Volume 10, #1______________________________________________________________ January, 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 December 6, 1990, over 500 senior managers attended Digital’s State of the Company Meeting in Merrimack, N.H. The meeting focused on Digital’s new management reporting system, quality, open systems products and strategies, and the production/mainframe market. The following articles are summaries of the speeches.
State Of The Company Address by Ken Olsen, President
Ken Olsen Responds To Questions
More On The Management Reporting System by Lyn Benton, assistant corporate controller
Total Quality Management And Its Role In Making Digital More Productive And Innovative by Frank McCabe, vice president, Corporate Quality and Process Technology
Banking As An Integration/Application Business Unit by Norm Goldberg, vice president, Banking/Investment Business Unit
The Open Systems Company By Bill Demmer, Vice President, VAX VMS Systems and Servers
Ultrix Opportunities by Kurt Friedrich, group manager, Open Software Group
Open VAX VMS Systems by Ken Swanton, Marketing Manager, VAX VMS Systems and Servers
The Production/Mainframe Systems IBU by Bob Glorioso, vice president, Information Systems Business
Update On The Mainframe Market by Rich Whitman, group marketing manager, Information Systems Business
The High Availability/Fault Tolerant Market by Fernando Colon-Osorio, Corporate Consulting Engineer and Manager of the High Availability/Fault Tolerant Business Unit
Doing the Whole Job
We’re trying to do a job that, I believe, is more complicated and difficult than anybody’s ever tried to do in the past.
We have friends and competitors in the computer industry who have concentrated a large organization in a very narrow market niche. For example, Apple and Compaq have done very well that way. Their strategies are elegant, simple and easy to understand, but terribly dangerous. I’d hate to be in their shoes or the shoes of any other niche players right now. If they ever make a mistake, or if one competitor does things significantly better, they could have serious problems.
We always have concentrated on the hard things because that’s what we want to do. That’s where we find safety, fun, excitement and challenge.
We’re trying to take a large company that’s involved in many diverse activities, and integrate these activities as one, so we present one integrated image to the customer. IBM does that to a large degree, but we’re trying to do it much more extensively.
We face many of the classic problems of big companies. First, we have groups who don't
communicate with one another. That happens naturally in every company, but we have to fix it.
Customers want a complete job from us. They want planning and design. They want hardware, software, networking, applications, installation, training, and support. We need to do the whole job, but we are acting like separate groups.
Engineers probably generate much of the problem. They like to define their product as a box they can specify and measure and for which they set the price and the cost with little thought of the systems’ cost. Then they let the Field design systems with it, and the Field ends up with a price book so thick no one can carry it.
Today, Sales has to take all of the myriad of products, and design a system, price it and put a bid in. Engineers say, "If we only had more competent sales people!" But engineers couldn’t do the job as they, themselves, defined it.
Doing the whole job takes a lot of planning to make sure every detail is done for every job we do for every customer.
This is in conflict with what we say we have learned from our experience and from the experience of Russia. Centralized planning stifles creative thinking and the taking of responsibility.
Centralized planning means centralized budgeting and control. It eliminates internal competition and risk taking, and it discourages products and ideas that have not already been proven by someone else.
The goal we have taken for ourselves is to avoid the rigidity of central planning that controls everything, and limit it to the planning necessary to be sure we take care of all the activities to which we are committed, and that we have the tools for every group to accomplish their part of the operation.
It is interesting to consider some of the most complex problems in the universe, like feeding New York City. Can you imagine anybody in their right mind allowing that entire operation to go on — the heating and feeding of eight million people on one little island with only a handful of bridges and tunnels coming to it -- without any plan? Yet somehow it works, and they all get fed.
Now nobody in their right mind would take a fourteen-billion-dollar company and allow dozens of business units to run independently. Obviously, you've got to control everything.
What would happen if we started controlling New York City? They’d all starve, as is happening in Moscow.
You see the conflict. Every organization has a tendency to build up staff and committees and optimize everything. But we can’t run the company centrally.
Our Faith in Technology
We have faith that we will win with technology. If we ever back down on technology and postpone it or stop it, we’ll go into a downward spiral like others before us. We won't tolerate the idea of backing down on our investment in technology.
People ask, "Why do you invest so much in products that come out so fast and increase the capabilities of computers so fast at a lower price?" Our answer is, "We have faith."
For 33 years this company has believed in and depended on the idea that as we make more capability for a lower price, the demand will grow even faster. We are not going to stop producing computers that are faster and cheaper and more powerful, at lower price. We’re not going to stop investing in technology. History has proved this is the right course, and that’s part of our faith.
What Happened to Our Profit?
People ask us what happened to our profit. We didn’t plan to make so little profit. We had a plan that produced profit, but that plan was dependent on selling more dollars of computing even when we were offering great technological improvement.
In the last few years, our product offerings have grown enormously in speed, which is only one of the measures of capability. In terms of MIPS (millions of instructions per second), the systems we’ve sold have more than doubled each year for several years. And we have faith that demand will continue to increase even faster than that — but not every year.
Obviously, when there’s a recession, or there’s a slowdown in the end-user business, which we’ve seen for several years now, demand does not grow that fast. That’s why things have slowed down for us. But we still have that faith in growth.
We budget and plan for being a growth company because we believe that demand grows as we improve the performance and price. That’s our faith in technology.
What is our weakness if we do so well and have such faith in technology? In putting technology first, we put much of our energy and money in technology, and we never quite get around to doing the rest of the job.
All the products we dreamed of having two years ago we have today, and we have products coming two years from now that are absolutely wonderful.
We have the products, but we don’t tell anybody. We don’t even tell our own Sales department about them. If we have any more money available, there would be more technology to do.
The answer to many of our problems is to break the company into pieces, each one as a separate business unit. That means each business unit has the obligation to lay out a business plan which demonstrates they will take care of every detail.
We have the best hardware, software, and networking; but, for every application, job, and solution we offer, not only do we have to offer technology, we have to offer the myriad of details necessary to define the job, specify the solution, educate our salespeople and customers, design the system with precision and elegance, price it with wisdom, deliver it, install it, and do all the things necessary to make the customers happy.
That is why we are breaking the company into business units and each business unit presents a plan, and the Executive Committee and the Board of Directors listen to each plan.
