customer's problems. When a customer says they want a workgroup server that's not mission-critical just doing file, print and low-level applications work, chances are a very cost effective solution for that customer would be an NT platform. We will not engage in the product development side of that equation in CSO. We'll outsource that to our friends in the Personal Information Products Group (PPG) in HP. They have the volumes and low cost structure that business demands.
Microsoft seems to set standards by creating market demand. How can you keep Microsoft from doing to HP what Intel did to IBM -- capture the more lucrative part of the technology sale and leave the low-margin part to HP?
It's definitely by outsmarting them, and it's not going to come easily. The laws of economics point towards a mass market product winning more market share than a more tailored value-added product. Henry Ford did it in automobiles, Bill Gates is doing it in software, Andy Grove is doing it in microprocessors. The two areas we can focus on is performance and everything that means -- reliability, scability, robustness, manageability -- and acknowledge that leading edge is a moving definition. What's leading edge today, and only available in an MPE or Unix environment, we have to believe will sooner or later probably be available in the commodity product. We're not trying to build a wall around our turf. We're going to stay at the edge of innovation space.
The second key to the strategy is not only to be known as the best alternative to Wintel when you need more performance or functionality, but the most interoperable with Wintel. We can more seamlessly integrate a solution than anybody else, with common manageabililty, common messaging backbones, a common application develoment environment, better networking, support packages and consulting that helps figure out how to deploy and manage networks. We want to outsmart our competitors in being the most interoperable partner, the best logical choice beside a Microsoft environment where that environment doesn't do the job.
Is every operating environment proprietary? If not, how can you add value without making it so?
Every system has one or the other of a proprietary feature. There are two fundamental issues underneath that word. One is a level of differentiation. At one end you've got unique, only available from one vendor. On the other end you've got multiple vendors. It's the mainframe model versus the Microsoft model. The other issue is who is determining what is going into the product. Proprietary in that sense one company determines, versus a committee getting together. This is where we get hung up -- the least differentiated system you can buy from the most vendors is Windows. On the other hand it's the most proprietary, because Bill Gates decides what goes into Windows and nobody else does.
For years HP said that this open-ness was a feature that would distingush a product and make it better -- even to the point of pitting HP-UX against MPE/iX. I get the feeling that this afternoon is the beginning of new era, and that's over.
I'm glad that came through. We don't have to go out and sing the "open systems means Unix" song now. Singing that song created a lot of frustration for the MPE community. The fact is that MPE/iX can be just as much as big a player on the Internet as Unix or Windows or anything else. People will choose it based on its functionality.
That is helped tremendously by the pervasive presence of the Internet. People are realizing there's no way you can have a unified, single architecture in the future. You'll be connecting to the Internet, which by definition has everything from 5-year-old Macintoshes to IBM mainframes and everything in between. It's forcing the hand to cut out the business of whether everything is uni-architecture, and what's open and what's not. It's going to force the issue back to what are the interfaces, can I connect, what functionality comes in the product. If you've got something that's unique, you can charge for it.
HP must secretly grow weary of answering the question "Is the HP 3000 fading away?" Why do you think the channel partners in particular keep asking it?
They are asking it, and I absolutely understand why. Channel partners want the closest possible alliance between their marketing strategy and their vendor's. Anything they have to look like they're going out of that alignment, it's a double hit. Not only do they have to spend money themselves to create the demand and the image because the vendor's not doing it for them, but there's also the double risk the customer may see both messages. They the customer may say they're uncomfortable, because the partner and the supplier aren't saying the same thing.
I can relate to the fact that in doing a good job of singing the open systems means Unix and HP-UX" song over the last few years, we are now much more sensitive to the fact that we don't need to sing that song as loudly. Also, we can come back and listen to our channel partners and ask "How can we deliver a message where you can just come an tune it, and not have to explain a bunch of inconsistencies about why HP and you don't seem to be aligned."
What's to be gained by not telling HP 3000 customers today that they'll have 64-bit systems in the future?
