Volume 10, #5 June 1991
"MGMT MEMO" was written by Richard Seltzer in Corporate Employee Communication for the Office of the President. It was written for Digital’s managers and supervisors to help them understand and communicate business information to their employees. You can reach Richard at email@example.com
On May 15, 1991, over 500 senior managers attended Digital’s State of the Company Meeting in Merrimack, N.H. The meeting focused on Digital’s four main businesses - VAX VMS systems, commodities, systems integration and services. The following articles are summaries of the speeches.
We are in an industry that changes very fast. What we can put on a single chip of silicon today is exciting and significant. But every year we need fewer people and less space to produce more computing, and that trend is continuing at an unbelievable pace.
This means that we always have to be ready for change and have to be absolutely rational, logical and careful in our analysis and planning. We have to be honest with ourselves, learn from our mistakes, and keep seeking the truth.
Today, we’re in four businesses: commodities, VAX systems, systems integration and services.
Commodities includes personal computers, workstations and networking. Low price and high quality are critical factors in this business. It’s a brutal market - prices are going down dramatically - but it’s an exciting one, and we have chosen to be in it.
Our VAX systems are unique. VAX computers may not always be the lowest cost nor the fastest, but they can do things that nobody else’s systems can. We have the most modem computer architecture and the most modem operating system. It is designed to last forever and to expand over a wide range of computing capability. Our VAX products still account for the vast majority of the company’s revenue.
We need to remember that as Rose Ann Giordano has said, "Our customers want us to tell them what they need." We need to remember our basic historical strengths and build on them.
We can also put together very complicated systems to do special jobs, using both our VAX and our commodity products. We have enormous assets that we can use to solve the complex computing needs of our customers - to make complex computing uncomplicated through systems integration.
Our services businesses are very well run and profitable. Their people go to school for two weeks each year, and at some point in their career go to school for six months. They try things, recognize what works and what doesn’t work, and then correct their approach.
Only once in our history were we the company with the fastest and least-expensive computers — that was when we first opened our doors. We changed the computer industry not by having the fastest and lowest priced products, but rather by solving customer problems and taking care of every detail for the customer. We marketed by taking care of the customer. A few years back we made the mistake of over-emphasizing speed and price in our strategy, and lost sight of the need to take care of the infinite number of details that the customer needs.
In the tradition of science, you make an hypothesis, run experiments, then redo your hypothesis, run experiments, and redo your hypothesis again. Some people spend their lives learning and redoing their hypotheses and moving toward the truth. That’s where the ideals of religion and pure science meet. St. Paul said - keep searching for the truth. Pure science says - keep trying and learning.
Unfortunately, some people once they have "found truth" spend much of their time figuring out rules and regulations to impose on other people. And, unfortunately, some computer scientists say - once we figure out the answer, no other attitude or policy is allowed; once we’ve made a pronouncement, there’s nothing more to learn.
We cannot afford the luxury of being closed-minded. We always have to be open to new ideas and prepared to change our approaches in technology, marketing and sales.
When IBM came out with their AS400, we laughed at it. It had no great background of customers who loved it like we have with our VAX family. It did not have a wide array of software like we have. It did not connect well to personal computers. It wasn’t a UNIX product.
The one advantage IBM had was a scientific approach to marketing. When their original! approach failed, they made new hypotheses and more new ones. They marketed that system by what they learned from their mistakes. Today it’s a $14 billion business.
Our commodity business is exciting. We’ve done some magnificent work there. But we still have to learn how to make money in that arena. In that part of our business, we have to watch every penny, to be the most efficient, and to keep our costs and our prices as low as possible.
The recently announced ACE initiative is very important in this regard, providing us with exciting opportunities and difficult challenges. In the workstation business today, we face three major competitors - Sun, IBM and Hewlett-Packard. We got together with a number of other smaller workstation vendors and also personal computer makers (21 initially, now more than 30) and agreed on a wide set of standards - including graphics, human interface and networking. The standards they agreed to are basically ours but were developed in open deliberations. And with these standards, UNIX* and MS-DOS** software will be truly transportable from one vendor’s systems to another and also from personal computers to workstations. This initiative basically ties together the personal computer business and the workstation business, including UNIX workstations. Software that runs on personal computers also will run on workstations.
This is a magnificent, beautiful strategy. But it means that two or three dozen companies will be making the same two computers, based on Intel and MIPS chips. When everyone makes exactly the same machine, to be successful and profitable, we need to operate with a new business model. We have to change how we do things, recognizing that price and quality are essential in this commodity business.
If I were going to start a computer business today, I could not think of a better asset to start with than the VAX VMS family. In the last twenty years, almost every computer company, except Digital and IBM, stopped work on worldwide networks. IBM, and Digital with its VAX systems, are the only ones who can nicely, gracefully, and safely tie together everything in the world. We do it better than IBM because we set about to make it simple and straightforward.
Today, large companies - our customers - want to tie together everything in their world. Their biggest problem is sharing software, data, mail and drawings, and developing work teams across the world. Some work teams are permanent, some are ad hoc, some work with words, some with pictures, and some design automobiles and airplanes in different parts of the world. Worldwide networking — the kind we have with VAX VMS systems — is a key part of the world’s needs in computers. VAX VMS systems also offer security — security from mischief and security from loss of data.
We offer high availability with our clusters so that we can recover quickly before harm is done if a component fails. We offer fault tolerance, which means that within the fault tolerant system, any failure will not stop the computer. And we offer catastrophe-tolerant systems where a computer system is duplicated twenty to forty miles away and tied together with fiber optics, using FDDI, and everything in the main computer is duplicated at the remote site. This means any failure, regardless of how catastrophic, would not stop the computer, because it is all ready to go in the duplicated system.
VAX systems are growable. The system used to control the logistics, travel and cargo for the whole U.S. military was done on VAX computers. The system was designed to have 400 terminals, but in the build-up for the desert war it grew to 4000 terminals, many of which were in the desert. This growth was easy and quick, without a glitch. Software was modified in the desert to do special things, and for many hours on end, they were processing 44,000 transactions a second on a distributed network, which is probably a claim no one else can make.
From the beginning, VAX and VMS were designed to be very disciplined systems. They were designed with very carefully spelled out interfaces, so the modules could be changed, they could grow, and new features could be added. This means that VAX VMS systems we have today are much different from what we had fifteen years ago, but it is still the same VAX family.
Part of the design was the assurance that software written ten years ago would play today, and also will play ten years from now.
The VAX family is also designed to cover a large span of sizes. It was originally specified to cover a size range of one thousand fold, but we cover much more than that today. Today we make a VAX system that is a small workstation, and a VAX system that is a mainframe or a supercomputer. The same software that plays on the workstation will play on the mainframe.
People who do large scale computing and computing that will tolerate no failures have one goal in mind: they want a system that works and never crashes. In general, they do not put specifications on the system. They want us to tell them how to do their computers.
Customers today also want us to tell them how to simplify their computing. They have piled application on application, and database on database in their mainframe systems, and they want us to simplify these so they do not feel intimidated and terrified by the complexity of the system. Here we have enormous advantages. We can tie computers together, tie a whole world together with networks, and we can do all the things they need within the system — whether it be transaction processing, or modem databases, or mainframe computing.
The VAX family is not only the major source of income for the company, but it is also a major part of our strategy for as far into the future as we can visualize.
We are investing today to sell today and to be profitable today. We have in our hands everything we need to be very successful in the commodity market. And with our VAX products we have everything we could ever dream of having today. We have the technology, experience and people to succeed in systems integration.
Basically, we’ve made enormous changes in the company and, in general, the changes are working very well. We are learning to adapt to this new and challenging business environment.
At the end of the meeting, Ken Olsen, president, fielded questions from the audience and elaborated on points that he had made in his address. Here is a sampling of his answers: Marketing
To get $14 billion in revenue from the AS 400, IBM is reputed to have invested $3 billion in marketing. Are you willing to make a comparable marketing investment for the same return?
Marketing means a lot more than advertising or slogans. It means knowing everything about the market, "divining" every way in which that market needs computers and what the customer needs that only you can supply.
"Divine" implies mystery. When you take a witch hazel forked stick and hold it correctly and walk along, it’s supposed to point down suddenly where there’s water under the ground. That’s called a "divining rod." By "divining" I mean discovering something that isn’t obvious, by serious study, intensive work, understanding everything and knowing what the customer needs. Divining means studying, worrying, thinking, talking with every customer, knowing all the technology, knowing what we should offer and knowing what the customer needs. It does not mean making market surveys.
Marketers should tell hardware people what to do. They should define the product in every detail and make sure the hardware people make it or that it’s available. That’s a key part of marketing. The marketer then takes responsibility for getting that product exposed to the customer.
Marketing defines the product, gets the product, and makes sure the customer understands the product and the message. Then when requests come in, the sales people do the work.
The job of marketing is broad. Marketing should drive the company. Marketing influences hardware, software and services people and, above all, stays on top of customer needs.
