by W. E. Pete Peterson
Copyright 1993, 1998 W. E. Peterson
By now we had our software business down to a simple routine. Every year we released a new version of WordPerfect, and every year we doubled our sales. It was almost as easy as printing money on our own printing press. Except for two small clouds on the horizon, our future looked bright.
Almost everything having to do with the release of 4.1 went smoothly. At first we billed it as a minor release, because 4.0 was doing so well. We were afraid that if we made too big a deal of the new release we might scare some companies into thinking that 4.1 was very different from 4.0. Often when a significantly different version of a product was released, a company would conduct formal evaluations of the product to determine whether they would use the upgrade or switch to another product. If our customers perceived 4.1 to be a minor release, we hoped they would decide an evaluation was unnecessary. We did not want to disrupt their purchasing patterns now that they were buying our product.
Our customers quickly saw through our "minor release" strategy, but they did not slow their purchases. They went crazy for the new product. The first quarter 4.1 was available, which was the fourth quarter of 1985, our sales jumped more than 50% over our best previous quarter. After InfoWorld gave us another perfect score, sales went up to $10 million for the first quarter of 1986. We were now dead even with WordStar and had enough momentum to pass them. Our only problem was filling all the orders that were rushing in. The local printing companies could not print the manuals quickly enough to keep up with the demand, creating a backlog for update copies as long as six weeks. As a result of the 4.1 rush, our PC market share was now somewhere between 20% and 25%.
With WordStar almost out of the way, we were ready to set our sights on other competitors. For the short term, Wang was still an important player in the word processing industry. Personal computers could not yet match all of the capabilities available on the larger Wang machines, especially in the areas of networking and security, and many of Wang's customers felt they could not afford to abandon their sizable investments in Wang equipment. Law firms were especially loyal to the old machines, because many lawyers were not ready to have computers on their desks. They liked the old way of dictating their documents to their secretaries and having final versions produced by the firm word processing center.
Wang was trying desperately to keep its business going. It was a multi-billion dollar company, accustomed to profits greater than the revenues brought in by the entire PC word processing software industry. Little multi-million dollar companies like Micropro and SSI were devouring its business, but Wang was not prepared to cut its size quickly enough to compete in the personal computer market. It was a dinosaur, unwilling to adapt to the new environment, and all that was left for it to do was survive until its money ran out.
Another big competitor in the short term was IBM with Displaywrite. Most of the dedicated word processing companies like CPT, Dictaphone, Lanier, NBI, and Wang had not been able to make a transition to the PC word processing world. Many had decided to close up shop rather than compete in the new, low-priced market. Some, like Wang, had chosen to ignore the PC and continue on as if nothing had changed. A few companies, notably NBI and IBM, had tried to write software for the PC, but most had not been able to sell many copies. IBM was the only company from the dedicated word processing world to also have a significant market share in the PC word processing world. Although they did not publish their software sales figures, we were fairly sure IBM's market share was close to ours. In spite of the fact that Displaywrite was usually found to be an inferior product by the press and by their customers, IBM continued to sell a lot of software. There was still some magic left in the IBM name.
In the long term, Microsoft was the only significant threat to our moneymaking machine. It was our only competitor with the resources, talent, endurance, and understanding to write a good PC product. Micropro had the resources and the endurance, but was famous for losing its talented programmers. Ashton-Tate, which purchased Multimate, also had the resources, but similarly could not keep good programmers. Lotus had most of the qualifications to write a good product, but did not understand the market. The spreadsheet maker chose to write a word processor aimed at the technical user, but instead missed every market. IBM had more than enough resources and endurance, but did not understand how to write good software. IBM shuffled its programmers in and out of large programming teams and evaluated team performance on the basis of the quantity of lines of code written rather than the quality of the code. At one point the company broke up the entire Displaywrite development team by moving the project from Austin, Texas to New York. If Beethoven could have written a good symphony by putting one hundred composers in a room and requiring each to write ten musical phrases that he would later piece together, then perhaps IBM could have written good software. I think IBM did not understand that creating beautiful software, like creating beautiful music, was an art form and not a manufacturing challenge.