That's good question, because we're sure pay a price for it in the form of the frustration we heard today. What we gain fundamentally is a value we hold even higher than living through this frustration: integrity. If we don't know the answer [about going to 64 bits[, we would rather say we don't know and we'll tell you when we do than give you the answer we know you want to hear. It may well be the outcome. As painful as this period is, in the long run our channel partners and customers will come back and say "HP didn't tell us what we wanted to hear just because it was easier."
One channel partner I talked with said the HP 3000 and HP 9000 are in a tight business unit today because "they're both declining businesses." Assuming that partner is wrong, how are those two businesses different?
We're trying to make the differences as small as possible, but there are some fundamental differences. They're in terms of degrees to which the industry at large is feeling comfortable in the development environment of one platform versus the other. They know that MPE is owned by HP. We can make MPE as open as we like, and we are making it as open as we can with interfaces, web server features. A lot of people regard that as an open environment, as indeed it is. But there's also the catch -- we decide what goes in it. We listen hard, but it's out there in the public domain and it's not being put onto anyone else's systems. That creates a certain perspective in the development community, in particular people who are developing tools, compilers and languages. They don't want to get too close to an environment owned by one vendor.
Given the steep hill that HP 3000 application providers are facing, what's the business case for continuing to sell things like TurboStore and DeskManager? Aren't there other markets where HP can apply its software efforts -- markets where there are stronger players and more areas for innovation?
I suspect those were areas where MPE was deemed as being under the control of one vendor, and the industry at large didn't rush forward and say they wanted to do mail services or store features for MPE. Our strategy is to first and foremost promote partners products that will do the job.Can HP do anything to increase the number of applications available for the HP 3000, or is it up to the channel partners and the customers in the market to make this happen?
We absolutely have to put a stop to our sales reps who are saying that our [MPE] channel partners should they should be selling the other platform. Our job is to enhance the efforts of our channel partners. We got that message loud and clear, and I think the field management team has the onus to really make sure our sales force adds values and doesn't detract.
We're not going to create them ourselves. We can make it very clear that this is a long-living platform that has a healthy, robust installed base, and that we're interested in promoting it and new applications. There are sad stories where we've put a lot of effort into specific applications only to run up against partners saying that the economics of doing those applications on the Unix platform, and MPE just didn't make the cut. I'm convinced there are still people out there to whom the 3000 still represents a good market opportunity that are not yet on it. We need to make sure we recruit and support them.
What do you think about the idea of an MPE developer's conference like the one HP supports for HP-UX?
That's an excellent idea. I haven't thought through the implications of what it might take for us to run it, but it might be exactly what we need to get more visibility.What do you think of the idea of an Open Systems Division, instead of two divisions competing for resources like CSY and GSY? What's to be gained by leaving the divisions as separate entities?
We are in fact sharing more and more of the programs and processes that go on in those two divisions. It is a possibility. I'm not announcing any reorganization here and get the divisions worried, but it's certainly one possibility that they could merge together. We're seeing some common tools and products being developed. They're also sharing a lot of programs. The 3000 division has developed some real competance at managing an installed base. In spite of all that we heard in the roundtable, I think we basically have a pretty happy customer base. GSY has a big installed base too, and they certainly have other challenges like fighting it out with the rest of the Unix community. Maybe we could have CSY do the installed base marketing for our Unix program. We're looking at smart ways to leverage the investments. That's part of what we've asked Glenn Osaka to do -- look at putting those two divisions together.
IBM appears to be targeting the HP 3000 customer who's hungry for applications on the HP 3000. Given the resistance these people have to Unix, why keep selling it to those who consider "anything but Unix" their strategy?
I'm not surprised that IBM's trying to sell AS/400s to anybody and everybody. I would if I were in their position. They sense a lot of 3000 customers are evaluating the impact of Unix or NT on their future and trading it off against staying in the environment they're in. We're not afraid of that invasion of our turf, but we are certainly reacting aggressively to it, to make sure those customers stay HP customers.