Basically, we want to have all the advantages of a big company, and all of the advantages of a small company. We will treat every unit as if it was a small company and have them work together to have the advantages of a big company.
Each business unit will be measured just like any small company. Out of this, if it is done right, will come entrepreneurship.
In the positive sense of the word, an "entrepreneur" is someone who if given responsibility and assets, lays out a plan and will perform miracles to make it work. In contrast, if a central planning group lays out a plan and funds it, it won’t work. People see it as the central planner’s plan. They feel that they never have enough money and that the plan was done by someone who didn’t understand the problem.
A good entrepreneur, who has a plan and takes responsibility for it and for correcting it when it is wrong, will make sure it works.
In order to make this system work, I have five principles:
o Everyone works for the Business Units. That is, the president, the financial vice president, the engineers, personnel, etc.
o The individual who proposes, does what he or she proposes.
o No one without responsibility tells others what to do.
o Budgets are sacred and stable, and not continuously "up for grabs." o Each business unit reports to a person who’s responsible, and not to a committee.
The business unit takes the responsibility to run that business. If things turn down, they cut the budget, use less money, and make a good business plan in the end. But we don’t re-shuffle people’s budgets to follow our latest fad. We don’t thrash around people’s budgets continuously. If you ask a business unit or an engineering group to carry on the job, to hire people, make commitments, lay out plans, they have to assume their budget is sacred. They, however, have to make corrections if things go wrong. But an unstable budget is devastating to an organization. If your resources are always at the risk of being taken away by somebody else, you never can be effective as a business unit.
Our system is very simple. We have three groups of business units. One group generates products — such as central processors, workstations, disks, servers, semiconductors and customer services. The second group consists of those who integrate and market and develop special products which are unique to their market. The third group is the people who sell. Each business unit group makes profit on the cost it incurs.
The account manager is a business unit manager who makes or loses money on the investments and sales he or she makes. There is no budget in the districts, only in the accounts.
What happens to the district and regional managers? Their job is to help the account teams and act as quality control for the account teams. They are measured on how well they help.
People who incur the cost have to justify the profit made on that cost. People who are out getting applications have to justify how to make a profit with them. Everybody makes a profit. If the customer is not willing to pay for it, we don’t do it.
Obviously, with freedom comes the opportunity to bend the rules sometimes for the good of the overall project. But, in general, every plan shows how every cost we incur will make a profit.
At the end of the meeting, Ken Olsen, president, answered a number of questions from the audience. The following is a summary of a few of those questions and answers.
Why is it that all the promotions at the higher levels in the Field seem to go to people
in Field Service?
First, that’s not entirely true. Remember that Bob Hughes, who is heading our North American sales, is definitely a salesman. He didn’t come from Field Service.
On the other hand, Field Service for many years has had a rule that everybody goes to school two weeks a year, and once in their career goes to school for six months. They also have had the business unit approach from the start. Each unit lays out a budget and is held responsible for that budget, as if it were a business.
If some other groups don’t believe in education, don’t believe in running things like a business, they shouldn’t be surprised to see who ends up at the top. For people to want to get ahead, the lesson there is that education and experience pay off.
Will revenue, including OEM revenue, be reflected in the account P&L?
We measure business units not for credits they own because they happen to be sitting somewhere. We measure each as if it were a business. We don’t use the word "credits." "Credits" means arguing for something you didn't deserve. We want to give information to the business unit managers on how well their investments pay off.
If a business unit invests heavily in a sale, and then gives the order to a third party, who gets a 30% or 40% discount, and the net result is Digital loses badly on that order, you don’t get any credit for that order. On the other hand, if you invested in an order and the company made a profit, it’s the profit we measure.
The Accounts make the investments in getting the business from their account anywhere in the world. If an order for an account comes in a remote town, the local district does not get credit for that order. Districts don’t get credit for orders; they just get credit for the help they give. The account has made the investment, and they use that order to justify the investments they made in that account. The measurement systems are primarily there to help the Business Units run their business.
We’re simply measuring the return on the investment made. In the product-line days, somebody owned the medical business, and somebody else owned the education business, and they’d argue violently to make sure they got credit for everything that they possibly could. Then, they spent the money wherever they felt like spending it. That’s not good business. We're measuring return on the investments we make: investments in selling, in integration and in product development. That’s the obvious business reason for accounting.
You mentioned that technology in terms of MIPs is growing at 120% a year, and that market
demand is growing rapidly too, but not every year. Doesn’t that mean that in recession,
if the market doesn’t grow that much, someone’s going to get hurt?
Yes. In every industry - automobiles, airplanes, snowmobiles, and computers - there’s a large number of manufacturers at the start. At one time there were about 500 making personal computers. At one time there were about 260 making minicomputers. There were also a good number of companies making mainframe computers. Now, in each of these categories, there is only a tiny number of companies still left. The number has shrunk enormously already, and it is unlikely that there will be anywhere near the number we have today if this recession lasts longer.
We plan to be one of the survivors. That is why we are in a strong financial position. We probably have one of the strongest balance sheets of any company in the country. The way to survive, particularly in a recession is to invest heavily in new products and maintain a very strong cash position.
We are embarrassed that our profit has not been higher lately, but our primary goal is to maintain our strong lead in technology and our strong financial position and, therefore, to be one of the survivors.
Networks are strategically vital to Digital. If this assumption is true, why are we not
planning and investing in the appropriate level of training for the Field? It seems it’s
owned by too many people with conflicting agendas, and no responsibility to the budget.
There may be problems in this area that I don’t see, but I will tell you the problems I do see. To a large degree we developed Ethernet as it is known today. We told the world how to make it work. We made a standard, but then stopped pressing our advantage and went on to more exotic technology.
We’re going to turn that around. Now, we are going to be the low-priced producer, the low-priced seller, and we’ll settle for nothing less than a very large share of the Ethernet market.
We can do things better than anyone else. If you look in the back room here, behind the demo equipment, you’ll see what has been produced by the group making inexpensive Ethernets. It’s beautiful, compact and trivial to use. The number of units is growing. We want to make thin-wire Ethernet, using our components, the standard of the industry.