Although Microsoft had a better appreciation of how to write good software, in 1986 they had too many irons in the fire to concentrate much on our market. Though they were much larger than we were, we always had at least three times as many programmers devoted to word processing. Perhaps what held them back more than their lack of resources was the fact that they had an arrogance about them, believing they were smarter than everyone else. They were a group of intellectuals with degrees from the most prestigious institutions, who had trouble believing that a group of ex-BYU students could beat them. On one occasion, one of their programmers witnessed a demonstration given by Bruce at our COMDEX booth. The programmer, upon seeing some of the features, interrupted Bruce to tell him that some of the things he was demonstrating were impossible to do. It took real arrogance for the Microsoft programmer to dispute something he could see with his own eyes, and it was this kind of arrogance that showed in Microsoft's product.
For 1986 we set our sights on IBM and Wang and directed practically all of our marketing efforts towards large accounts. We felt if we could capture the large accounts, that schools and small businesses would follow their lead. To this end, we aimed all our advertising directly at large corporations. While our reps still visited dealers, they spent more than half of their time working with large accounts. We provided big companies with free demonstration copies, extra telephone support, and special pricing. We let them know we either offered products or intended to offer products on IBM's entire line of computers.
IBM helped us win over many of its customers. Since 1983 we had offered WordPerfect directly to IBM employees for $125 a copy, and by now probably more IBM employees used WordPerfect at home than used Displaywrite. In 1984 we let IBM offices buy WordPerfect at the same low price, and until 1989, when IBM forbade its offices to use our product without written permission, WordPerfect was likely on as many machines in IBM offices as was Displaywrite. Around 1985, IBM started buying our software as a distributor, reselling WordPerfect to its own customers. Even though IBM preferred to sell Displaywrite, many of the company's employees and offices recommended our product.
While these recommendations helped, the biggest favor IBM did for us was not to improve its product very much over the years. Although IBM's customers wanted to remain loyal and were willing to put up with poor performance and an awkward interface, we were adding too many new features for those customers to stay with Displaywrite. Each new release of WordPerfect put more pressure on a company's decision makers to dump Displaywrite. IBM could not improve its product quickly enough to prevent its customers from jumping ship.
If the computing world would have stayed as it was in 1986, we would have had little trouble from our competition. There were, however, two big changes coming which would eventually give Microsoft a chance to catch us. The first was the laser printer and the second was the graphical user interface, or GUI (pronounced gooey). These two improvements were still small threats to our business, but support for both would require an enormous amount of work. At the time, it was a little discouraging to think that just as we were beginning to win the word processing game, the rules were changing.
Adapting to the new rules left our programmers with no time to enjoy the success of 4.1. Laser printers were starting to sell, but our three year old printing technology was not capable of supporting the new printers well. Although we could produce a document which looked like it was typewritten, we could not produce a document that looked like it came from a printing company. We were limited to only eight fonts in a document and only one font on a line. Our proportional spacing support was not perfect, and our program had no provision for kerning (the process of fitting proportionally spaced letters closer together) or leading (adjusting the space between lines). The program also had no support for graphics (pictures, figures, or drawings). To support these features, we had to redesign and rewrite the entire printing portion of the program and much of the screen portion. It was like 3.0 all over again, only this time it was a lot more work.
The job was huge. Every function code in the document needed to be changed. All measurements, including the ones for margins and tabs, needed to be in inches or centimeters, instead of lines and spaces. Fonts needed names and point sizes, like Helvetica 14 Point. WordPerfect needed to learn how to download fonts to printers and to support font cartridges. It needed to know the width and height of every character in every font.
By February we knew we would never finish the project to meet a Fall COMDEX deadline. We were afraid of the consequences if we broke our Fall COMDEX routine, so we decided to split our WordPerfect DOS developers into two groups. One group, the smaller of the two, worked on 4.1L, a version set to release in the fall of 1986, which would use the old printing technology. The other group worked on WordPerfect 5.0, which was what we called the product that would include the new printing capabilities. We hoped that 5.0 would be ready for release in the fall of 1987.
