by W. E. Pete Peterson
Copyright 1993, 1998 W. E. Peterson
Children Masquerading as Adults
After Don left, we were like little orphans without any supervision. We knew almost nothing about management or marketing, but we did not care. We were free to do whatever we wanted to do, whenever we wanted to do it. The fighting was over, and as long as we did not run out of money, we knew no one would bother us. Like little children playing dress up, we did our best to look the part and to imitate the way grown-ups played the business game.
If we had wanted to practice medicine, we would have needed a degree and some experience. If we had wanted to practice law, we would have had to pass a difficult test. Running a company, however, did not require proficiency tests, a college degree, or any relevant experience. While running a business profitably is arguably as difficult as any other profession, almost anyone is allowed to give it a try, whether they are qualified or not. It is like an inalienable right, available even to the foolish, the young, and the senile.
We never considered hiring professional managers to help us. In fact, the professionals were the enemy. They represented "business as usual," which meant working for an overbearing boss, fighting political battles, and living with contention. When one banker suggested we look for someone with experience to run the company, we found another banker. Having gained our freedom, we did not want to lose it. For my part, I enjoyed not having anyone looking over my shoulder to make sure I wrote each letter correctly or said exactly the right thing to a customer.
With Don gone, Bruce and Alan asked me to take his place. Alan was to look after development, Bruce was to concentrate on our international business, and I was to do the sales and marketing for the US and Canada, while keeping an eye on finances. None of us knew if I could handle the job, so I started on a temporary basis. If I felt comfortable doing the job, I could keep it. If I had trouble, I was supposed to help find someone who could do the job, and I would then work for that person.
I asked for Dan Lunt's help because of his technical background in computer science and engineering, which was something I lacked. Dan's personality seemed to me to be better suited to marketing than programming anyway. He had trouble sitting in front of a computer for eight hours a day. After a couple of hours of programming, he would wander around interrupting someone or looking for something else to do. I was impressed with Dan's sales experience--he sold real estate part-time for one year while he was in college. It was not until years later I found out he had never even sold a house.
Dan and I did not change much at first, except to establish standard discounts for all resellers. Having to run to the file cabinet to read a contract and find the right discount before quoting a price was not practical. As quickly as we could, we tried to give all our resellers a 40% discount, but it was not something we could do overnight. In some cases it took us a few years to renegotiate the old contracts.
If we had had more experience, we probably could have found a lot of things which needed improvement, but we were doing well just to learn what it was we were supposed to be doing. It took us a year to figure out how to run the business. Although our sales were not increasing significantly, our profits looked a lot better once the large bonuses had stopped. We were never in any financial danger, with expenses averaging about $50,000 a month and sales averaging about $65,000.
In spite of our flat sales, Bruce and Alan were not afraid to hire a few more programmers. Alan continued to teach at BYU, so he was in a perfect position to recruit BYU's best computer science students. He was a tough grader, and anyone getting an A in one of his classes was a candidate for a job offer. When Alan found students with exceptional talent, he and Bruce would offer them part-time jobs and put them to work on one of the programming projects. By the time they graduated, their work had helped the company grow to the point where we could afford to offer them good salaries. Eventually we used Alan's old grade sheets to search for good programmers who had already graduated.
Alan Brown was one of the BYU students we hired to work part-time. He was brilliant, especially at math. He could do amazing things like perform hexadecimal computations in his head. I think he had a photographic memory. He intended to go to the University of California at Berkeley after graduation to get an advanced degree and do research, but we hoped to one day offer him enough money to change his mind.
He started work in the storage room, using a door stacked on some computer boxes for a desk. The boxes which supported the door belonged to SSI's new IBM Personal Computer. The IBM PC was announced late in 1981, and we bought our first one in February of 1982. Alan Brown's first job was to see if it was possible to convert SSI*WP to run on the small IBM PC. Because there were no programming tools for him to work with at first, he did his initial programming in machine code. Programming in machine code is similar to eating rice with tweezers: both take a lot of time and patience. By March, Brown was convinced we could make SSI*WP work on the new computer, so we decided to go ahead with a PC project.