Digital’s stock has been selling 20% below book value, with a total market value of $7
billion. That’s cheap. With AT&T in the process of acquiring NCR, is there a real danger
of someone taking over Digital?
Life is filled with dangers. There’s always a danger.
Now, we’re gentlemen and ladies. We’re gentle and nice. But if someone did come after us, we’d fight tough - very tough.
Although I am humble, I do sincerely believe that we are critical to this country.
The issue of MGMT MEMO reporting on the Heald Pond Woods Meeting said that the company’s
goals are 22 percent return on equity and 16 percent return on assets. Surely, a profit
ability goal with no mention of revenue growth rate at which this is to be achieved is relatively useless in guiding business management action? Surely, our business objectives
include both revenue growth and profitability according to market opportunity and cash
The answer is yes. I believe that you add both growth and profitability when you measure the company. To a first approximation, you measure the corporation on the sum of the percentage of growth and the percentage of profit, or the percentage of return on assets. The test of that theory is, when you have zero profit and grow a large amount, you would rate that equally with a company that had zero growth and a large profit. I believe that’s true.
Thus far, we don’t get agreement internally on the details of how you make this measurement. What is important is that it does cost to grow. If you don’t grow, you can cut your costs a lot.
To the stockholder, this means growth is worth as much as profit, so we should evaluate both growth and profit.
The reason people don’t agree with this measurement is that experience sometimes seems to show that when you grow you make more profit. What experience does show is that if you grow more than you plan, you are very profitable. If you grow less than you plan, you are very unprofitable. But, over a period of time, growth does have cost.
How do the Business Units set the price?
The Marketing/Integration Business Units will set the price. They have the responsibility for planning the details, the investments and costs necessary to develop, sell, and deliver the product. Therefore, they have the responsibility to set the price.
In their price, they will leave a certain amount for selling or for giving discounts. Discounts and allowances will only be given when there is a savings made in the efficiency of selling. Discounts will be given by the account team, not by the hierarchy or the Field administration.
Salespeople incur costs and make a profit on the costs they incur. We’re not going ask them to make another 22% on the total system. If we did that, we wouldn’t be competitive.
But, of course, if we have the opportunity, if we have a unique product, we will charge more for that because it’s worth it to the customer.
Do business units stand alone, or can they share profit and losses in certain circumstan-
We allow freedom. We count return on investments. Sometimes business units will lose money, sometimes they will make it. If they lose money sometimes, they’d better make more at other times.
Basically, I’m giving our senior managers the opportunity to run businesses and show they can succeed and make profit with their value-added efforts.
The new management reporting system offers Digital a competitive advantage by helping our decision-making environment. It should give us the small-company advantage of being fast on our feet, while having all of the resources of a large company.
We are looking at the company in three different dimensions — the Product and Service Creation Units (which includes software, services and hardware), Application and Integration Business Units, and Customer Accounts. Business unit managers will be accountable for setting their direction and responsible for the results.
To be successful, a business unit manager, as an entrepreneur, must determine the level of activity the customer needs and is willing to pay for. Entrepreneurs cannot be told what they have to buy and at what price. They expect and demand the best in class.
Overall, the new management reporting system will help us understand where we are making or losing money and why we all do the activities we perform today. In simple terms, it will help ensure the customer is willing to pay for our activities.
We arrived at this system by considering the dimensions of Digital and their levels of complexity. The three dimensions are technology, uses of the product, and customer relationships. Most companies, when they get large, divisionalize around the dimension of their business that presents the highest level of complexity, so they can give focused attention to the needs of customers in that dimension. Digital, however, is very complex in all three of those dimensions. We produce thousands of complex products, customers put them to thousands of uses, and we have a multitude of different kinds of relationships with customers.
When a company has high complexity in all three dimensions, it is foolish to tell one dimension to manage the other two. Instead, we want each business unit manager to act as an entrepreneur would, with the freedom to focus on any one of the elements that make up Digital’s value chain.
A business unit might focus on basic technology or solutions or applications or accounts. Each business unit has the right to propose a new idea, without being checked by the other two dimensions.
At the center of our new management reporting system is the Executive Committee. They are going to provide the leadership to help us integrate and decide which directions represent profitable growth potential for future years,
But the goal of the new management reporting system is not just to provide information to Ken Olsen or the Executive Committee. Rather it is to provide the right information in a simple enough format to the business unit managers. This information must be straightforward enough so people can remember it and work with the various alternatives.
We will tell business unit managers what profit they have earned. We will take revenue less the cost of the input they have received and less the cost of their own specific activities. If the cost of input and activities is greater than the revenue, that means we haven’t added enough value, from the customer’s viewpoint. In other words, the customer is not willing to pay for the business unit’s work; and the business unit is not contributing to Digital’s profit. In that case, the business unit manager will be asked, "Why are we in this business?"
That may seem very simple, but in today’s system we cannot do that effectively because of the number of excuses related to allocations and "helpers." The business unit managers, as entrepreneurs, will have to demonstrate the ability not only to grow but to grow at a profit.
From a planning and budgeting standpoint, we have developed a framework to help us go forward while we working the details. Over the next three months, we are going to use the information that has been developed and the work that people have done in terms of sample profit and loss statements and simulations. But we do not have the luxury of planning a year and then implementing. With the stock trading as low as it is, we must start it now and work the exceptions later.
We will provide entrepreneurs the information they need immediately. It may be rough and approximate, but it will be information that they have not had before.
By the third quarter, we should be about 60% complete, providing information that can be used for seeing trends and looking at "what-if" scenarios. We will continue to improve the available data to whatever range the business unit managers say they need.
The point is — we are beginning now.
We have to make significant changes in the way we think about planning. The business unit managers will provide the lead with their new ideas, their enhancements to current ideas, all in terms of a profitable growth scenario. The functions will respond in terms of benchmarking, indicating where they are on their journey to best-in-class, and understanding variable and fixed costs. The Executive Committee will continue to review and update corporate models for FY91, FY92 and beyond. Understanding external factors as a context for the decisions they make, the Executive Committee will provide advice and counsel on business unit presentations and provide the framework in which we will do budgeting, in the fourth quarter, for next year.
The focus of "total quality management" is delighting customers on a sustained basis and achieving undisputed leadership in total cost. Our objective is to use quality in this sense as a major market differentiator.