The L in the 4.1L stood for the word "legal," because most of the new features in the release were aimed at the legal industry. We hoped that 4.1L would get us into more law offices and that new features for line numbering, improved paragraph numbering, and an automatic table of authorities would be the last nails in Wang's coffin. For customers outside the legal market, 4.1L would truly be a minor release.
Neither of these new versions included support for a graphical user interface. We did not have the resources to improve the printing and add GUI all at once. Of the two technologies, our customers were more interested in support for the laser printer, and doing that job provided more than enough work to keep our developers busy through 1987. We had no choice but to hope that the impending GUI revolution would take some time.
At the time there was no clear GUI environment winner. IBM with TopView, Microsoft with Windows, and Digital Research (the CP/M guys were still around) with Gem were all trying to become the GUI of choice on the IBM PC platform. We were pulling for anyone but Microsoft to win. As the company most likely to give us trouble in the future, we did not want them in control of the GUI environment.
Whenever a customer or a writer from the press asked me if we intended to support Windows, I liked to compare Microsoft to the fox in the story of the Gingerbread Man. In the fairy tale, the Gingerbread Man jumps from the oven and runs away from the old couple that baked him. The old farmers try to catch him, but he yells the famous line, "Run, run as fast as you can, but you can't catch me, I'm the Gingerbread Man." As he runs away, he passes many animals who try to eat him. As he runs away from each animal, he repeats the line "Run, run as fast as you can," etc. Unfortunately, he comes up short on running room when he reaches a river. At the river he meets a fox, who offers to take him across the river so he can escape his pursuers. Although wary, the Gingerbread Man jumps on the fox's back and accepts the ride, believing the fox's promise not to eat him. Half-way across the river, the fox lets his back sink into the water, and, claiming weakness, asks the Gingerbread Man to jump onto his shoulders. The fox sinks lower in the water, forcing the Gingerbread Man to move to the top of the fox's head and eventually to the tip of his nose. Just before the Gingerbread Man reaches the safety of the opposite shore, the fox opens his mouth and gobbles him up.
WordPerfect was like the Gingerbread Man, running to stay ahead of the competition by adding more and more features each year. When we came to the GUI river, Microsoft asked us to trust them and let them take us safely to the other side. They approached us, as well as a number of other popular software companies, under the guise of collaboration. They offered us help if, in return, we would develop a version of WordPerfect for Windows. While we knew Microsoft wanted Windows to succeed, a feat which would require the development of Windows-based applications, we realized there was little chance that they wanted us to beat them in the word processing market. In spite of their assurances that Windows, rather than word processing, was their top priority, I did not want to trust them. I had trouble understanding how they could keep a straight face while giving us their sales pitch. They were an aggressive company, and their appetite for success knew no bounds. They were a sly, rich, hungry, and intelligent fox. I was not going to encourage SSI to accept their offer if there was any hope that another company might give us a ride.
By the spring of 1986 we were starting to do fairly well in Western Europe. We had versions of WordPerfect for every Western European country except Italy and Portugal, and those translations were in the works. Our royalties back from our European distributors amounted to about $5 million a year, and at least $3 million of that was profit. Although we were not well-known in England, Germany, and France, we were enjoying significant market shares in most of the other European countries. In some of the Scandinavian countries, we had market shares greater than 50%.
Now that we were a forty million dollar company, there were times when I felt a little overwhelmed. I appli ed and was accepted to BYU's Executive MBA program, thinking I had to know more if I was going to effectively run a large corporation. I decided not to go to night school, however, once I realized SSI would likely be a one hundred million dollar company before I was even half way through the program. For better or worse, I would have to rely on what I could learn from books and from our trials and errors on the job.