It was not a difficult decision. I remember when a few of us were eating lunch at the local Sizzler, an eavesdropper interrupted to say, "Are you guys writing a word processor for the PC? So are we." So was everyone else. At least 200 companies would introduce a word processor for the IBM PC within the next two years. Most of these products would never make money, because the market would get very crowded very quickly, but our risk of failure was not as great as most of these other companies. Already many of our Data General customers were planning to buy IBM PC's and hoping to use SSI*WP on them.
There were other small computers we could have supported first, like the Apple II or the Victor 9000, but the IBM Personal Computer seemed to have the greatest chance for success. IBM, especially in 1981, was well-respected. They seemed big enough and powerful enough to do almost anything they wanted.
We would have liked to have had the first word processing product for the IBM PC (as someone said, "If you are going to march in a parade, you want to try to go in front of the horses."), but we had a number of hurdles to get over before we could put a lot of programmers on the project. Our first problem was deciding whether to write the new version in assembly, which was the language used on the DG version, or in a high level language like C, Pascal, or Basic. Writing software in assembly language is somewhat analogous to baking from scratch rather than using a mix. The programmer has more control over the ingredients of his program in assembly, but the programming requires more work and more lines of code than it would if the programmer were to use a high level programming language. Most of the high level languages were not well suited for writing a word processor, although we did consider using C. We had trouble, however, finding a reliable C compiler for the PC. This, and the fact that programming in a high level language usually produced a program which was bigger and slower than one in assembly, convinced us to use assembly.
Once we decided to write the new version in assembly language, we were forced to wait until May for an assembler to come on the market. Once the assembler was available, we still needed a good program editor. EDLIN, the editor that came with DOS, was not worth using, so Alan Brown began getting P-Edit ready to work on the PC.
In the meantime, the other programmers were not wasting their time. They were working to improve the DG version, and by now we had released version 2.0 of SSI*WP. They added a lot of new features, including spelling checking, footnotes, and some basic arithmetic which worked on columns and rows of numbers.
We had no systematic way of deciding what features went into a particular version of the product, but many of the improvements came from the suggestions of our customers, who were constantly calling with requests for more features. If something was easy to do and made sense, it usually made it into the next version. If it was very difficult or not often requested, we would usually put the feature on a schedule for a later release. Sometimes if a customer absolutely had to have a feature right away, and if they were willing to pay a few thousand dollars to fund the work, we would take their money and move the feature to the top of the list.
While our software was getting better, our relationship with Don was getting worse. Even though we were paying him as a consultant, we did not see him at all for the first few months of the year. After a vacation in Hawaii, he went to England as a consultant for his old employer, Itel. At Don's suggestion, Itel was starting a distribution company in Europe to sell software products for Data General computers. Don's plan to become a software distributor was ahead of its time. Eventually software distributors would become the primary means of getting software out to dealers and customers, but to do it in Europe, in 1982, with software only for DG machines, gave Don little chance of success. There were not enough products or customers to support the venture.
Even though Don still owned one third of SSI, he chose to have his new company distribute our archenemy, TIPS, to its dealers and customers. As I remember it, he asked us for exclusive rights to sell SSI*WP in Europe, but when we dragged our feet, he decided to go with the other product. It was very strange to have a member of our Board of Directors promoting a competitor's word processor. We were not too happy to see press releases and TIPS sales literature containing Don's statements of endorsement for their product.
When Don came back to Utah for a short visit in April, we called a special meeting of the shareholders to remove him from the Board of Directors. Surprising us again, Don voted to remove himself from the Board and nominated me to take his place. At the end of the meeting, he wished us well and encouraged us to make him rich. He told us some day we would pay him $250,000 for his stock. We offered him $30,000 right then, but he laughed. Don went back to England for a few more months until his distribution venture was abandoned.