What is the single largest obstacle between where we are and where we want to be as a company?
I don’t believe in making a list of priorities. It confuses people. It implies that we all have to do just the one thing at the top of the list. If you do that one thing, then success comes like magic. But, as a company, we have to do many different things at once, particularly in the area of marketing, which should drive all the rest.
We’re making major investments to create new channels for distribution: technical OEMs, components alliances, etc. How does this fit in the New Management System?
It fits very well. Many third parties offer expertise in areas where we don’t and having them sell our products is very important to us. Our goal is to make a profit and to grow. Channels are one of the tools to do that. Where channels help us, we do channels. But where they’re not profitable, we don’t do them. The New Management System helps us identify where we make a profit.
* UNIX is a trademark of UNIX System Laboratories, Inc.
** MS-DOS is a trademark of Microsoft Corporation.
Today Digital is a full-service company with two basic system thrusts: the VAX VMS family and the UNIX environment. VAX systems are alive and well and will continue to grow; and, at the same time, Digital is serious about UNIX. Other speakers today will cover UNIX. I’ll concentrate on our VAX VMS directions and how we plan to return Digital to growth similar to what we saw during the 1980s.
A few weeks ago our "VMS partners" - some of our field service support people - came to Nashua, N.H., where they had an opportunity to talk with Ken Olsen and Jack Smith. One of
them was also selected to talk to the Executive Committee. The biggest concern they expressed was the fact that the outside world is starting to believe that the VAX family does not have a long-term future. As we move ahead on our dual strategy approach, it is very important to emphasize that the VAX family is alive and well, at the same time emphasizing that we are serious about our various UNIX approaches.
Recently, a special model of the VAX 6000 system, a militarized version built by Raytheon, was sent aloft in a space shuttle and began orbiting the earth at over 17,000 miles per hour. That means we can now say that "VMS is the fastest commercially available operating system in the world" — another first for the VAX family. But what really matters here is dependability. Certainly a system that can be booted and operated in support of a space shuttle mission can be considered dependable — a system that customers are willing to bet their future on.
In addition to dependability, we offer scalability, with the broadest range of compatible computer systems. We also offer upgradeability that is unsurpassed in the industry. And our well-known networking capabilities and our Network Application Support (NAS) effort give us a high degree of interoperability and portability across many disparate architectures.
Putting together those unique attributes, we can say that VAX VMS is an "open commercial- strength system" — probably the only such system in the industry.
IBM offers many commercial-strength system capabilities, particularly with its AS 400 and its MVS systems. But IBM only has one product line that supports the open system environment, which is the new RS 6000 with its new AIX operating system. AIX, the IBM variant of UNIX, is clearly not up to the level of supporting commercial strength applications today.
The rest of our competitors, most of whom have now switched to UNIX, all support the open environment, but none of them yet have the commercial strength of our VAX VMS products.
We are working to make VAX VMS systems even stronger. For FY92, we have a full range of development efforts to support VAX VMS products in the open systems environment, as well as activities intended to bring better price/performance across our entire VAX product line.
The POSIX interface, which we announced will be available this year, is currently undergoing external field tests. We also have committed to support the distributed computing environment sponsored by Open Software Foundation (OSF). In addition, we will seek "X Open branding" early next year. X Open is a European consortium that does rigorous testing of various hardware/software platforms to make sure that they meet the definition of open systems that they have defined. We already have achieved that goal with our ULTRIX
system, and within a year we expect to achieve it with our VMS System. These are proofs of the seriousness with which we are approaching the open systems environment with our VAX VMS products.
Meanwhile, in calendar 1991, we are taking action to improve the price/performance of the VAX product line. Several weeks ago we made significant price reductions in our VAX 6000 family. That was intended as a tool to help our sales people sell against such systems as the IBM AS 400.
To me, the most disturbing aspect of the AS 400 success story is that, according to outside consultants, about 90% of their $14 billion in sales were installed without any serious competition being involved.
Although IBM initially targeted the AS 400 at the small business environment, we have seen countless cases where IBM used the AS 400 to offload or downsize their 3090 mainframe installations. That’s where customers see major savings. Those are accounts where Digital should be able to pick up new business applications with our VAX 6000 series.
Aside from price/performance, the next biggest concern of our customers centers around our layered product pricing policies. Today these policies are based predominantly on system capacity. A customer who upgrades a CPU is automatically hit with an increase in price for previously installed layered products. They would like to see a pricing policy based more on actual usage. Our goal this year is to change our policies to allow for pricing of layered products based on usage, as opposed to automatically basing prices on system capacity.
At the same time, we are selecting targeted markets where we would like to capitalize on the unique strengths of our VAX VMS family and where we are making strategic investments for the future. Two of these areas are distributed production systems computing and network server capabilities, ranging from our historic timesharing to the more complex network server environment. This is our VAX VMS "flavors" program.
As an example of a complex networking server environment, consider a test case we have under way now with a West Coast University. In their configuration, they have a VAX system supporting Macintoshes, personal computers and UNIX on desktops at campuses throughout the state. The VAX system not only provides the server activities for all those desktops, but also does the client work that is needed to access an IBM DB2 database that sits remotely at their headquarters. This is an area where Digital can capitalize on the unique strengths of VAX products-dependability, networking and commercial strengthsupporting customers as they begin to integrate their enterprise.
We also plan to announce later this calendar year some major revisions to our VAX product line that will significantly improve price/performance. And far greater advances are scheduled for calendar 1992. Basically, over the next year, we will completely revise the VAX family of products, giving us much better hardware platforms to support the unique strengths that we offer today with VAX VMS systems. And that’s just the beginning.
Next, we will begin introducing RISC technology into the VAX product line, bringing enormous performance improvements. While we will probably not see revenues from our new RISC-based product line until FY93, we want to make sure customers understand today that this is where we are heading. We can tell our customers that they can look forward to a long product life if they invest now in VAX systems. If they invest in VAX VMS products today, those investments - in applications, databases, user interfaces, peripheral equipment, training, etc. - will be good throughout the remainder of this decade, and into the next century. In other words, our strategy for future products is an important marketing tool for now.
Last week I delivered the keynote address at DECUS (Digital Equipment Users Society) in Atlanta, Georgia. Afterwards, many people told me how pleased they were to hear that that there is a long life ahead for the VAX VMS family.
We are telling customers that we will use RISC technology from the top to the bottom of the VAX product line and that the new technology will be phased in over a two year period. Along with the hardware platforms, we will also phase in the functionality for porting VMS software. And as proof of our ability to accomplish this, we already have more than ten advanced development units of our RISC architecture operating in the U.S. and Europe today in support of this program.
In the 1980s, the VAX VMS family was the main source of growth for Digital. Our goal in the 1990s is for VAX VMS products to continue to lead Digital’s growth.
We are calling FY92 the "beginning of the VAX Renaissance." Historically, the Renaissance was a time of rebirth, new excitement, revival and invention. In our VAX Renaissance, we are inventing "commercial-strength open computing." Over the next six months, we are taking all of the leadership functionality and dependability of VAX VMS systems, putting standard open interfaces on top of it, and bolting much faster engines underneath it.
Adding open interfaces to VAX VMS systems will make it very familiar for programs and programmers, who are used to other open systems. The new engines will be the fastest in the world in their price range for many commercial applications. But these aren’t the big news. The major uniqueness is that we will create the only open fast system with "commercial strength" functionality. For example, VMS software provides far and away the best software dependability on the market, so you never lose your data, thanks to features like clustering, shadowing, journaling, two-phase commit, failover, recovery and fault tolerance. VMS software also continues to provide the most flexible environment for easy growth, change, and multi-vendor support. And, VMS software is still years ahead in complex networking.
We are not just inventing a better mousetrap. In FY92, we will focus VAX/VMS products on the three markets where they will win most easily.
First, we will re-focus on timesharing. The word "timesharing" has fallen into disfavor of late, but that hasn’t stopped IBM from being very successful at it. In January they told the world that they had sold $1 billion worth of their new RISC machine, the RS 6000, in just six months. At first glance, it appeared that they were being remarkably successful in the workstation market. But half of those systems were sold as timesharing machines, and another quarter of them as servers. Only a quarter were workstations. Also, the bulk of their $14 billion in AS400 sales were for simple timesharing applications. The irony is that VAX VMS timesharing is far better than either of those and anything else on the market. There’s enormous potential for us if we re-focus on this area.
The second target for us is the emerging distributed production systems market. Here is where we go after the back-room, high-cost, mission-critical applications that often run on mainframes. As these applications grow, evolve, and are rewritten, these customers are looking for more cost-effective alternatives. But, they want the dependability they are used to on a mainframe. Once again, VAX VMS is the ideal solution.
In between these two markets, we are targeting the emerging market for "commercial strength" servers. We are already recognized as the best minicomputer to serve personal computers (PCs), winning the PC Week survey the last two years in a row. We will extend our product leadership here, and increase our marketing focus. In each of the three markets we are targeting, our goal is to have: the undisputed world’s best products and services, the happiest customers, the number one position in market share, growth and profits.