I continued to go to my dad for help. He had done a lot of formal planning at Harris and offered to help me understand the process he used. Dan Lunt and I flew to California to listen to what he had to say. He showed us this very complicated flow chart, describing his planning process. To meet the first step of the process a company defined its purpose, goals, and objectives. The next step required the company to predict how well it would do in meeting its goals and objectives if no changes were made to future plans. If the company's predictions fell short of the goals and objectives, which was usually the case, then the third step was to determine how to improve performance. Depending on the time frame, the company might increase or improve its development or marketing efforts, or it might consider acquiring other products or other companies. My dad emphasized that each of the steps required good communication from the top to the bottom of the organization. This communication would ensure that everyone in the company understood and agreed to the purpose, objectives, and goals, that the forecasts were accurate, and that everyone would try to make the new marketing, development, or acquisition plan work.
The approach was entirely new to me. For one thing, we started each year without any specific sales goals. Bruce and Alan never asked for a number, and I enjoyed not having to be held to one. I was also afraid that focusing on a sales goal would hurt rather than help. If we focused instead on working hard and doing a good job, we could succeed even if we did not hit an arbitrary number. Success is measured by expectations, and I wanted us to feel successful if we did our best and worked well together. I did not want us getting discouraged if our numbers ended up 2% short of some prediction.
We also had no thoughts of acquiring other products or other companies. I liked to keep things simple. It was the only way I could deal with the complexities of running the business. We had enough trouble writing, selling, and supporting our own products without having to integrate in another company or another product line. I did not want to ask for problems we did not know how to solve.
Although much of the process my dad explained did not apply to us, it was obvious we needed a written definition of our purpose and objectives. It was no wonder our employees had trouble understanding how the company worked and what they were expected to do, because we never told them. We could live without the sales goals and the acquisitions, but we definitely needed to publish and teach what it was we were trying to do.
As we set out to define our purpose and objectives, it was obvious we needed to change our name. Satellite Software International did not fit us and certainly did not tell the world what we were doing. SSI Software was only a small improvement. WordPerfect, Inc. was the obvious choice, especially if we ever intended to focus in on word processing, but for years Dan had rejected the name because of the way it sounded on the telephone. If someone was not careful about their pronunciation, WordPerfect sounded like "We're perfect." We spent a lot of time considering names like Artistry Software, Futuresoft, ProSoft, and even Image Resource Company. We finally gave up trying to find a different name, however, deciding instead to be careful about our pronunciation. In 1986 we officially became WordPerfect Corporation. We used the word "corporation," because it was about the same length as WordPerfect, and the two words looked good together with one on top of the other.
By now our employees were scattered all over the city of Orem. The customer support department was located a mile to the southeast of our main offices, grabbing practically every square foot of an office park called Lincoln Square. The number of support calls had jumped to 8,000 a week with the release of 4.1 and was continuing to rise as sales increased. The accounting and publications departments along with SoftCopy, our manufacturing company, were about one mile to the northeast, and were expanding as well. The Macintosh, Amiga, Atari, and VAX groups would all end up at the very south end of Orem, and the DG group would move across and just down the street from our main offices. By the end of the year we would have more than 300 people in the company at eight different sites.
There were some advantages to having the company spread out like it was. The distances forced the managers of the various departments to take on more responsibilities and solve more of their problems themselves. We also saved a lot of money. If we had tried to build a facility big enough to house the entire company, we would have had to borrow money or find investors, both of which were against our nature.
We were, however, to a point where we could afford to begin building a complex that would bring the company together again. The site which was most promising was the research park in northeast Orem. Orem City had purchased a 110 acre orchard in a residential section of the city and had used federal matching funds to put in the improvements for a business park. In spite of the objections of the surrounding residents, the city officials intended to build and lease office buildings in the park and use the profits from the venture to increase the city's tax base. We liked the location, but when they offered us space in 1985, we did not like the terms. Instead of offering incentives to move in, as some neighboring cities did, Orem was requiring a very long lease at what we thought was a high price. No other companies were interested either. Novell, which had been the most sought after company for the park, had moved to Provo instead. The orchard had a beautiful infrastructure, with streets, sidewalks, streetlights, electricity, water, and a sewer system, but no tenants. Except for the fruit, the primary benefit of the research park was to give local teenagers a secluded place to park late at night.