I liked my new role as a vice president and member of the Board of Directors of a young software company in a new industry. After years of wondering what I would do with my life, I had found my place. My job was so interesting and challenging that it consumed my every waking moment. I was past the point of being a workaholic. I would go to the office on Saturdays and stay much longer than was necessary. I would take piles of trade publications and literature home to read each night. I dreaded Sundays, three day weekends, and vacations when I could not go into the office. I was addicted to the software business and could not get enough.
In the spring, Dan and I decided it was time to advertise. We went to the local bookstore and bought a couple of books on advertising. After reading a few chapters, we wrote some copy and placed our first ad in a DG trade publication. When a few leads came back, we thought advertising was a snap. That was the first and last time we thought advertising was easy.
Our approach to advertising became our normal approach to doing business. Whenever we felt we needed to do something new, like pricing a product or working out a direct mail campaign (direct mail is the polite name for junk mail), we would buy a book, read a few chapters, and then do the best we could. We never thought to use an advertising agency, a public relations firm, or a consultant (Alan liked to repeat the joke that a consultant was a guy who borrowed your watch so you could pay him to tell you the time). Occasionally we were willing to get help with legal and accounting matters, but never when it came to matters of marketing, sales, or management.
If we had trouble finding a book on a subject, we tried our best to figure things out on our own. Figuring out how to sell our software so that we would retain the legal rights to the product was one such problem. There were no easy answers and a lot of conflicting legal advice. Some lawyers suggested using trade secret law for protection. Others thought copyright law applied. Some suggested applying for patents. The situation reminded me of the times I was unprepared for an essay test in school. If I did not know the answer, I would write anything and everything down in the hopes of getting part of the question right. Attorneys were doing the same thing. They did not know the answer, so they suggested a little of everything.
In cases like these, where easy answers were unavailable, we generally copied what IBM was doing. We decided to "license" our software rather than "sell" it and try for copyright and trade secret protection. We granted the customer limited rights to use the software, but not to own it. No one knew if this arrangement would stand up in court, but it was the best decision we could make at the time.
Late in the spring we moved from our cramped space behind the donut shop into what were for us some very impressive offices. Novell, which started out in Orem as a computer manufacturer, was losing money and had decided to move to a less expensive location. Following their departure, we moved into Novell's vacated offices, enjoying their new wallpaper and carpeting.
I went to my first DG reseller conference about this time. This was my first time at a resort hotel (the Camelback Inn in Phoenix), and I enjoyed the speakers, the conversations in the hallways, the club soda at cocktail parties, and the beautiful surroundings. There was so much interest in SSI*WP that people followed me everywhere, even to the men's room, to ask about it. We were having no trouble winning sales away from TIPS and AZText.
While our customers and dealers seemed to love our product, we did not get a lot of respect from the experts. Occasionally a consultant would stop by the office to take a look at the product. Usually it was because a hardware manufacturer was looking for a word processor to purchase and call its own. Very few of these consultants liked our product. They would ask stupid questions like, "You don't look anything like Wang, so how do you expect to sell your product?" It drove me crazy that the consultants rarely knew how to type and never took the time to learn how to use our product. After only a quick look, they filed us away in their "losers" folder. Their lack of respect gave us motivation to work harder and prove them wrong.
Although we did not understand it at the time, we were one of many small companies disrupting the established practices of the computer industry. As one consultant put it, we were a renega de. Our product was strong enough that some DG resellers (not DG, but their resellers) were going after word processing bids with Data General machines. Although on a very small scale, we were starting to take business away from established companies like Wang, NBI, and IBM.
The profitable, well-established computer manufacturers like IBM, Digital Equipment Corporation (DEC), Wang, Hewlett-Packard (HP), and Data General were not used to outsiders coming in and talking to their customers. In fact, the standard purchasing process followed by customers and vendors made it almost impossible for new or small vendors to sell their products.