In July, we will attack these target markets by training the sales force on a revolutionary new way of selling. We are working with Howard Woolf and hundreds of people around the company to create a "sales attack guide", which will highlight systems customized for the targeted markets. For each target market, we have optimized the software environment provided. For instance, for our timesharing market — the terminal server -- we have created a lean, low-cost VAX system, by removing elements such as the database and PC connects, which are not essential in that market. At the other end of the spectrum, for distributed production systems, we have included all of the dependability features -- the shadowing, journaling and clustering -- as well as the system management features needed in the production system environment.
After we optimize the software environment for each target market, we then assembled the best "fighting systems" for each. For example, the sales person interested in distributed production systems turns to page five of the attack guide, and finds the five featured computers with which he or she can most easily. It shows the price, including: hardware, software and service. It shows the metrics that are important for this market. And it features the key software, hardware and service options for this market. The guide also highlights the top 100 or more applications, including which market and fighting system they go with.
This approach should make VAX VMS systems simple to buy, simple to order, simple to sell. It also helps us deal with one of the most difficult problems of systems engineering today — the fact that with all the components and options we offer there are millions of possible systems a customer could order. It is impossible to test all possible combinations to determine their performance characteristics. With this focused approach, for each of our featured systems we can provide detailed test results; so for instance, someone who buys a VAX 4200 system as a desktop server for Macintoshes will know how many of them it supports.
We will continue to offer, a la carte, all of the richness of choice and functionality of the complete VAX family. But our goal is to have 80% of VAX revenue coming from the the sales attack guide. That kind of focus will significantly improve our efficiency in engineering, manufacturing, sales, sales training, support and service. It will cost us less to do business this way, and we will pass the savings on to the customers who buy this way rather than a la carte. At the same time, this approach should help us sell added software and service, by highlighting the particular software and service needs of the targeted markets and making it easy to order.
There will also be a renaissance in VMS applications in FY92. Starting in July, we will be investing $65 million in a VMS application program, working closely with the Marketing Business Units (MBUs) and also Dave Grainger’s channels organization. Our first goal is to sell what we have. And so, we will focus on the top 100+ applications in the "sales attack guide", making sure that both the third-party application provider and the customer get the technical support they need, and also making sure the application provider has the incentive to sell on VAX products in FY92.
One of our other application initiatives is "priming the open VMS pump." We will show people how easy it is to move their UNIX software to open VAX VMS platforms - and automatically pick up "commercial strength" functionality. When we bring out our next open product (MOTIF) in a few weeks, we will have over 100 application vendors with us. We plan to do the same with POSIX at mid year. We also are doing a lot of work in the channels area, going after Value-Added Resellers (VARs) and systems integrators.
FY92 is also the year when we once again aggressively promote VAX VMS products. We will let the world know what "commercial strength" means - what VAX VMS systems can do that no one else can do. Our promotion will be closely tied to the target markets we are pursuing in the sales attack guide.
In FY92, the VAX Renaissance begins - the rebirth and new excitement in VAX VMS products. We are not waiting for the next generation RISC-based VAX products. Today’s VAX systems are the most dependable and flexible systems in the world. They have the best networking. They will soon be very fast, and "open." But what differentiates VAX VMS the most from the competition is commercial strength dependability, flexibility and networking. We have 10 million customers using VAX products today. We plan to delight all of them and add millions more.
The discussion of the commodity business begins with three four-letter words: cost, cost, and cost. In this context, cost is not product cost alone, but every cost incurred from quote to collection.
The world is changing, and Digital must change also. We have to become leaders in the dynamics of these changes. That begins with changing our mindset around the economics of our business and focusing on cost.
Commodity computer products are high volume and very price sensitive, but we can put serious engineering and marketing effort into these products to add genuine value and differentiate them from the competition.
For instance, you can buy a wide range of different kinds of toasters. The least expensive model with the least functions might sell for $10. Another with more features, or specialized, say for toasting bagels, would cost more, but not a lot more. There are differences, and people will pay for them. The trick is to get more on the price line than you add on the cost line.
The challenge we face is to produce the product at very low cost and to price so we get paid for the value we add. To do that, we need to have effective business models in all parts of Digital - from the selling through the logistics. That won’t be easy, because it requires new thinking.
We are moving into a new business that will stand or fall on how well we achieve commodity costs and value pricing. This commodity idea has to cut across the entire corporation, in everything we do. Teamwork will make the difference.
We need to adopt new thinking, work together, change how we organize, how we behave and how we measure. We are developing new business models, new ways of doing business and new ways of managing our business. We need to develop a new winning business approach in the marketplace, and we need to do it very quickly as we move into these new businesses. The world isn’t going to wait for us.
Our new components business is an opportunity for major incremental revenue — selling new products to new customers and reaching new end markets. The assets - patents, production capacity and engineering capability — are already in place, and we have significant up-side potential. The higher production volumes should ripple cost reductions throughout the company.
The supply chain of the computer industry is composed of the base technology suppliers and the system manufacturers who serve the end-user market. We have the opportunity to supply technology and component products to both. This way we can gain incremental marketshare in the non-Digital end market by getting our products incorporated into their products. Given that the non-Digital end market is currently 95% of the industry, this opportunity is enormous.
We have significant strengths to be a major player in this area. We have patents, technology, high quality, high volume world-class manufacturing capabilities, world-class engineering, and worldwide service and support. We are already selling thin film disk heads to three major customers and next quarter we will meet a production rate in excess of one million per quarter. We also have a major contract for private label video terminals. In other words, we are off to a good start.
This business is very competitive. We have to sell very high volumes directly to very large accounts. We need very low selling costs to succeed in this kind of business. We must establish strong engineer-to-engineer relationships, building our reputation and connections. We also need to build very strong long-term business partnerships. The companies we are dealing with are interested in the long haul. They are looking for suppliers who push the technology ahead, who will help make them competitive in the future. We know that we are competitive when we can make a profit operating as an Original Equipment Manufacturer (OEM).
Aside from the incremental revenue, the increased market presence through success in this business can also help us leverage industry standards. And quick market feedback from these customers can energize us to build ever more competitive products.
The Japanese have been doing this for years. In fact, one of their basic rules is that they sell the things that they manufacture for themselves to others to assure their competitiveness.
We must match the realities of the market, which are cost, cost, cost and price, price, price. We need to run the business on low gross margins. This means that we need high- volume customers where sales yields can exceed $10 million per year, with very targeted, low cost marketing and factory-to-factory logistics. Then through excellent asset management, we plan to achieve return on assets (ROA) high enough to make us a leader in profitability for the company.
This is not an investment business. We are going to pay as we go, making a profit on an incremental basis immediately. For us to succeed, teamwork is crucial. We are creating a strong team of dedicated sales people, many of them veterans who go back to the days when we often sold on a technical basis with a sales-engineer style of effort. We have a world-wide focused marketing organization and technical and sales support directly from the Product Creation Units.
We also need products that fit the market. For example, Larry Cabrinety’s VIPS Group recently signed a $30 million contract to sell video terminals to Olivetti. For this project, we modified the logo, the color of the unit, the microcode and the keyboard to match Olivetti’s unique needs. In the disk area, we have a new 800-megabyte unit that will be shipping in FY92. A number of system manufacturers are interested in it and have ordered initial samples. The Networks Group just signed a major manufacturer to use our Network Manager Software.
We also see opportunities for products that conform to industry standards. As an example, the ACE initiative will produce major hardware and software products that we can sell to other system manufacturers. Ten Product Creation Units have already set up dedicated
efforts to work on this Components initiative, and five others are developing their plans. The product list encompasses our entire range of peripherals, options, network and software products and their sub-component parts, and represents a major profit opportunity for us.
To succeed, we need high volumes, low costs, strong teamwork and strong financial controls. We will charge for what we deliver, and to quote Grant Saviers — "We have to squeeze the nickels so hard they turn into quarters."
In summary, the components initiative is a strategic opportunity to increase the company’s profits and market share and it also should force us to be a world-class supplier at many levels. For our long-term growth, prosperity and survival, it is important that we use this initiative to get very close to the pulse of the market.
The installed base of personal computing in the world is about $340 billion and the installed base of mainframes and minicomputers is about $660 billion. That means there’s a total of nearly $1 trillion of installed computing in the world and one third of the installed base, of everything shipped from day one and still in place today, is in personal computing. There are 71 million personal computers in the world, and another 25 million are going to be sold this year. Today, out of every dollar our customers spend on information technology they spend 55 cents on hardware, software, networking and services that relate to personal computing.
Our objectives are quite simple. We want more than our fair share of this business. And we want to use this buying pattern of our customers to gain access to new opportunities. Selling the desktop leads to selling the network, which leads to selling the servers.