In every enterprise, there is a value-added piece involved in meeting customer needs, and a non-value-added piece, where good people and assets are consumed doing things customers don’t want. We want to eliminate the non-value added activities and empower every manager and employee towards value-added activities only.
A value-added enterprise concentrates on meeting customer’s expectations — both their explicit demand (what they say they want), and their latent demand (where we help them interpret and determine their real future needs). Products and services are designed concurrently, not serially. The goals are set up front and 80% of the quality issues are determined right at the onset. The products are built in an environment of just-in-time
manufacturing processes, with the statistical process control in place. With service organizations that are value-added in their approach, customer problems are solved the first time, rather than on multiple visits. A value-added enterprise has highly trained sales people who are very responsive in dealing with customers. The whole company is highly focused towards the customer.
Digital’s approach focuses around four major initiatives intended to empower every manager and employee to be entrepreneurial in meeting customer needs, and innovative and disciplined in driving for continuous quantum improvement across the corporation. We are striving to exceed customer expectations without defects, without waste, with best-in- class practices, and total employee mobilization.
Those initiatives are:
o Voice of the Customer,
o Six Sigma, and
o Just-in-Time Cycle Time.
Voice of the Customer is our effort to make sure we "do the right thing." The other three initiatives are intended to make sure we do that in the most efficient way. Benchmarking aims at using "best in class" practices. Six Sigma is an approach to eliminating defects. And the "just-in-time" cycle time approach makes sure we compress our time-based activities.
Voice of the Customer consists of four steps:
o To make sure every one of us is totally passionate and relentless about solving customer problems while making a profit.
o To make sure we have a profound understanding of our customers’ needs and their perceptions of Digital’s products and services.
o To work with the customer to try and anticipate the customer’s future needs and make sure we meet them.
o To express those customer needs and how they can be executed technically.
Everybody else in the industry is trying to "satisfy" customers. We have to find tangible ways to differentiate ourselves by "delighting" our customers. The first step in achieving that is "contextual inquiry," where marketing or business unit managers work jointly with the development groups and with customers. In the classical approach to marketing, you ask customers what they want, and then respond to their explicit demands. But customers tell every supplier in the industry the same thing. To differentiate ourselves, we have to do more than that. We have to observe how the customer works, and imagine, with the customer, unique and entrepreneurial ways to solve their long-term needs. Then, having gathered all of that information, the marketing and development people, using the customer’s language, lay out all the technical ways of executing those items and then make the various trade-off decisions. This way, we come up with a specification and move forward to execute it quickly. This process forces heavy integration between the business units and the development groups and gets them all much closer to the customers.
Having figured out what is the right thing to do, the next step is to make sure we have best-in-class practices in areas that are critical to our success. Benchmarking is a formal way of looking at other companies to find out who has the best practices and then incorporating those in our own company.
The first step in benchmarking is figuring out the critical success factors for a particular organization. For example, in some manufacturing groups, the critical success factors are: customer acceptance, cost, quality, and time to profit.
To succeed at this, senior managers have to be personally involved, particularly in the first steps, to avoid the danger of benchmarking activities that we shouldn’t be doing at all. We need to know our own operations very well, and know the leaders, not just within our own industry, but worldwide. Then we must incorporate the best practices and set the new standard and gain superiority.
We want to encourage people to help identify major areas where our what we do differs significantly from best-in-class. Those are opportunities we want to go after, and we should encourage people to come forward and identify them.
Our Six Sigma initiative focuses on rapidly reducing defects and ultimately eliminating them. This includes "design for excellence," design for reliability, availability and cost. We want to get on a path of ten times growth in reliability every four years.
"Six Sigma" is a term from statistics. In the real world, no process can be exactly repeated time after time. There is always some variation of results around a norm. The goal is to reduce the amount of variation and also to set specifications at realistic levels so that there are almost no instances that fall outside allowable limits. A process with a "six sigma" distribution has only three defective parts per million.
We have already seen significant success in this area in the Storage Group, which is on a path of 60% reduction in all its in-process defects this year.
Historically, Engineering would design a product, and pass their requirements to Manufacturing. Manufacturing would then try to make it. Variations in manufacturing processes relative to the requirements would mean that a certain amount of material would end up wasted as scrap or rework. By integrating the efforts of the development and manufacturing people, roughly 80% of the quality issues can be resolved right at the beginning.
This forces a creative tension in the setting of product and process requirements and goals, designing the product and manufacturing process simultaneously to jointly meet quality and cost goals.
This joint effort does away with a lot of the uncertainty in the system. Manufacturing no longer needs to maintain extra capacity to handle rework and scrap, and costs are reduced exponentially throughout the product cycle. (Typically, what would have cost a dollar to fix in Engineering, costs $10 to fix in Manufacturing, and $100 to fix in the Field).
Six Sigma is not limited to the technical domain. It is applicable to every organization and every individual throughout Digital. It is an empowering concept which personally challenges each of us to eliminate defects from the execution of our tasks. We want all employees to look at the value-added they provide the corporation and find ways to improve their work process.
For example, Don Gilbert, who is responsible for janitorial services in the Springfield plant, mapped all the processes involved in this work and then benchmarked them. On his own time, he visited all the hotels and plants in the area. He came up with ideas for how to improve janitorial practices at his facility. He planned how to implement them and went to his boss and the plant staff and showed them how he could provide significantly higher levels of services for a lower budget. That’s the benefit of empowering individuals at all levels in the organization, letting them drive for excellence.
I believe we’re already a leader in the industry in cycle time. The focus of our "just in time" and "A(delta)T" programs is to compress all our time-based activities even further. This includes time to market, time to process an order, time to manufacture through a plant, time to respond to a customer problem, etc.
For example, in the Hudson plant, introduction of just-in-time manufacturing and cycle time techniques helped reduce a process that had taken five weeks down to one week. That 80% reduction in cycle time led to improved responsiveness to customers and dramatically reduced the level of scrap.
Today 25 manufacturing plants in Digital have very active Just-In-Time A(delta)T processes in place. Typical results to date across the company are 60% reduction in work-in-process time in manufacturing.