Customers did not make a decision to buy a computer one day and simply pick it up the next. First they sent out an RFP (request for proposal) or an RFQ (request for quotation) to invite interested vendors to submit bids for a new computer system. The request contained a list of requirements for the system. The manufacturers would read the requirements carefully, and if they thought they could qualify, would submit proposals and price quotations. Usually one of the computer manufacturers was a favorite going into the bidding process and would try its best to help the customer write the requirements in a way most favorable to itself. This was one of the reasons IBM salespeople spent so much time getting to know their customers. The vendor that helped write the RFP could usually tip the odds of winning the bid in its own favor.
The proposals which the customer received rarely contained a mixture of products from more than one vendor. IBM sold only IBM products, DEC sold only DEC products, etc. Since the customer actually expected the system to work, vendors did not want the bother of mixing and matching a group of incompatible products. The proposals included hardware, software, systems engineering (any special programming services), training (teaching the customer's employees to use the system), support (answering questions from the customer's employees), and maintenance (regular servicing to keep it running smoothly and repairs when the system quit). This was one-stop shopping.
Once all the proposals were in, the customer and the interested vendors went through a mating-like ritual, filled with meetings, presentations, demonstrations, promises, and a few rounds of golf, until finally one vendor was awarded the bid. Typically price was not the primary consideration. Though an IBM computer could be as much as twice the price of a comparable DEC computer, and a DEC computer could easily be twice the price of a comparable DG computer, IBM outsold DEC, and DEC outsold DG. Reliability was more important than price, and the winning vendor was usually the one, of those that seemed to have the ability to fulfill the requirements, that had the best reputation.
Once the bid was awarded, the customer and the computer vendor were bound to each other forever, or at least until the next RFP cycle. A detailed contract governed every facet of their relationship. If the customer wanted to add something new to the system at some point in the future, the contract required the customer to get permission from the vendor. Typically any warranty or representation of reliability was voided with any unauthorized addition. Even if the vendor gave its permission for something new from a third party, such as SSI*WP, the vendor usually blamed any problems in the computer system on the added software or hardware. This is one reason why we had to be so committed to customer support from the beginning--we always had to prove we were not the culprit if a problem came up.
The customer paid a high price for the reliability and convenience of one-stop shopping. Typically a vendor would not make all the components in a system, so it would take other companies' products and private label them. (Private labeling was when a vendor would put its own label on the outside of another manufacturer's product and sell it as its own.) To buy a NEC printer from NEC might cost $2,500, but to buy the identical printer from Data General with a DG nameplate might cost $5,000.
Our existence and success were a threat to these cozy relationships and high prices. Instead of pushing customers to buy a new computer, we tried to convince them to add our software to their old one. Luckily, Bruce and Alan had chosen to create SSI*WP on a Data General machine. DG struggled to win bids against DEC and IBM, so they were more willing to allow a third party to get involved. We were both an ally and an enemy to DG, which accounted for the love/hate relationship we had with them through the years. We had to tread lightly and be very careful about what we did and said around their customers to receive their cooperation.
As the largest and richest company, IBM benefited the most from the old way of doing business. Ironically, IBM was probably most responsible for bringing down the old establishment. When they went outside the company for the operating system for their new PC, they opened the door for other vendors to sell to their customers.
The operating system (OS) is the first piece of software which runs on a computer. This software does just what its name describes--it operates the system. The OS software controls all the pieces of the computer--the processor, the screen, the keyboard, the disk drives, etc. If you turned a computer on without an operating system, nothing would happen. With an operating system, the computer comes to life and is then ready to run other software. Other software products running on top of the operating system were usually called applications.
In the single source, one-stop shopping old world, almost every computer manufacturer wrote its own operating system. This guaranteed that a computer from one manufacturer would never be compatible with one from another manufacturer. It was these incompatibilities which helped perpetuate the single vendor solution.
Probably because they were in a hurry to get the PC out the door, IBM decided to look outside the company for the PC OS. CP/M from Digital Research, the most popular OS for small computers, was the logical choice, but Digital Research somehow dropped the ball when IBM came calling. Bill Gates and Microsoft picked up the ball by buying rights to an OS very similar to CP/M for $50,000 and then adapting it for the IBM PC. When IBM decided to buy the PC operating system from Microsoft, and when they allowed Microsoft to sell the same operating system to other companies, they opened the door for other companies to copy their machine. Eventually the same computer would be available from a number of vendors, giving customers the chance to shop around and get more competitive bids. The PC would turn the computer establishment upside down. Microsoft, which was then a little software company near the bottom, would one day end up on top.