It’s a big opportunity, but, of course, the business is extraordinarily competitive. Remember, the definition of "commodity" in the computer industry today is a 16 megahertz, 386 SX sitting on your desktop.
About half this market goes to the top half a dozen or so companies, such as IBM, Compaq, AST and Dell. The other half is owned by a lot of companies which have no more than 2% marketshare each.
It’s interesting to note that there is an enormous price differential - as much as threefold - between identical models of a PC from first tier suppliers such as IBM and Compaq, and from a Korean supplier. They offer the same box, same specs, same chips inside, and probably different disk drives and different memories for very different prices. This is a very important lesson for us. No doubt the market will squeeze down in terms of the dynamic range of prices. But we have the opportunity to put the brakes on that change in the marketplace by innovating, creating uniqueness and marketing. Using the toaster example, we want to build $6 toasters and sell them at $20, not $5 toasters and sell them at $10.
Dell Computer has an interesting business model for selling personal computers. Headquartered in Texas, they’re one of the fastest growing players in the PC business. They’ve grown to be a $500 million company in just a few years. They’re making money and, according to a recently released report, they’re number one in customer satisfaction. They do all their business on the telephone. There’s no direct selling and no distributors, (though they are working on adding distributors soon). Service and support are clearly distinguished from products in their pricing model. They subcontract their on- -site service. They also have just one price. There’s no such thing as list price and street price and letting customers negotiate the best deal they can. The single-price approach saves them sales time and effort and, builds customer trust.
Michael Dell doesn’t claim to be the price leader, but rather the "value leader." His message is "the best value in the market," not the lowest price. Basically, Dell maximizes the perceived value and charges for it.
They’ve taken the Avis strategy — strive to be number two in size but number one in customer satisfaction. That has helped distinguish them from the hundreds of small players in the market.
Our strategy is to become the leader in network personal computing. This is a message that builds on our strengths in servers, operating systems (VMS, UNIX* and OS/2**), local area networking, wide-area networking and clusters. Our network management products allow customers to handle enormous networks of personal computers. We can draw on all our strengths in networking, systems, service and support.
We can make PC networking easy in ways that truly delight customers. Today, PC networking is fairly complex. The average PC has more horsepower, capacity and disk storage than the average VAX-11/780 system. Some of them even approach the power of a Cray. There’s lots of different application software. MS-DOS is not the world’s most complete or wonderful operating system. As a result, parts of this networking job really are "rocket science." That challenges us to have the support, training and expertise in our organization to do the whole job for the customer.
Yesterday we introduced some important leadership products. We have a notebook-size, 20 Megahertz computer based on the Intel386 SX chip. Weighing just six pounds, it comes with a full VGA screen, mouse pad, built-in 60-megabyte hard disk and standard 1.4 megabyte floppy disk drive. We also introduced a 33-megahertz laptop that can have an expansion slot for an Ethernet or token-ring board for interconnecting to a network. Both these products were optimized for use in a networking environment. The stars of the show were two 486 systems. The DECpc 433 workstation, based on the Intel486*** chip, runs at 33 megahertz and comes with up to 64 megabytes of memory and full support of MS-DOS and Windows. Over 200 Independent Software Vendors (ISVs) have announced support for this product, which sells for under $6,000. And the "Tower" — the DECpc 433T system — offers 27 MIPS. It is limited by current MS-DOS software to 3 gigabytes of storage but has the capability to go to nearly 10 gigabytes.
We were able to develop these products in a remarkably short time by working in partnership with Olivetti for the notebook and laptop, Intel for the workstation, and Tandy for the Tower.
Today, the PC marketplace is in confusion. The channels and methods of distribution are going to be completely restructured over the next few years. There’s a significant move towards networking. The timing is right and we are very well positioned to take market- share.
We’re building on strong products that create more value at less cost. We’re striving to be predictable and to stay ahead of demand. And we’re going to win by clearly communicating our messages.
* MS-DOS and Windows are trademarks of Microsoft Corporation.
** OS/2 is a trademark of International Business Machines Corporation.
***Intel386 and Intel486 are trademarks of Intel Corporation.
Over the last 18 months we have learned a lot about the "commodity" business from our development and marketing of the multi-processor applicationDEC family of computers for small businesses. This versatile computer is based on the Intel486 chip, the same chip that powers the DECpc 433 system that was just announced. It is ideally suited for running the Santa Cruz Operations (SCO) UNIX operating system - a multi-user operating system that is extremely popular around the world and is now an important part of our overall UNIX strategy.
When we were planning our entry into this business, we learned that it is very important to build what the customer wants to buy. That may sound trivial, but it isn’t. Our small and mid-sized customers want to buy applications, not technology or computers. And many of the applications they want to buy run on SCO UNIX. That’s the strength of SCO UNIX. It’s a good operating system not because of any technical details, but rather because it runs lots of applications. Since SCO UNIX ran on Intel-based systems, we designed our product around that chip.
These products we are developing are not "commodities" in the sense of soy beans. They are high volume and price sensitive, but they are also differentiated. We can add value and price for that added value. We differentiate these products on the following dimensions: functionality, time to market, quality, cost, distribution, service and support.
Quality is a very important differentiator in this business. It’s particularly important for those people who did not consider it the first time they purchased. Customers learn by their mistakes. With the applicationDEC family we invested in quality as a way to differentiate us from the competition - in particular, functions that customers can see and appreciate.
For example, we are the only company in the industry with a product like this that has completely concurrent diagnostics. That’s extremely important for increasing manufacturing quality and also for simplifying service because this is a very powerful and complex product. The product consists of up to four 33-megahertz, Intel486 systems. That amounts to almost 100 MIPS of processing power, with over a gigabyte of storage in the cabinet itself. It can typically handle up to 100 concurrent users running SCO UNIX. You want a good high quality concurrent diagnostic with a product like this.
To keep our costs as low as possible, which is essential in this market, every part of the system - from the power supply, to the printed circuit card cages, to all the chips, to the board layout - was put out for bid, with internal groups competing with external vendors. Groups inside the company rallied around lowering their cost and being more competitive in the open market and, in many cases, won the bid.
Also, in this business strategic alliances are important. We have to get away from the Not Invented Here (NIH) syndrome — the assumption that we have to do everything ourselves. With the applicationDEC family, we formed two strategic alliances which were key in enabling us to bring the product to market quickly and in lowering our overall investment. Naturally, we worked with SCO for their version of UNIX. We also sub-licensed technology from Corollary Systems for the multiprocessing extensions that allowed us to put more than one Intel486 processor in the system and have it run seamlessly.
In addition, to lower costs and get to market auicklv. we had a good concurrent engineering plan. We worked very closely with people at the Digital manufacturing plant in Taiwan, both for the manufacturing and also for the manufacturing engineering work needed to get the product ready for market.
Functionality is another important differentiator. Our primary goal was compatibility with applications that ran on single-processor systems. If we broke the compatibility in any way and those applications couldn’t run seamlessly on a multi-processor unit, our entire design would be a failure. But within the scope of absolute applications compatibility, we added speed and expandability. The multi-processing allowed the system to run multiple times faster than the latest, fastest Intel chip. Multi-processing also extended the architecture* opening up application areas that previously could not run a singleprocessor Intel UNIX system. We also packaged the system to make it easier to add peripherals and to diagnose problems. We also focused on performance. All computers that have the same Intel chip and run SCO UNIX don’t necessarily have the same performance. We fine-tuned the system design for performance.
Time to market was also important. Two approaches are important here. First, spend time up front to make sure you have good specifications. And second, don’t change them. The temptation to "make improvements" is always great, but changing specs delays the product. By not changing the specs, we were able to bring the applicationDEC to market in record time.
Basically, to be successful with a "commodity" product of this kind, you need the right product, the right marketing plan, the right selling strategy and the right channels. In the case of the applicationDEC product those channels are being developed by the geographies. You also need the right service strategy because our ability to provide the comprehensive range of services clearly distinguishes us from the "clone" manufacturers, such as Compaq and Dell. That’s a tremendous asset that we need to capitalize on to differentiate ourselves in this business.
We can differentiate ourselves. We can have the lowest costs and the hottest products. There’s no reason we have to sacrifice that. Digital can be a mainstream, mainline player in the "commodity" business, as we are in our traditional business.
The Local Area Network Ethernet (LAN) Hardware business is a growing, profitable and strategic business in a competitive and maturing market. To continue our profitability and to gain market share, we need to incorporate a set of business behaviors which we are identifying as commodity-like and which are focused on solving customers’ problems.
The customer environment that we are exploiting is the millions of PCs and workstations that are being used today as standalone computing appliances. The users who bought these appliances are discovering the benefit — and the necessity — of being connected to share computing resources, data, and peripherals; and to exchange mail and messages. They need to communicate both within the work group, externally to other functions in their company, and externally to customers and suppliers. The Local Area Network is the entry point for these customers to networking. Their LAN buy is the second major purchase after the personal computer and is a key strategic business for Digital because it’s positioned between the personal computer and the enterprise network.