The A(delta)T concept involves breaking a process into its value-added and non-value added elements. You determine the theoretical level, which includes only the value-added piece. Then your goal is to arrive at that theoretical level, and then to challenge that level and strive for further improvements.
We have trademarked our term "A(delta)T," and half a dozen Fortune 30 companies are in the process of licensing and using it.
We’re also teaching our vendors. One of our major vendors for storage parts implemented it and three months later they had reduced their cycle times by 70% and reduced their costs by 17%.
The foundation of the Total Quality Management Process is employee involvement and management leadership. This process requires the mobilization of the entire work force to drive towards continuous quantum improvements. You bring teams together to strive for innovation and to implement it. We have such teams in place already across most of the manufacturing plants. As we now drive this approach in the service and sales organizations, we expect to see increasing teamwork and continuous improvement.
In closing, we as managers are responsible to lead the company to achieving best-in-class in every dimension. Today, there’s a tremendous ground swell of people getting involved in this. But we need every senior manager in the company to commit to get themselves educated in these areas and to provide vigorous leadership that empowers the people who are doing benchmarking and reducing cycle times. We have to help employees deal appropriately with the changes that occur as a result of those types of dramatic improvements. We need to integrate total quality management into all aspects of our business.
Always start with what’s important to the customers, and you can’t go wrong as you work your way through this.
Since 1987, the banking marketplace has been growing 6% to 8% per year in computer expenditures. Our marketshare has been growing at 20% compounded, and our revenue at 25% compounded. Two of the top ten banks in the U.S. now spend more money with Digital than they spend with IBM. Today, about half the money that is transferred in the banking system in the world moves on Digital computers. In other words, we are really making an impact in this market.
Chemical Bank operates their entire retail branch network using VAX computers. They expect that about 70% of their profits are going to be on the retail or consumer side of the business.
A year ago we installed a new trading room at Bankers Trust in Broadgate London. They are using that system — including 250 VAXstation computers — to trade about $20 billion a day. Bankers Trust came to us with a problem, and we jointly developed the solution. From that project, in addition to the sale, we developed our DECtrade product, which we now market to the rest of the banking industry.
There is enormous opportunity represented by emerging countries that do not yet make full use of electronic funds transfer. I recently had dinner with the brother of the new president of Czechoslovakia. I asked him, "How do you move money into Czechoslovakia as you want to set up new businesses?" He answered, "Very slowly." Six weeks is the average time to transfer money into Czechoslovakia. That is a great opportunity that we can now take advantage of.
Banks are facing many problems today, such as consolidations, mergers, and even failures, which creates ever-changing opportunities. In this environment, banks need to use technology to get a competitive advantage. They are not interested in technology for technology’s sake. Their goal is service and responsiveness. They are going to buy computing solutions from vendors who can show them how to increase the level of service and responsiveness they give their customers.
Banks buy solutions, not hardware. We need to understand the their problems. We need to worry about what they have already installed, and lead with our advantages, such as Network Application Support (NAS) software, and our ability to integrate. Rather than have them throw out what they have currently installed, we want to help them grow from our base. We need to look at their business from a global, round-the-world, round-the-clock perspective.
Success in this arena requires having a complete team focusing on these solutions. That team includes the Application Business Unit, Sales, Services and Digital Customer Centers (DCCs).
We all need to listen sensitively to the issues and problems of customers. We have to involve them as participants in the planning process. And we need to design platforms that are based on standard products. We can’t make money reinventing the wheel each time. We need repeatable solutions that we can sell over and over again. We can implement these solutions with the Enterprise Integration Services (EIS) organization, working with both us and the customer from the beginning through customizing and installing.
From this approach, customers get innovative, but practical solutions that solve their problems. They get total system integration. As a result, customers view themselves as valued Digital partners.
Information about money, markets, and customers is going to be the key strategic weapon for the 1990s. We have the opportunity with our products and with our solutions to be leaders in that marketplace.
Last month we announced that we would add the POSIX interface and other open standard interfaces to our VAX VMS product set. That means Digital can now call itself "the open systems company." In other words, every product line that Digital supports today supports open systems. With the current pressures around open systems, Digital should take advantage of this opportunity.
What do we mean by open systems? Customers would like to be able to implement any of their applications in a vendor-independent way. They would like to be able to develop an application on one hardware/software platform, then move it over to other hardware/soft- ware platforms when they choose to. This goal can be accomplished if indeed these hardware/software platforms provide the open, standard interfaces that the applications are designed to.
In other words, if we provide a set of open standard interfaces on the operating system level, a customer or a third party software company can develop an application that meets those interfaces and then be able to run it on any hardware/software platform that provides those interfaces. That is called "portability," one of the major features of an "open system."
Today, Digital supports its open computing environment through its Network Application Support (NAS) products. The goal of the NAS strategy is to be able to interoperate across any hardware/software platform, independent of its structure and its architecture. It enables us to support client-server activity and to integrate an enterprise, that has multiple computer architectures and computer systems from multiple vendors.
Today Digital has hundreds of third parties working to develop applications using NAS capabilities. In this case, NAS products are software packages coded to standard interfaces that allow various systems to interoperate and allow applications to interoperate with other applications across the network.
Some people have the mistaken impression Digital is not really serious about UNIX* and is just trying to push the VAX VMS alternative as its major operating system. Others believe
that Digital has switched to UNIX and that therefore VAX VMS software is a dying system. Both of these impressions are wrong.
Today, Digital’s computer repertoire includes both VAX VMS environment and ULTRIX/OSF operating system environments. These are the environments for which Digital will provide full service support.
In addition to being a founder of of the Open Software Foundation (OSF), which supports open interfaces as the vehicle for open systems, Digital is also the major supplier of the technology that OSF is putting forth. Our ULTRIX strategy is to be the first and best with implementations of OSF technology.
Today, applications written using the POSIX interface for any ULTRIX system, and later this year they will also be able to run on our VAX VMS system. Likewise, applications then developed for VAX VMS systems when using the POSIX interface can be easily moved over to our ULTRIX system or to other companies’ UNIX systems.
This is a key element of achieving open systems. That, coupled with NAS support of interoperability across a multi-vendor environment, provides the foundation for Digital to declare itself as "the open systems company."