By August of 1982, Alan Brown had P-Edit working, and every programmer in the company, including Bruce and Alan, started work on the PC version. They targeted November 15 as the earliest possible release date and worked almost around the clock to finish the program quickly. The PC version was almost identical to the DG version, except that the PC version did not need the code to handle multiple users. We also changed some of the more violent terms used in the DG version, such as "abort" and "kill," to words such as "stop" and "cancel."
We were in a hurry to get the program ready. Throughout the summer we watched as product after product was introduced. Easywriter, from Sorcim (their name came from the word micros spelled backwards), beat us to market with IBM's help and blessing. IBM worked with Sorcim to get Easywriter ready early, and both companies sold the product under its own name. In this case, the advantage did not matter, because both versions were bad and bombed immediately. Volkswriter, a low-priced word processor, came out and captured the low end of the market. Because the program had a very limited number of features, the programmer who wrote it could use Pascal, which was available before the assembler, without worrying about the program's size.
WordStar came out around the middle of the year and immediately captured a large share of the market. It had been the most popular product for CP/M machines, and on release, became the most popular product on the PC. WordStar made it out early mostly by luck. The programmer who originally wrote the program for CP/M had already left Micropro, the company which sold WordStar, by the time the PC came along. The rumor was that the Micropro programmers assigned to translate the product had a difficult time understanding the CP/M assembly code, so they did a quick and very literal translation to PC assembly. It was similar to translating a book word by word from one language into another without trying to understand the meaning of the sentences. Though it was hard to believe, the program worked, although with a few bugs. By the time we came out with our PC product, WordStar owned at least 75% of the market, and the twenty or so other word processors already on the market shared the rest.
One big problem we had getting ready for the PC release was finding a new name for the product. SSI*WP was not very catchy. I liked the name WordPerfect, but I could not get anyone to support me. I thought of the name while pulling into a parking space in front of our offices, when I had one of those "Ah Ha!" type experiences. I liked the name, because it reminded me of the idiom "letter perfect" and described something which was word for word correct. I rushed into the office, sure that everyone would go crazy and love the name, but no one did.
Months passed and we still could not agree on a name, so we decided to hold a contest among the employees to name the product. Whoever came up with the winning name would win $100. From a long list of nominated names, we all voted for our favorite. Word Plus and ProWrite received the most votes, and WordPerfect came in somewhere near the bottom. In spite of a very poor showing, I put WordPerfect on the list of names we gave to the attorney for a trademark search, just in case we could not use one of the winning names. It turned out there was already a word processor with the name Word Plus, and there was a printer with the name Prowriter. Because we had taken so long to decide on a name, we had too little time to start over. Under these circumstances, WordPerfect became the unpopular winner. The name was so unpopular, in fact, that no one paid me the $100 prize money.
Despite its initial unpopularity, the name proved to be a good one. It was so positive sounding that it made any criticism sound untrue. It was like naming a soap, "Makes You Look Younger," so the competition would have to say something like "Use our soap instead of the Makes You Look Younger soap." "WordPerfect" also sounded like a very good product.
That fall I called on ComputerLand's headquarters to see if they were interested in carrying our product. Their chain of franchised stores was the largest and most important part of IBM's distribution channel at the time. We were willing to offer ComputerLand an exclusive right to sell the PC product if they would promise to pick us up, but, luckily, the buyer I talked with was not interested in the offer. I did ask him, however, how many copies a moderately successful product might sell in their stores. The buyer said they had very few moderately successful products. They generally had only hits and misses. A miss sold almost no copies, but a hit could sell two to three thousand copies a month. As he spoke, I quickly did the math in my head. Three thousand copies was about $500,000 a month. I could not imagine getting that much money from just one customer.