We are also exploiting a set of expectations about customer wants: ease of use, price- /performance and choice of vendor. The customer wants to connect to the LAN as easily as plugging in a telephone. It shouldn’t require any set up, any special knowledge, special tools or specialists, and it should make use of the lowest cost wiring available. Moves and changes should be as easy as unplugging from one wall jack and plugging into another. The LAN should be user centered — approachable, obvious, and convenient.
LAN customers are knowledgeable buyers. They are sensitive to price, performance and features as variables in a value equation. This puts a premium on our ability to price for value in the market; to be able to offer the lowest prices when appropriate and, when we consciously add value through performance, features or services, we price appropriately higher.
Customers also want choice of vendors. Like the telephone, there is an expectation that Local Area Networking products be compatible with those from multiple vendors, and that they interoperate using standard interfaces and protocols. Customers reject any possibility that might lock them in to a specific vendor.
Choice also applies to outlet. Like the phone appliance, customers expect to be able to buy the product from the outlet of their choosing, phone and catalogue sales in addition to direct and distributor sales.
Finally customers want "DUF." That’s the inverse of FUD - Fear, Uncertainty and Doubt. Customers now have FUD and don’t want it - they want DUF -Dependability, Uncomplication, and Flexibility. DUF is also the Digital Unique Factor, one of our competitive advantages over the plethora of small companies in this business. DUF is a set of value-added attributes which we bring to the product and to the relationship between the customer and ourselves: quality, value, worldwide service and support, and asset and investment protection.
Today’s $5.4 billion LAN hardware market is expected to grow to almost $7 billion in 1994. This market is segmented between Ethernet and Token Ring. Ethernet is, simply, a superb high-speed LAN technology and has historically outperformed predictions. We intend to continue this performance and to take market share for Digital with our new cost-focused Ethernet products. Moreover, we are continuing to aggressively drive Ethernet technology with better and lower cost product. For example, we believe that with our knowledge and expertise we can soon deliver an Ethernet chip with a cost advantage of 10 times that of Token Ring. With that cost advantage, we can put an Ethernet interface into PC’s and workstations at the same cost as a serial line interface today.
Our strategy for profitable growth includes:
o Quality - "Six Sigma" benchmarking with the goal of being best in class in all aspects of the business;
o Lowest cost producer - fewer parts, smaller size, lower power, fewer manufacturing processes;
o Easiest to use/install product solutions - utility-like, no tools, 10-page installation and operation guides, 10-minute installation, self-diagnosing, self-correcting;
o All channels - available where the customer wants to buy it
o Fast response - new product every six months, adjustment to market trends every three months, 24 hr. delivery; and
o Pricing — no frills entry price for the low end, value price for performance, features and service.
The camera industry provides an interesting example of a business that underwent both a customer usability and a commodity revolution in the 1960s, 1970s and 1980s. The early 35MM cameras through the 1960s were oriented to the knowledgeable and professional user. They were clumsy collections of multiple pieces requiring manual setting of f-stops, shutter and film speeds and required strobe units with shoulder mounted power packs. They were sold through specialty photographic stores and were high priced.
Then in the 1970s Canon, Olympus, Nikon and others came out with semi-automatic cameras. These cameras eliminated manual exposure control and provided for mounting unitized strobes on a hot shoe. The ease-of-use attributes and smaller overall size resulted in purchases by a whole new class of non-professional customers and sparked the migration of channels from specialty photographic stores to department stores and chains. These new cameras produced professional quality pictures and came at half the price of their predecessors.
Then came the 1980s. This generation of cameras added automatic focus, self threading and zoom lenses and incorporated the strobe in the body of the camera. LCD digital read-outs and audio feedback were added to tell the users just about everything they could ever want to know. Now, whole new classes of photographers, the disc and Instamatic camera users, migrated to the larger and better 35MM format. They got professional quality at a bargain price. The cameras became available from every conceivable outlet including mail order and the local convenience store. The camera market had become a commodity market. More than price, however, the customer got quality, availability, choice and ease of use.
It is the combination of these attributes which grew the market. In the process, customers became discriminating about value. I believe these lessons apply to Digitals commodity networking business.
Beyond the goals and strategies, the foundation of our business is a set of core competencies out of which we build the product. First is our ability to shrink the product in size. We have called this silicon integration. It is more, however, in that we have taken a holistic view of the product, blurring the lines between the software and hardware disciplines. We are relentless in seeking code efficiency and asking whether code need be in RAM or ROM or in silicon. As the product shrinks so do our costs.
Second is our core competency in networking protocols and electrical interfaces. We derive our ability for innovative execution and reliable product from these skills.
Third is design. Design is a unique differentiator. European design of automobiles or kitchen appliances and Japanese design of electronic appliances are key differentiators for these products. Design in network hardware - particularly as it complements the overall ease of use, approachability and applicability - helps sell product.
Fourth is human factors - creating network hardware that is user-friendly. We want customers to have confidence in and to be able to install and set-up their desktop interconnects by themselves.
On April 8, our new family of Ethernet smart hub products exploded into the marketplace. These products were designed with the customer in mind. They are the embodiment of our core competencies and incorporate our key strategies providing quality, lowest cost with outstanding ease of use. Concurrently we are actively developing all the channels which our customers want to use.
The hub replaces the Satellite Equipment Room rack at one-tenth the size. It needs no special ventilation or cooling. It mounts right on a wall, any wall or, if the customer prefers, in a rack. With its cover, it’s suitable for placement in an office. To expand or reconfigure a LAN is now literally a snap. The modules plug right in, and automatically configure themselves. You don’t have to brine the network down, and vou don’t need a system manager to reset parameters to put you on-line. This design gives us significant differentiation from the competition and, we believe, high customer appeal while solving the customer’s problems.
The April 8 announcement has had broad and favorable press coverage. The press picked up on our commodity-like advantages: ease of use, quality, price/performance, and choice of distribution channels. Perhaps with these products we are opening the era of the "point- and-shoot" style of networking.
April 9 was an important day for Digital. That day the Advanced Computing Environment (ACE) announcement was held in New York City and Brussels. On stage were Ken Olsen for Digital, Rod Canion for Compaq, Bill Gates for Microsoft, Bob Miller for MIPS and Doug Michaels for Santa Cruz Operations. Bill Gates said, "This is the most significant event in our industry since the announcement of the PC." That quote was picked up everywhere, worldwide.
We jointly announced a new RISC specification and two major operating systems — OS/2 version 3 and and a unified UNIX system (Open Desktop) from the Santa Cruz Operation. Twenty-one companies committed to this initiative on that day, and a dozen more have signed up since then. The major PC vendors and major production system vendors such as Bull and Control Data, NEC and Sony all committed to this set of standards.
The MIPS architecture, which Digital already uses in its RISC workstations, was selected as a part of the hardware specification. The data format known as "Little Endian" was also selected. This is the format used in the 70 million PCs and also the 10 million VAX users today.
Open Desktop incorporates major Open Software Foundation (OSF) technologies and will follow the Distributed Computing Environment (DCE) from OSF, the Distributed Management Environment (DME) from OSF, and Motif. We are putting full ULTRIX binary compatibility on this product; so all our ULTRIX applications will run unchanged.
Bill Gates announced OS/2 Version 3, a robust and very powerful new product. This 32-byte operating system will support all of today’s OS/2 and MS-DOS applications through Windows. It is also POSIX compliant and it will run on Intel386 and Intel486 platforms. They also have stated that it will run on the MIPS RISC architecture, with software developers’ kits in the hands of Independent Software Vendors (ISVs) by the end of this calendar year and end-user production systems by mid-1992.
What does this mean for customers? This means full, binary compatibility for the RISC line from the desktop to the data center. We have worked very hard the last two years to get 2,000 applications on our RISC ULTRIX systems. There are also thousands of SCO UNIX applications on Intel386 and Intel486 systems. And there’s about 25,000 to 30,000 applications on MS-DOS. In other words, with systems that comply to the ACE standards, customers will have access to about 40,000 applications.
This is an ISV’s dream. They will make one port and know they have access to 20 other vendors. This will change the way they do business. From now on, ACE will probably be their first or second port.
ACE will change the dynamics of the industry. Customers will be able to select their hardware by price, quality and functionality and know that by buying ACE-compliant systems, they don’t have to change their software at all. This means they will have investment protection and the end of having to rewrite applications over and over again.
What does it mean for Digital? All of these vendors have embraced Digital’s RISC UNIX strategy that we’ve been working on for three years: OSF, MIPS architecture, Little Endian data format, and our TURBOchannel bus. We don’t have to change what we’re doing. We continue to move ahead, and we have an enormous opportunity right now because we are so far ahead of the rest in the development and availability of ACE-compatible products.