*UNIX is a trademark of UNIX System Laboratories, Inc.
All the various flavors of UNIX software available in the marketplace today are in a state of migration. Our ULTRIX Version 4, which we ship today, is based on the same Berkeley UNIX code that is used by Hewlett-Packard and SUN. We are going to the Open Software Foundation’s OSF/1 system, as are IBM and Hewlett-Packard. Meanwhile Sun is moving to USL’s System V.4.
All of these current products are competitive, and all are in the process of moving to a new base. When it comes to the base operating system, it is really a pretty level playing field. In other words, we should not let operating system issues get in the way of selling in the UNIX space.
In fact, Sun’s transition to V.4 is a major change. They have given all of their Independent Software Vendors (ISVs) a thick portability guide, indicating this is not going to
be a very smooth transition.
Sun’s history is very interesting. They have successfully marketed their company as a leader in standards, yet Sun’s pattern is becoming clear. The press and analysts are starting to say that Sun believes in open standards only as long as they propose them themselves. Anything they don’t propose, they fight or ignore.
Digital offers all of our technology along with other resources of our Engineering groups to all the standards groups. We accept whatever the standards groups declare is the standard, and then we are very aggressive about implementing it. To me, that is the definition of an open systems company.
Our ULTRIX software is gaining momentum as the open systems leader. In the RISC/UNIX line, from Q1 of last year to Q1 of this year, we grew by over 300%. That is evidence that our message and our products are working for us.
The cynics who thought we weren’t serious with UNIX are starting to change their minds. The OSF consortium is working. OSF is real, and it is releasing technology very quickly, generating faith and enthusiasm from the press and customers. Meanwhile, Sun is becoming increasingly isolated from the rest.
Before OSF, standards bodies would take years to draft a standard. OSF has changed all that. OSF supporters believe that a good decision made now, based on the best available technology, is much better than a great decision made later. Rather than striving for consensus, they are looking at all the technology, making decisions, and moving forward quickly. Of course, if a standards body does endorse a standard, OSF supports it and incorporates it into their technology. X/OPEN’s Portability Guide (XPG) is a good example, and OSF/1 fully complies with this standard.
Today, with ULTRIX V.4, we have a great implementation of UNIX software. We also have great RISC hardware and over 1800 applications. The press recognizes the DECstation 5000 computer as the best workstation available and the DECstation 2100 machine as the best workstation bargain. These are tremendous sales advantages.
In a few years, the only major operating systems left (excluding PC software) will be VMS, UNIX and the IBM proprietary systems. The smaller vendors, such as Wang and Unisys, are all converting from their proprietary systems to UNIX. But VMS and IBM proprietary systems clearly have very long futures ahead of them. They have significant capacities, and no significant shortfalls.
Nevertheless, there will be a lot of movement in the customer bases. When customers move, they are not likely to change from one proprietary system to another. 1 doubt that we will lose much VMS business to IBM’s MVS and vice versa. If anybody decides to convert from these proprietary systems, they are going to go to open systems. So when customers do decide to move from VMS or MVS software, we should be there with our UNIX products to catch that business.
We should remember, too, that as the other smaller vendors migrate their installed bases from proprietary systems to UNIX, Digital has a great opportunity to get this business with our ULTRIX/RISC products. Some of these accounts may demand the capabilities and features that are currently available in VMS and IBM products. And they may not be ready to move to UNIX. In that case, we have a distinct advantage with our VMS standards message.
Meanwhile, Sun and Hewlett-Packard are selling UNIX aggressively. Their customers love UNIX, and if we are going to do any business in these accounts, we have to sell UNIX.
The bottom line is that the bulk of Digital’s opportunities for growth will be in the UNIX area. We have to use RISC/ULTRIX systems to go after this new business and growth. We need to learn it, and we need to sell it well.
When I first heard the words "Open VAX VMS," I must admit, I had to laugh. The thought of "Open" and "VAX VMS" in the same sentence seemed contradictory. It takes a major mindset change to think about Open VAX VMS systems, because it is a major change, and a huge business opportunity.
Many in the press and, unfortunately, many customers are beginning to ask about VAX VMS software and its future. They are wondering: is VAX VMS software a proprietary, mature product, that is primarily sold to the installed base, whose growth years are behind it?
My message today is that what is happening in engineering right now, the things that we are adding to VAX VMS software in the next 12 to 24 months, in the area of openness and prlce/performance, transforms it into a major growth product, because we believe we will have the premier commercial strength open system on the market.
Our strategy can be summarized into three points: Open VAX VMS software is open. It retains superiority in commercial strength functionality and quality. And, it has leadership price/performance.
There are as many different definitions of "open" as there are people. Today VAX VMS software is very open in the sense that you can connect almost anything to it. We support any desktop from terminals to PCs, Macintoshes, VAXstations and UNIX workstations and IBM mainframes in our Network Application Support (NAS) environment.
In 1991, we will add key open interfaces to VAX VMS software, when we add MOTIF and POSIX. For example, POSIX on top of VMS software will allow you to much more easily move UNIX software to the VMS system. Today, there are several different types of UNIX in the marketplace, but it is fairly easy for a software developer to port his or her application from one to another — for instance, to or from Digital ULTRIX software, Sun/OS, HP-UX or IBM AIX. But it’s difficult to get to and from VAX VMS software. POSIX radically changes that. POSIX makes VMS software a member of the portability club. It becomes as easy to port your open application to VAX VMS software as it is to another flavor of UNIX.
At our Open VMS announcement last month we actually had a Sun workstation next to our new VAXstation system, and we showed, with an early version of our POSIX interface, how you can easily move the application from one to another. But if you could simply move it, how interesting would that be? The amazing thing is that you retain the underlying commercial strength functionality and quality of VAX VMS software.
What do we mean by commercial strength? We have the most dependable systems in our VAX VMS family, as we are the only ones to have clustering, shadowing, journaling, fail-over, recovery, two-phase commit and fault tolerance. We have the flexibility of client-server systems, timesharing systems and now mainframe systems with unique compatible growth. Our network prowess is obvious from local to global to network management. Our Cohesion software development environment allows you to develop large scale applications in up to half the time of competitors. You retain these commercial strength capabilities in the open environment. For example, when a Sun application runs on open VMS software, it automatically, without recoding, picks up clustering, shadowing, journaling and fail-over. That’s what allows us to become the commercial strength open system.