Dan took charge of advertising and planned to have advertisements in the PC magazines on October 15. This was a little early, but back then it was a common practice to advertise a product before it was ready to release. Some companies went so far as to advertise software even before they started the programming work, making sure they had orders before they made their development investment. At least our software was close to completion when we ran the advertisements.
At Bruce's insistence, we committed all our savings to the ad campaign announcing WordPerfect 2.20. I remember how dramatic and daring it seemed when he said something like, "We'll spend $100,000 on the roll-out if we have to." The ad we came up with looked like the cover of a science fiction magazine. It had a lightening bolt coming out of a man's head, going through his hands, continuing through a computer keyboard, and writing words on a piece of paper. The ad was beautiful to look at, but it could have used some professional help. We were amateurs, and it showed.
We officially announced the product with a mailing to dealers and the press on October 15. Terry Brown, an old friend who was hired in 1981 to help with documentation, prepared the mailing. He also worked to get the manual ready to go to the press on November 1.
When WordPerfect actually started working on November 18, there was little time to celebrate. There were a lot of bugs to get out, and the speller and the sort packages were still not finished. At the last minute Dan decided we had to have an Epson printer driver (the Epson printer was one IBM sold for the PC with its label on it), so he went back to programming for a few days to write the driver.
WordPerfect 2.20 for the IBM PC shipped the day after Thanksgiving. It was a good thing WordStar was number one, because we could not have handled a large sales volume. We did not know how to assemble packages or how best to ship them. We did not take the time or spend the money to have a typeset manual or plastic template made. There were many things we had to learn before we would be ready to take over the market.
Just moments before WordPerfect was supposed to ship, we found some problems with the installation instructions. At the last minute we typed up new pages with the word processor and then photocopied, cut, punched, and put them in the binders. If somebody wonders whether they have a first edition copy of WordPerfect for the PC, they can know for sure if the first five or six pages look like bad photocopies.
Our entire manufacturing facility fit in a room about 25 feet square. We purchased a couple of disk duplicators, but the machines could not keep up with our orders. For the first few weeks we caught up by paying our older kids a dime a disk to duplicate the disks at home. We also hired Alan's daughters to put the replacement pages in the binders until the new manuals were ready. We were still a small company with only 18 employees. Alan was still coming in after hours with his older kids to clean the offices for extra money.
Our orders department consisted of two people taking orders at desks in the hallway at the center of our building. We put some shelving up on the wall across from where they sat to hold the ready-to-ship packages. With the order desk in the middle of everything, we all knew what was happening. We all liked to watch as the manufacturing people, led by Dan Lunt's brother Ron, assembled the packages and placed them on the shelves. We knew how well sales were going by the height of the stack of invoices on the order desks and the number of boxes waiting to be picked up. If manufacturing fell behind, we all pitched in to help them catch up. It was an exciting time.
I remember the first WordPerfect sale. I had come into the office for just a few minutes on the day after Thanksgiving, and as I was walking out the door, the phone rang. On impulse I picked it up. A man at the other end of the line said he had seen our ad and asked to buy a copy. As I wrote down his credit card number and his address, I was thinking how much easier it was to sell PC software than DG software. The experience was almost intoxicating.
In spite of the crowded word processing market, our weird ad, and our cheap-looking manual, our phone seemed to ring off the hook. Sales for the last quarter of 1982 jumped from a quarterly average of $200,000 to $450,000. We finished the year with sales of a little over $1,000,000. That was not much more than the $800,000 we had made the year before, but we had done it on our own without Don's help. By trial and error, we had learned how to sell our DG products and how to introduce a PC product. We were only a small group of friends, relatives, and neighbors with little experience, but we were profitably running a million dollar business while having a good time.
Our peaceful existence as unsupervised orphans was about to come to an end, however. The stable, slow-growing DG business was about to be overrun by a very unstable and rapidly growing personal computer business. Soon we would not have the luxury of solving problems and making decisions at a leisurely pace in an isolated environment high in the mountains. We would have to grow up quickly to keep up.