The final specification is expected in 30 to 60 days. It will be given to all the initiative members under non-disclosure, as we move ahead building our compliant machines. But the DECstation Series 5000 machines we’ve been shipping since April 1990 are already ACE-compatible. They have the MIPS architecture, Little Endian data format, TURBOchannel bus, and ULTRIX software going to OSF. If you sell a customer a RISC ULTRIX system from Digital, the software they write on it today will run unchanged on ACE-compliant hardware tomorrow. We’re totally compatible. We are first. We have a leadership position. And we’re going to continue to take advantage of that leadership position now!
What are the opportunities for us? Customers who selected IBM’s RS 6000 or Sun SPARC systems in the last two months have asked us to come in and talk about ACE. They realize this change is significant and they want to take another look. The U.S. government is considering including ACE in their specifications.
Some ACE vendors have already asked us if we can act as an OEM, selling them the TURBOchannel Graphics option, VME options, FDDI, and other options we already have available. We can ship these components to them and they would put their own label on them. This would help them get to the systems market and would give us increased volume in the components business.
We also have a tremendous opportunity to sell our latest layered software, which now will be able to run on other vendors’ ACE platforms.
We also see systems integration and service opportunities. Some ACE vendors only sell through indirect channels. We will find these products in our accounts. We already know how to service and to integrate these products, and customers will need our help.
The ACE specification also allows room for innovation in such areas as graphics and multi- media. We intend to lead in innovation as well as bringing in our expertise from the VAX VMS area in symmetric multi-processing.
To most of the other ACE companies, who were great friends on the day of the announcement and are fierce competitors today, ACE represents a revolution. They have to get their strategy together and quickly design products. For us, ACE is a natural extension of our RISC UNIX strategy. We’re already there.
This initiative is open to anybody who wants to join. It is based on standards and provides investment protection. One of our key messages is that we have products that are compatible today. Nobody else can say that now. For software developers, our platforms are the logical choice for developing ACE application now.
As Mark Shulman, a well-known analyst wrote three days after the announcement, "For the first time in a decade, Digital has now positioned itself as a major player on the desk." We have an enormous opportunity here, and we must take advantage of it.
In Digital today, it seems that systems engineering is done everywhere by everybody. That may be our opportunity and our issue.
To understand the role of systems engineering, consider a building. What this building looks like to you depends on your interest in it.
If you are going to meet someone there, when you enter the building you might expect to find a lobby and an elevator and someone to receive you. If you opened the door and saw a dark tunnel, you might be reluctant to enter. You want to feel comfortable. Maybe the building has some special helpful features. The name of the company you are going to visit might be on the elevator. Or there might be an automated "people mover" at the door that will take you to the office where you want to go.
You’d have a different perspective if you were going to construct the building. You want to use world-class components for this building. Maybe you buy world-class windows from one vendor and world-class window frames from another, and discover there is a one centimeter gap between the frame and the glass on every window. These are world-class parts, but they just don’t work together. Say the air conditioning system is overloaded, and the heating system doesn’t work. You soon realize that buying world-class components doesn’t necessarily make for a world-class building. You have to think in different terms. You have to realize that all of the pieces in the building have to work together to make the whole and that somebody has to be responsible for making sure that they work together. They either have to specify the components or select them so that they work together.
In designing this building as in systems engineering, the pathway to success lies in paying attention to all the details.
Let’s look at systems engineering from the many perspectives we have at Digital. Architects specify and define the design of the components and how they will fit together. Engineering people design components to meet the specifications and fit the systems environment. They also may have to deal with marketing people whose job is to translate customer needs into system requirements and to characterize the actual performance of particular systems so we know what they can do for customers. We put applications on the system and make sure the applications work. When we sell the system, we configure it so that it solves the customer’s problem in a specific way. Then if extra help is needed to tailor it for the customer, our EIS people can do that custom work. We may recruit Manufacturing to make sure that we have problem-free installations and delivery. That’s our internal view of systems engineering.
The technical systems people try to put components together in such a way that they work together, provide more capability than any of the individual components alone, and satisfy a specified set of needs. Today there is no clear boundary line between hardware and software - 1 don’t know of any chip built today that doesn’t have software involved with it. The art of designing a chip in such a way that it works on a board is really systems engineering work and requires taking a systems view, including considering the customer perspective.
From the technical viewpoint, putting a chip on a board, creating a module that does something specific and interfaces with other components, putting modules together in collections that we might call central processors (CPUs) are all important steps in systems engineering. Collecting operating systems, net-ware, applications, storage and so forth is another important piece of systems work. Solving a business problem by collecting all of the pieces together is also systems work. But the question remains — what does the customer want? Like someone entering a new building, the customer may have a very different perspective of our final product.
The customer may say, "I don’t care what you do, but you have to put your technology together in such a way that it solves my problem." In other words, to that customer any activity that solves the customer problem with your technology is systems engineering.
We have to do our work in such a way that we satisfy what the customer is looking for. In this case, we would like to do our systems work so we satisfy the broadest range of customer needs with the minimum amount of work - which sounds like a good way to make money. A customer also wants to be flexible, to be able to change things to meet changing business needs.
We look at the system as chips and boards, collections of parts and testing, but the customer is looking for a way to solve a problem. These are both valid views.
Considering these ideas, let’s clearly define some of the things that we do. First thing, "systems marketing" means developing customer product requirements for the various Product Creation Units, based on the "Voice of the Customer" — the first phase of our four-phase quality program - and other market research. This includes the package, price, position and promotion - the complete solution for a broad class of customers. In addition, systems marketing supports the Field with marketing, technical support and training.
"Systems engineering" means defining and specifying, testing and qualifying all of the pieces. The pieces are typically designed to go together and to satisfy a broad range of customer needs. They take the ideas for building a system for a certain class of customers and make sure all of the pieces fit together and that they indeed do work as we expect them to.
The next phase is "systems integration." Integration requires systems engineering, but instead of engineering a system for a general class of customers it is focused on a specific set of needs for a particular customer.
Systems integration should leverage off systems engineering work. And systems engineering should leverage off the component work. That’s one way to help keep down development costs. If we all understand where we are heading and we all work together, we should be able to make more money.
For example, in the Information Systems Business Units, we are trying to make sure that the specifications for the parts are correct, understanding how they go together to satisfy a class of customer needs. We have to work together with multiple groups within Digital — the groups that make the components and groups that focus on particular industries. We also have to work with EIS because we can provide them with some technology and sometimes even some people that they need. As we do specific integration work, we learn what kinds of systems technology we might want to invest in to make the system integration work go better.
On the one hand we have the opportunity to develop tools that can lead the evolution from custom, individual integration work towards commoditization. On the other hand, we have to ensure that the components we are building fit together, so we don’t have a centimeter difference between the window and the frame. We need to exploit these two opportunities.
Our challenge is to make the best parts, to make them profitably, and to make sure they fit together in systems. And the systems have to make more profit than the sum of the parts because of the value we have added and for which the customer is willing to pay.
Over the last year, we’ve identified three common customer buying habits and built business models around them: commodities, production or solution systems, and systems integration. Services cross and are included in all three, but have their own unique business model.
In the commodity model, price is all-important. No cost is added unless it’s absolutely necessary. The capability offered has to be easy to sell, easy to install, easy to use. Indirect channels are important. Service may be a major marketing advantage but it is most often priced separately. Customers may buy through this model when they seek to downsize their systems or to buy smaller, less expensive computers than they have in the past, or when they seek what they perceive to be "vendor independence."
In the production or solution systems model - the VAX VMS model - the focus is on addressing customer business problems, through functional or departmental solutions. This is a traditional area of strength for Digital, where system dependability, reliability, scalability and interoperability are important. Customers are looking for solid, tested, applications availability and total cost of ownership. Here the customer is willing to pay for the value. This is where IBM succeeded with the AS-400, due in large part to their systems marketing effort.
We still need to understand all of our costs in this production systems model, and be as crisp and efficient as we can be. We also need to identify the value of our systems marketing work, particularly in addressing the key customer issue in this area - protecting their investments not only in hardware and systems software, but also in applications, data and training. In this model customers really want our solutions expertise. This is where the statement "Our customers want us to tell them what they need," is particularly cogent.
Systems integration requires a comprehensive approach. Here we need to continue to improve our ability to reuse the experience and expertise we get from one customer to meet similar customer requirements. To reduce our risks and costs, we want to go into a project with 70% of the solution already thought out, but priced for the comprehensive customized solution. As in the production systems area, these customers clearly want our solutions expertise as well as our capabilities in multi-vendor integration. This is the area where the Integration Business Units (IBUs) and systems integration groups are working most closely.
The IBUs are one of the dimensions of Digital’s New Management System. They have two roles. They assure that the right portfolio of applications, the right application strategy, the right solutions capability is in place for a specific market. Through systems engineering, they put together the integrated solutions that are necessary for that market and then make sure that the company makes a value-added profit in that particular area. They are also responsible for the company’s profitable plan that includes all the products, applications, integration, service strategies and distribution strategies for a particular market or class of customers. This is where the systems and systems integration marketing work is most apparent.