The third part of our strategy is price/performance. Many people give us credit for having the best balanced performance in the industry, from top to bottom. From speed, to I/O, to storage connectivity, we have exceptional, balanced performance. Unfortunately, in certain market segments, particularly in our traditional technical segments, which still today represent over 40% of our business, some customers say that VAX VMS products are slow, not as fast as RISC machines running UNIX. Over the next two years this will change radically. We expect our performance will pass all of our competitors, as we add leadership RISC technology to the VAX VMS environment.
In summary, we are opening VAX VMS software, retaining our commercial-strength functionality advantage, while leapfrogging the industry in price/performance. To me these are all sort of unbelievable things, but the group who is making these claims is the group who invented the VAX family, invented VMS software, invented MicroVAX computers, and brought you most every VAX system you have had in the last fifteen years. This is the plan. This can be a very strong marketing message, as well.
Let’s consider what this message means to two typical customers. The first is a loyal VAX customer who says, "I love my VAX VMS software environment, but I’m worried about its future because it’s closed and slow." This customer might have millions invested with us, and millions more invested in software, data and trained employees. They are getting pressure from their management to consider other systems. IBM and others are throwing all kinds of information at them raising questions about where the VAX VMS environment is going, and they want reassurance. For them, open VAX VMS systems is the ideal message, as it addresses both of their concerns. It gives them the confidence to buy today, knowing they are staying on the leading edge of technology.
This message is also very well received by the customers of the proprietary vendors, such as Unisys, Data General, Bull, and even IBM. These customers are often very uneasy about the future, as unlike with VAX VMS their vendors are not rapidly adding openness and leadership performance. They are concerned about being dead-ended. For example, one Unisys customer told me, "I’m running a mission-critical application on my Unisys platform. Unisys is trying to sell me their UNIX platform, but I just don’t think it has the functionality and quality that I need. It’s just not ready. I’ve considered VAX VMS systems. You clearly have the functionality I need. But while you serve good food in your prison, I don’t want to be in prison again." In other words, this Unisys customer recognizes the high quality of our software but is very concerned about a "closed" proprietary environment.
Once again the ideal message is open VAX VMS systems. They can have the necessary commercial strength functionality now, while buying into an open and price/performance oriented environment as well. If this customer wants even more openness, because open interfaces on top of VMS software isn’t enough for some reason — then we of course have our ULTRIX products. The same standard open interfaces that we are putting on VMS software, are already on ULTRIX software. And so you don’t get stuck on either side, as you can switch later. I believe sales reps will sell a lot more of both VAX VMS systems and ULTRIX systems when the sales rep can talk about both.
Our new organization — the Production Systems and Mainframe Systems Business Unit is focusing on a large market that we have, for the most part, not been tapping.
Roughly half of the total available computer market falls into the general category of "production systems." We often use the word "mainframe" in the same context, referring to "mainframe kind of systems."
When we first started looking at this kind of business many years ago, Carl Gibson, in a now classic memo, referred to "the mainframe of mind." In other words, the customer has a certain kind of expectation not only from the computer system, but also from the vendor who delivers it. And that combination basically calls for a new set of behaviors on our part that we have not been practicing too much in the past. We have to get into that "mainframe of mind."
In the production systems business, the customers are typically the MIS directors, chief information officers and CEOs who get involved with mission-critical purchases.
The major competitor in this arena is IBM. IBM does this kind of work very well, in all the categories that these customers require. Tandem has learned how to compete in specific parts of this marketplace. The Japanese big three are very interested in the marketplace, but they haven’t got all the resources in place to compete yet. They have so far limited themselves mostly to being plug-compatible to the IBM environment.
To a certain extent, some of advanced science applications fall into the same category. Convex and Cray compete for those applications. But our principal competitors are IBM, Tandem and the Japanese big three.
The customers want a real partnership with their suppliers, in which there is some sharing not only of the rewards, but also of the risks. The people who make the buying decision are betting their lives and careers on that decision. They need that decision protected.
We have many capabilities in our architectures and our systems to protect the customer — such as the broad range of products and performance that we can provide as the customer grows or centralizes or decentralizes. We can provide them with the environment they need, at the particular time that they need it, without rewriting software or retraining people. That is a major advantage. We also have an "availability continuum" which is unmatched in the industry.
But customers need more than that. They are looking for solutions to their business problems — a full range of solutions. They want partners who will look beyond their companies’ own specific offerings and help with other things as well. They are looking for mainframe class services, as well as products — sendees of the kind that our Enterprise Integration Services Group (EIS) can provide.
They need fundamental products that are highly reliable and highly available. They need software that works. They need assurance that as they adapt to new technologies, they won’t have to go through major transitions internally and spend time in that transition and lose money in the process. They need to make sure that the "Digital" they are working with acts toward them as one company, with a single focal point for them to contact. They want their account manager to be the boss with access to all the resources from Digital. They want that account manager to be able to get the products, services and support, and, most of all, the responsiveness that they demand in their environment.
At DECWORLD in Japan last week, I learned that some of our larger customers now believe that Digital is a viable vendor across all of their application needs, not just their science and technical needs. They are talking to us in a different way now. They are looking for vendor stability and for performance that meets their needs and the ability to grow. They want us to be able to respond to them. We have to leam that their systems environment is their business.
These companies are looking for flexibility. Many are consolidating systems. Many are distributing. They want to be able to do both, and we have the best products to satisfy both of those requirements.
We used to think of reliability and availability in terms of hardware. Hardware is part of it, but our customers are concerned about the total system, and how the application plays in their environment. They don’t measure availability in terms of "is the central processor running?" but rather "Am I running my business?" That’s a different model. Fault tolerance is important, and disaster tolerance is becoming very important.
In Japan, because of earthquake problems, the government has just mandated that major corporations must have disaster tolerance sites. That is an opportunity for us because we are the only supplier with a simple, elegant solution.
To have a "mainframe of mind," we have to be responsive to every customer problem associated with their system, regardless of whether the problem originated with us or a competitor or a third-party supplier. Often, these customers are not price sensitive, but they are very service and responsiveness sensitive. When we bid for the system, we have been make sure that we include charges to cover our costs for doing that and make a profit at the same time.