Late in the summer of 1986, I decided to call the city to see if they were desperate enough to abandon the lease requirements and sell us a few acres outright. I offered to pay $25,000 per acre if we could own the land. After some negotiating, they agreed to let us buy 22 acres for about $22,000 per acre, if we would agree to certain other conditions. We would eventually purchase about 95% of the park and build a number of buildings on the site, though we would add buildings only as we had the money.
In August of 1986, 4.1L was close to completion, but, much to our chagrin, one of the trade publications leaked the story of the release before we were ready to ship. It was probably only fair after the way we had used the press to slow down the sales of 4.0, but we were still upset about the leak. We were not prepared for sales to go down in anticipation of a new release. We decided right then to change the version number, if for no other reason than to make the publication look bad.
We also decided to add a few more features. Alan Ashton, Doug Lloyd, Layne Cannon (a programmer and another ex-student of Alan's), and I had taken the 4.1L version on the road to show to four or five of our largest accounts. After looking over the new version, these customers all wanted more. They were willing to wait another year for 5.0, if we were willing to add a few more features to 4.1L. Most of the changes were fairly easy to make, like increasing the number of columns supported and adding two new types of tabs. We added the new features in about thirty days and were still ready to release by COMDEX.
The changes made our manual a lot more interesting. Originally, a few update pages would have handled all of the 4.1L changes, but with the 4.2 additions, we needed a new manual. Unfortunately, we did not have enough time to produce one. At first we sent out the old 4.1 manual with update pages, which the customer had to insert in the proper place. It was a problem for customer support to deal with two manuals, the one with update pages and the new manual which eventually came out, but we felt we had to ship before COMDEX. We were a little superstitious about what would happen if we broke our fall release tradition.
We hoped that no one would care too much about our printing deficiencies, but Microsoft was too sly to let that happen. Laser printers were becoming more popular everyday, and Microsoft was doing a better job of supporting them than we were. They were also doing a good job of making sure our customers found out about our weakness. In spite of their propaganda, our sales did not decline. Third quarter shipments had jumped to almost $14 million, and with the release of 4.2, they would jump again to almost $18 million.
The dark clouds on the horizon, laser printing and GUI, looked less ominous once 4.2 was released. Microsoft was having trouble getting a good version of Windows out the door, and our customers were not getting impatient for fancy fonts in their documents. It looked as if we would have time to catch up.
As in years past, we were not content to try and sell only a DOS version of WordPerfect. By now we had many other versions. Microsoft was betting that the larger computers would go away. They had versions of Word for the PC and the Mac, and they were working on a Windows version, but they had not publicly shown any indication of putting their word processor on more machines. Although we were inclined to agree that larger machines were not too important in the long term, we did think that supporting them in the short term would help companies choose WordPerfect as their standard. We wanted to be the word processing standard for the world, and we hoped to support all of the significant computing platforms. In 1986 we had software released or in the works for Digital Equipment's VAX, IBM's mainframes, Data General's minicomputers, and many UNIX machines. In addition to the big machines, we had software released for the Apple II and software in the works for Apple IIGS, Macintosh, Amiga, and Atari. We were also still supporting many of the old DOS machines that were not IBM PC compatible, as well as many different types of PC networks.
We were not too careful about how we made a decision to support a new platform. At times, we were completely out of our minds. For example, we started the Amiga project because a few of our Apple II and Macintosh programmers were sneaking source code out of the office, so they could translate it at home for the Amiga. Their plan was to bring a finished product back to work for us to sell. When Alan, Bruce, and I discovered our programmers felt strongly enough about the Amiga to port WordPerfect on their own time, we assumed that enough other people wanted to use the machine as well. We should have done at least a little research or taken a little more time with the decision. By deciding to go with the Amiga, we were taking on a lot of unnecessary and unprofitable work.
The DG division benefited from our efforts to have WordPerfect running on every platform that made sense. The DG group, which had been neglected for years, was allowed to hire enough programmers to bring WordPerfect up to date on that platform. They were also given enough resources to think about writing an office automation package.