The FY92 business plans - the systems marketing and systems integration marketing work of the IBUs — directly relate to the three business models. Each IBU is normally focused on commodities, production systems or systems integration, but contributes to all of them.
For example, small business fits the commodity model. Price competitiveness is key. Indirect channels are essential. Every penny of the business model is critical to making profit. And differentiation often occurs by crisp execution and implementation to market. There have to be many applications available through general programs like ISVs or through indirect channels.
In contrast, the VAX VMS systems are at the heart of our health care business. We work closely with Cooperative Marketing Partners and put together automated medical record systems or laboratory information systems.
In banking, the focus is more on systems integration. Leveraging our systems integration work, we are building platforms for retail banking, investment trading and so forth.
All the IBUs are responsible for differentiating Digital capability, adding value as perceived by a class of customers, creating demand for Digital’s systems and systems integration and defining our systems integration strategies for that class of customers. They also create the guidance for sales, sales support and systems integration people regarding repeatable solutions and integration templates.
Electronic Data Interchange (EDI) technology can allow our customers to link electronically with their customers and suppliers, for close communication and efficient exchange of the basic "paperwork" of doing business.
In this part of the systems integration business, it is relatively easy to quantify the benefit we can bring to the customer and to value-price our solutions. Our message to the customer can be precise and concrete. The benefit comes from reducing work-in-process inventory, administrative staff for order processing and cycle times for typical administrative processes.
In EDI, we see an excellent opportunity for revenue and profit growth. We need to deliver solutions that are complete in all the details of the base VAX VMS hardware/software platform and the base software products needed to conduct the EDI dialogue between two trading partners. We work with third parties who have business applications and integrate those applications into our EDI solution so that they can serve as part of a continuous process to move information from a company out to its suppliers and customers.
EDI represents a major change in the way our customers do business. Often, they redesign the business process in the affected function to take advantage of the new efficiencies, reduce costs, and improve responsiveness to their customers. When they hire us as consultants for a fee to help them understand their problems, as a side effect we get a lot of information that can help us in the selling cycle, allowing us to propose a complete solution while understanding what this solution is really worth to the customer. We deliver that complete solution in cooperation with multiple third-party application providers and service providers on a worldwide basis.
Our software and hardware engineering organizations create the base platform with VAX platforms and VMS, DEC Net and DEC EDI software. We characterize the components so we know how our base platform works with the various third party solutions. Then, we add the major services components.
Services is the part of the solution that really drives the sale. We offer a series of packaged services ("Q numbers") with fixed prices. These help us break a fairly complex selling cycle into a series of manageable pieces that are easier for our sales organization to understand. This makes it simpler to sell and faster to get an early order. In addition, we can sell the services up front and defer what normally would have been selling expense shifted over to revenue and profit production. And finally, we offer a series of very sophisticated services to do whatever the customer needs in enterprise integration. The Integration Business Unit (IBU) acts as the responsible party to coordinate all aspects of the team, and the IBU marketing organization drives training in the Field as well as demand creation.
Over the last year, we found that it is a long march from the beginning of the sales cycle to the complete order. We learned that we need to take in money along the way, rather than wait a year to go through the sales cycle for a total job. Today, we break up the total job into deliverable and billable pieces, and take orders along the way. We might start with an analysis of the customer’s business need. Next, we need to prove that we can do the job. We do that by driving a pilot built around a series of service packages. We usually go back then and do a much broader needs analysis. First, we looked at a specific function or application set. Now we broaden our horizons to how we would implement this type of capability broadly across the enterprise with hundreds of trading partners. Finally, we deliver the complete solution, including VAX VMS products, our leadership DEC/EDI base software, third party software, integration tools and additional services to implement and manage the solution. By breaking the solution into pieces, the cycle becomes more easy to sell and generates a steady stream of revenue.
This is high-margin VAX business which, in addition, gives us opportunities to sell desktop devices, networks and other components. The customer does not tend to quibble about price — these are usually low- or no-allowance orders. And the technique of breaking the sales cycle into a series of consumable bites and using services to turn what could have been sales expense into profit is a model that can be applied in other areas of our business.
PowerFrame is an application for improving productivity in the development of electronic products. A "design data management framework," it may include electronic design automation , mechanicalComputer-AidedDesign (CAD), embeddedComputer-AidedSoftwareEngineering (CASE) and technical documentation tools. Our PowerFrame product provides both the context and the visualization of the data created in the design process. Its uniqueness is in its ability to integrate applications or tools into a single open environment and its support for platforms from multiple vendors. Over the last three quarters, worldwide sales for this relatively new product have exceeded our forecast by a minimum of 40% per quarter.
In the commodity business, the PowerFrame application is sold into accounts with little or no Digital content. These are customers who buy whatever is fastest or in vogue. In the standard production systems business, we continue to expand our VAX VMS base. And in the systems integration business, we sell services around PowerFrame solutions.
We’re working on a mechanism to allow OEMs and large end users to do their own porting to new platforms, while allowing us to retain control over our core product. An added benefit to doing PowerFrame development on these multiple platforms is that we become users of the other vendors products and services and this gives us better insight into what these vendors actually deliver.
As a success story, in the commodity business, a company called Chips in Technology, located in California, uses Sun systems and IBM PCs to design chip sets for IBM PC clones. They had a CAD design environment that did not allow them to automatically keep track of design data, configurations or versions of designs for multiple designers. They were not and still are not users of Digital hardware. But today they use Digital software, running on their Sun and IBM hardware. Having broken into this new account with our software, we are establishing ourselves as the best at multi-vendor interconnects, and paving the way for future sales of hardware, software and services.
As an example of production business, Toshiba uses our PowerFrame application, running under VMS software, to integrate their diverse applications and platforms. Toshiba is a SunSPARCOEM.TheymanufacturelaptopPCs,Sunworkstationsandcommunicationsequipment at three plants in Japan. Toshiba wants to provide data management in their environment of VAX computers and laptop PCs. In limited production testing now, they believe in VAX VMS systems for PowerFrame server support. They are buying a VAX 6000 system to add to a cluster and plan to buy a VAX 9000 system next year. They do not believe that UNIX servers are a viable alternative to VAX VMS servers. For them, the benefits of VMS software are smooth network integration, security, reliability through clusters and management services.
As an example of the systems integration business, Saab had a design problem with cable harness designs for their cars. They had a manual system of schematic entry and wanted to automate. The problems to be solved were time to market of new designs; reliability due to lack of configuration control; engineering change order (ECO) support and security; rigidity of the design process; and inability to implement policies. The PowerFrame product doesn’t solve all of these problems, but it provides the base for services to be applied to form a total systems integration. Saab bought hardware from us and initial services, and based on our initial delivery of those services, they are increasing their individual hardware base and ordering further services. The PowerFrame product is open enough to allow Saab to integrate the tools it has chosen.
Today, over twenty major software tool vendors have signed up to make their tools run under PowerFrame software as a part of the PowerFrame Synergy Program. This helps us provide a broader base for services to be added later.
In summary, our PowerFrame product is the world leader in framework technology.
We have open VMS software, open systems, and also "open services." As an open services company, we do what our customers want us to do. We provide them with flexibility of choice and multi-vendor service solutions. We deliver a wide range of service support from the desktop to the data center, and we can manage the most complex information environments for customers.
Today, Digital supports over 8,000 products from over 800 vendors. We’re leaders in the multi-vendor environment maintenance world. We set the standards for excellence in international service capabilities, both in quality and in depth and breadth of our services.
Being an open services provider means that we must serve three business models -- commodities, VAX VMS systems and systems integration.
To deliver this breadth of services, we made a number of changes in our organization. In particular, we clustered some of our businesses to better be able to serve the different customer buying habits, to save money, to eliminate redundancy, and to make it easier for Sales and customers to see what we’re doing.
The Product Services Cluster, managed by John Rando, is our traditional services businesses. It includes Hardware Product Services (low growth, high profit), Software Products Services (high growth, high profit), and Desktop Services (explosive growth, just beginning to ramp up on the profit curve.) There are also businesses within businesses, such as our focus on maintaining non-Digital hardware and software.
Education and Consulting, managed by Pat Cataldo, is our knowledge-transfer cluster. Training continues to grow and remain profitable. But to keep that growth rate up we have to work on customized training in non-technical, non-traditional areas. For instance, perhaps we can train our customers on Just-in-Time for manufacturing, or provide training and support for a custom system integration solution.
This morning, Ken Olsen quoted Rose Ann Giordano, "Our customers want us to tell them what they need." There’s an important corollary to that — it’s fine to tell our customers what they need to succeed as long as we’re getting paid for telling them. Hence we created Digital Consulting Services, which includes Management Consulting, Organization Consulting, MIS and Applications Consulting. We’re learning to sell this expertise instead of giving it away. This business is growing 20% a year, and the profits are good.