We have been preparing for this business for a long time and in various ways. In the sendees domain, we have an Enterprise Integrated Services organization that can deliver everything we need. We have the kernel of products, people and processes. We need to make improvements in all areas, and we need to provide the things they need to easily plan and manager their environments.
We need to remember that a "mainframe of mind" does not necessarily mean we only sell a "mainframe" computer. We are selling everything that Digital offers in an environment where the customer has specific "mainframe of mind" requirements.
With our products, we offer a range of performance and growth, as well as networking capabilities.
The role of our new business unit is to focus attention on a class of customers and competitors and a certain category of needs. We have to specify those needs and represent them to the rest of the company. We have to help develop and implement the strategies that will satisfy those needs and delight those customers. We also provide some systems knowledge to the EIS organization and to the Application Business Units to help them deliver systems that meet the requirements of customers of this kind. We will provide marketing support to those areas, as well as to the Product Business Units.
And we have to help the account manager by providing a level of expertise to customers of this kind, when called upon.
Our goal is to profitably delight customers by helping them match their needs to the technology systems and services that we offer.
The "mainframe" is an environment, not just a box. It includes hardware support, sales support, software support and consulting support. Its throughput is measured not in terms of how fast the job runs, but how much work gets done. Throughput here really means organizational productivity.
The key characteristics of a mainframe are reliability, availability and maintainability. When we set out to design the VAX 9000 system, the IBM 3090 was the recognized best-inclass mainframe. Our goals were to do better than that, and we succeeded in terms of high throughput, numbers of users, systems performance, and balanced performance.
The VAX 9000 system offers large memory size (up to two gigabytes), along with high memory bandwidth (500 megabytes per second), high I/O bandwidth, and sustained performance. These are all mainframe characteristics.
Today the manufacturing problems are behind us. We expect that in Q3, we’ll be able to make as many VAX 9000 systems as we can sell.
We anticipated that during the first year most of the VAX 9000 business would come from upgrading the installed base. As it turns out a third of the sales to date have come from new business: new applications, new accounts, or competitive displacements, where our equipment replaces that of a competitor. Also, in a number of cases, we’re selling multiple VAX 9000 systems to the same account. That’s because these systems are being used in production environments, where they can't go down. In July 1988, just two and a half years ago, Digital got into the transaction processing market in a big way. Today we’re one of the top three vendors, and we’re taking on IBM, the top vendor, at the heart of its business. To date, 36% of the VAX 9000 business has come from transaction processing applications.
We just recently had a customer advisory board meeting in Florida, where we brought in senior managers from ten of our first accounts to hear what we could do better to be first-in-class. They told us that despite some of the slips in the program, the installation planning and implementation set a new standard. It was better than the IBM mainframes they had installed. They said the performance was better than they had expected. They felt that, as a company, we were listening to them.
They also noted that the mainframe business is a very different environment for Digital. It requires real partnership, not just taking the customer out to lunch. It requires helping the customer be successful, planning and sharing risks, and communication. We were criticized for our communication. We need to make sure that we "say what we do and do what we say." They were also concerned about commitment. When they bet their business on these applications, they want to know five years from now that we'll protect their investment and take them into the future.
They also emphasized the importance of "systemness." They spend as much on disks and I/O as they do on the central processor. They want all of the pieces from Digital to work, not just a fast central processor.
And in the area of service, they "don’t want MicroVAX support for a mainframe environment." They want mainframe support.
Overall the news was good. All ten customers agreed to be reference accounts, and overall were delighted.
We have the product today, and it is a VAX computer. It runs both VMS and UNIX software, so the operating system is not an issue. The challenge is. with focus and expertise, to get new mainframe business in FY91.
Today, only Digital offers scalable availability. That means that a customer can select a conventional (VAX 9000) system, a high availability (VAXcluster) system or a continuous processing or fault tolerant (VAXft) system. No other supplier offers that range of choice.
Tandem does have high availability solutions, but for fault tolerance, they offer an incompatible product line called the Integrity S2. Stratus has full tolerance, but they do not have the range of availability that our VAXcluster systems offer. IBM has a variety of solutions, but they are not compatible. Our products are.
High availability systems use redundancy to minimize downtime. Fault tolerance, on the other hand, means continuous processing.
For example, think of a light bulb. With fault tolerance, you would say it could never go out. With high availability* you have a spare bulb nearby and can switch quickly, if need be.
Only Digital offers both kinds of solutions, with compatible systems. That is fundamental to our product advantage.
A customer can select a high availability solution today and tomorrow, if continuous processing becomes critical to the operation, he or she can select a fault tolerant solution. And the code in both environments will be compatible at the binary level.
We are leaders in this area. We have installed more high availability solutions than Tandem and Stratus combined. We have installed over 15,000 VAXcluster systems, which include between 80,000 and 100,000 computers in total.
The success of clusters derives from their range of availability as well as their range of performance. By adding processors or storage disks, you can readily upgrade your system performance. Clusters also protect the customer’s investments because old processors and old storage devices can be combined with new ones.
Meanwhile a growing class of customers requires continuous processing. For them, the cost of the downtime is simply too great. We expect that this continuous processing environment will become the standard in the near future.
In the early days of computing, businesses always had a manual alternative for everything they did on a computer. They had run their businesses manually before, and if worse came to worse, that could make due that way for a while when the computer went down. But today move than 90% of business applications perform tasks that have never been done manually. If the computer goes down, there’s nothing to fall back on.
This is true in mission-critical applications like electronic funds transfer in the banking business. But it is also true of many retail businesses. For example, at your local video rental store, if the computer is down, the business probably comes to a halt. They have no other way to keep track of their transactions. In other words, fault tolerance is not a luxury. It is becoming critically important in many markets.
summary, the reason Digital
continues to be successful in spite of a very tough business
environment is that we do not just target niche markets.
Rather we have compatible products that cover the entire range
of performance and scalability and that run both VMS and UNIX
software. In the 1990s, success depends on solving not only
just piece, but rather the entire range of the customer’s
firstname.lastname@example.org privacy statement