Office automation was the hot product in the early 1980s, when large computers ruled the world. The goal had been a paperless office, one in which all employees had computer terminals on their desks and in which documents were written, memos were sent, and meetings were scheduled electronically using one software package. Wang had Wang Office, Data General had CEO (Comprehensive Electronic Office), DEC had All-in-One, IBM had PROFS, and the list went on. The OA market did fairly well until the PC came along. Then the automated office was temporarily forgotten, as most customers decided they cared more about 1-2-3 than they did about electronic messaging.
Our DG programmers became interested in office automation, because one of their word processing clients, the US Department of Justice, was interested in the concept. Thinking they had to provide electronic mail and electronic scheduling to keep the DOJ's word processing business, the DG crew began work on an OA system they called Library.
The DOS group was interested in the work the DG group was doing. They had released a product named Library, which was a collection of DOS programs that we did not know how to sell separately. The collection included a shell, which let the user list and start DOS programs by means of a menu, an editor (P-Edit), a calculator, a file manager (for finding files), a simple database, a calendar, and a game called Beast. We sold all of the programs together for only $129. We did not expect to make a lot of money, rather only that our customers could get good use out of the programs. The DOS Library programmers hoped to one day turn their program into an office automation package for PC networks by adding electronic mail and enhancing the calendar to do meeting scheduling.
For a relatively small company, we were certainly doing a lot of different things. We intended to one day offer a word processor, a spreadsheet, a database, and an office automation package for all computers that made sense, while commanding a significant market share. We also asked Dave Moon, who had worked on MathPlan and Library, to look into desktop publishing. We were probably as aggressive as Microsoft in our development plans, but we did not feel greedy, since we limited ourselves to applications. We seemed determined to try every software applications category until we found another winner.
Unfortunately, we never seemed to find a way to sell these other products in large volume. With MathPlan, for example, we tried increasing the advertising and dropping the price of MathPlan to $195, but sales would not go up. We also tried a $99 MathPlan promotion for WordPerfect users, but after one month the promotion fizzled. These and other failures encouraged a lot of finger pointing, with marketing people claiming we needed more features and developers saying we needed better advertising. I thought the problem was never committing enough programming resources to these other products to make them successful. Alan and Bruce thought the products were good enough to sell well if we were more aggressive with our marketing. I had trouble believing we should spend a lot more money on products which already had been given a reasonable chance of succeeding.
Fall COMDEX was again better than ever. By now André was turning into a real showman. He had constructed a large game wheel for the show, one which required the approval of the Nevada Gaming Commission. During each show, our actors invited four or five members of the audience to come up on stage and spin the wheel to win a prize. Since our products were the prizes, it gave us a chance to draw some attention to our other programs. André's idea worked. Our COMDEX booth was packed with people pushing and shoving to get a seat.
Near the end of the year, I received my copy of InfoWorld containing their review of WordPerfect 4.2. By now I had a ritual for reading the review. I would take the magazine home and wait until everyone else in the house was asleep. Then, in a comfortable chair, I would very slowly read and enjoy every word of the review. There was not much enjoyment this year, however. Although the review was very positive, we did not receive a perfect score. InfoWorld had toughened up their standards and marked us down because of our printing deficiencies.
We finished the year with $52 million in sales and pre-tax profits of almost $15 million. This made us the fifth largest PC computer software company, behind Microsoft, Lotus, Novell, and Ashton-Tate, but more importantly, it made us the king of the word processing hill. The high ground held by IBM from 1964 to 1978, by Wang from 1978 to 1983, and by WordStar from 1983 to 1986, was now ours. WordPerfect for DOS, our cash cow, had more than doubled sales again.
At our Christmas party, Alan was excited about our growth and optimistic about the future. Looking a lot like a child about to give his mom a Mother's Day present, he offered to take the entire company to Hawaii if sales reached $100 million in 1987. For those of us there, there was no question as to whether we would make it. Doubling sales every year was easy.Go to Chapter 8