The third cluster, Systems Integration and Support Services, is run by Max Mayer. It represents about a quarter of the company’s service capability. Here, we have combined our Computer Special Systems (CSS) and Software Services project capability to form Application Project Services. This is a very high risk, difficult business. We need repeatability and discipline. It is a very fast growth area. Profit is still a problem, but we’re getting better fast. Network Services is the organization that does the planning, design implementation and management of local area networks, wide area networks and global networks. It has excellent growth and good profit. Operations and Support Services includes, among other things, facilities management, customized support, recovery services, contingency planning services and disaster planning. This is a risky and very high growth business. They are on plan or ahead of it, but they are not yet where we want them on profit. Finally, the Systems Integration IBU is very high growth and is working to improve profits.
In Digital, business is "systems integration" if it has non-standard terms and conditions, some custom content, and requires a program manager to manage it.
The Systems Integration IBU is responsible for the business models: methods, tools, training, program management and rolling together the profit and loss for our company in systems integration. As systems integrators, we plan, design, implement and manage the customer’s solution needs. To do that, we use all of Digital’s open services, the products and software from the Product Creation Units, third-party hardware and software and of course, our various partners.
We can’t be a systems integrator to everybody. We have to focus our skills in support of the marketing strategies of the various IBUs. We work with the IBUs and Sales to define as a team what skills we need, where we need to grow and where we’re going to rely on our partners.
When you look at our performance in systems integration over FY91, there is good news and bad news. The good news is that this business grew significantly more than we planned. The bad news is because we underestimated it, we ran short of program managers, resources and trained people.
In FY92, we plan to train hundreds more program managers and finish the roll-out of the Digital Program Methodology: one jargon, one method of proceeding and one process for managing systems integration projects. We’ll also elevate the skills of current program managers because the programs are becoming so much more complex as our sales people and customers develop more faith in our ability to do this work. The program manager works for the account manager and is the key to the success of the account manager.
We will continue to develop third-party alliances, working with the IBUs and Sales to figure out who our partners should be.
We’ve implemented improved controls and discipline in our programs. We do more up front qualification of programs, and the Executive Committee reviews and approves large programs. But we have to do a better job of value pricing this work.
Two years ago no one knew we were in the systems integration business. Today we are ranked third, fourth or fifth (depending on who you ask) largest systems integrator in the world.
In FY92, we plan to improve our profitability by working closely with the IBUs and Sales. Our top priority is profit. We’ll work on our Application Integration Sets (AIS) "road maps" that tell us what applications to use to solve which problems with which customers, with which services, what departments and what products on which platforms. That should reduce risk and make us more predictable. We’ll also pre-configure systems and have platforms that we tailor so 70% of the systems integration work is done before we begin the project. We’ll make better use of our assets, including "reusable bid teams" - people who know how to win that business.
Finally, we’ll improve our global capabilities. Our Program Management Office has established global links, and we’ll improve on that and help work on global service pricing and service warranty. We’ll also improve our global training, methods and tools, as we focus on profit, profit and profit.
Systems integration is big and getting bigger. Today, we have roughly 1,000 projects and programs of all sizes going on at any one time. This is going to continue, and we’ve got to get it right. We’ve got to deliver the full range of services and products and learn how to do it at a profit.
We have a lot to be proud of regarding Digital Services and Systems Integration. Our services are flexible, with a single point of responsibility, the way our customers need it. We’re leaders in supporting a multi-vendor environment; and we’re well on our way to becoming the best world-class systems integrator.
In the computer industry, we see successive waves of profitability. In the first wave, we we were motivated by selling computer hardware. Our service, operating systems and applications were all intended to help us sell more hardware. That’s where the margins were.
As hardware became more of a "commodity," prices went down, and we couldn’t get the same kind of margins from hardware as before. Then we looked to software for the next wave of profits.
We expect that software will be the driving profit factor for the company over the next three to five years. By that time, commoditization will have struck there as well. We will still want to be profitable in software, just as we still want to be profitable in the hardware business; but the margins will be tougher. We expect the next wave of profits after software will be in systems integration.
Software differs radically from hardware in several important characteristics. It’s intangible, intellectual property. It’s hard to "show" someone software and hard to protect it. Also, software doesn’t break, so you can’t build a business around repairing it. If software doesn’t do what you want it to do, that’s because it has a design failure and needs to be re-engineered, not just repaired. Software also has enormously high incremental profit. If you sell one more copy of a piece of software, your incremental cost is nearly nothing — regardless of how valuable and useful that software might be to a customer. Finally, the selling of software depends on technical support by very competent pre-sales and EIS people in the Field.
To compute the revenue vou would get from a given piece of software, take the mice.
divide it by the number of technical hours required to make the sale, multiply that by the hit rate (because if you only make the sale the half the time, you need twice as much as support as you would with a 100% hit rate), and then multiply by the number of technical hours available.
To get more revenue, you can adjust various factors in this formula. For instance, you can change the price. Or you can adjust the training and packaging to reduce the number of technical hours required. You can help sales people make better decisions about when not to bid, thereby increasing the hit rate. You can also improve the training technology to get people in the field trained more quickly. We have the opportunity to make very good profit from software in FY92, provided that we can get those pieces of the formula right.
Our competitors in software use the following strategy: they call our sales force and say, "We’ve got this great piece of software to sell on your machine. There’s free technical support. Just call this 800 number, and we’ll move a lot of hardware boxes for you." Yes, it’s good to move hardware boxes, but it’s even better if we can sell our highly profitable software on top. In other words, we have to shift our orientation if we want to maximize the profitability of the total systems that we’re selling.
At the same time, the industry is evolving toward an "information systems utility." Digital started by selling to departments and still sells to them. On top of that, we now do a tremendous business in selling products and services to integrate a whole corporation. Once a company has succeeded in integrating its own departments, it now wants to integrate with its customers and suppliers. Then they’ll discover that the integration work they have done would also be useful to other companies in their industry, and they will sell that integration to their competitors, as we now see to some extent in the banking and airlines industries. After that comes a worldwide utility, which provides the same service but is not specific to a single industry. Sometime around the year 2000, we can expect to see consortia of worldwide computing and communications companies delivering hardware and software capability as a utility service. Delivery through that kind of utility would be the ultimate in commoditization of computing resources.
The market will expand because the major companies in the world will give much of the work of their MIS departments to the utility. That work probably represents between 35% and 50% of all the expenditures that are being made for computing today.
In the utility model, there are basically three places you can sell. You can sell to the end-user, which is a very competitive, highly commoditized market, centered around the human interface device. You can sell to the corporate information officer, who is very interested in command and control, accounting, billing and efficiency. And you can sell to the developer who writes the applications that are used in this utility.
Basically, as we sell corporate information integration, as the industry moves toward the utility model, the industrial or commercial strength of our software and services will have to increase.
The job of our software business is to create the infrastructure that will allow corporations to integrate themselves and will also allow for the growth of industry and worldwide information utilities. At the same time, we have to support the trend toward open, multivendor systems.
How can we do all this very complicated, expensive development work and still make money'
First, we have to commit to a set of standard interfaces, both existing and future, that will convince customers and the future information utilities that we will be able to interoperate with other parts of their infrastructure. Today those standards are the Application Environment Specification from the Open Software Foundation (OSF) and the X-Open Portability Guide, together with standards from the International Standards Organization (ISO).
Second, we have to produce and sell world-class products with rich functionality that is not available through the standard interfaces. If we don’t get these products to market quickly enough, someone else will, and we’ll fall behind in the development cycle where new functionality leads to new standards and then another level of new functionality.
Third, we need to work the process of advancing the standards even farther and getting our technology accepted as standard.
We have done very well in the first and third categories. OSF, for example, has bought more than half of its technology from us. We’re packaging the Distributed Computing Environment software for them, and we have a very strong bid in for the Distributed Management Environment software.
But, from a software perspective, we have to reduce substantially our time-to-profit and time-to-market. On the average, historically, we have taken four to five years to produce the first dollar of return on our software investments. That is an astoundingly long time. Today, we’re making money in software, and it’s a very attractive business for us. But the historical time-to-market of our products has been far too long. Fixing that is our major thrust right now.
In FY92, we plan to extend our existing, proven architectural capability and product delivery from the VAX VMS space to heterogeneous platforms, including ULTRIX of course. We will then be able to treat our software as a business and not just as a tool to leverage our hardware sales. We’ll be able to make substantially more profit on our intellectual property by judiciously selling it to our competitors and to customers who want to use it on our competitors’ hardware platforms. For instance, we are putting Motif, the OSF standard windowing interface, on Sun platforms and expect to make a profit at it. We’re interested in ACE, UNIX, IBM and Macintosh platforms and windowing terminals — all the places where we deliver Network Application Support (NAS) capabilities.
This business challenges our assumptions about how we’ve dealt with software in the past, but it is a wave of the future for us in making money. We have to package and sell our NAS software appropriately for these new heterogeneous markets. We’re currently working out details of the deliverables.