Chapter 4

Starting School

Back in my drapery days, I used to lose a lot of sleep thinking about all the deadbeats who owed the business money. One guy in particular made me very angry. He ordered and received a houseful of drapes and then refused to pay, claiming I had promised him the drapes as a favor. When we tried to get help from our collection agency, we found out the man was a disabled veteran and his pension from the military, which was his only source of income, by law could not be garnished. Occasionally I would see the man in the grocery store, and each time my frustration would return and I would lose a few more hours of sleep.

It took me seven or eight years to realize that my anger was not helping the situation. I needed to grow up and accept the fact that a small percentage of customers were going to try to find a way to avoid paying for their drapes. Losing a little money to deadbeats was just one of the costs of doing business. I had to learn to hand the deadbeats over to the collection agency and then forget about them.

When Don Owens started selling TIPS in Europe, we were similarly angry. How could we allow a major shareholder who knew our business inside and out to work for the competition? We wanted him out of the company immediately, but his asking price of $250,000 was more than we could pay.

Although we were young in terms of business experience, we were smart enough to go to a good law firm for advice. Our attorneys suggested we dilute Don’s holdings as a way of urging him to settle for a smaller amount. Key to a legal dilution was the agreement Bruce and Alan had with Eyring Research. The old agreement clearly showed that Bruce and Alan were the owners of the word processing software. Although Don may have felt otherwise, Bruce and Alan felt they had never given up ownership of their software to SSI. This claim put SSI in a very difficult position. If SSI only owned the right to sell and enhance the software, then Bruce and Alan were in a position to revoke this right. Luckily for SSI, Bruce and Alan were willing to sell their ownership rights to SSI in return for a big chunk of stock. Unfortunately for Don, the effect of the sale would be to decrease his percentage of ownership in the company.

Our attorneys expected Don to be very unhappy with the transaction, but they also expected him to eventually realize that he could not legally stop or reverse it. The attorneys hoped that once Don owned a significantly smaller share of the company he would sell out for a smaller amount. The possibility existed, however, that Don would bring a lawsuit against us all. The thought of being sued was a little frightening to me, because I did not yet realize that lawsuits, like deadbeats, were a normal part of doing business. I did not understand that a successful company was bound to have at least one or two lawsuits against it at any one time.

Our attorneys did their best to educate me in the ways of the real world. They explained that running a company was a lot like running a ranch in the Old West. Back then a rancher could not count on the judicial system to protect his interests, so he had to enlist his own hired guns to protect his property. If a rancher was weak or had no stomach for a fight, rustlers would steal his cattle and tougher ranchers would take his range. Likewise, if we had no stomach for a fight, our software would be stolen, our trademarks would be infringed, and our savings would be taken by customers claiming harm from our products. If we did not use our hired guns (our attorneys), our judicial system would not protect our interests. We were fast becoming a company with deep pockets, and we had to be willing to fight to prevent people from emptying them.

I was not ready to believe the attorneys at the time. I was still too young and too idealistic. Unfortunately, what they were telling me was true. Even if a successful company is fair and honest in every one of its business dealings, there will be a few lawsuits. The only way to avoid them is to stay unsuccessful and keep your pockets empty. As soon as you have something worth having, there will be someone else who will try to take it.

It was a hard decision to risk the lawsuit with Don, but Bruce and Alan were upset about his helping the competition. Once we decided to go ahead, the first step was to have an independent party appraise the value of the software as well as the value of SSI. Next Bruce and Alan made a formal offer to sell their software, which was appraised at about $500,000, to SSI in return for more stock. A special meeting of the Board of Directors was called in the summer of 1982 to consider the offer. Although Don was no longer on the Board, he was invited to the meeting so he would know exactly what was happening. Naturally he objected to the purchase, but the Board voted to proceed. Don had walked into the meeting owning one third of the company, but walked out owning only 5%.

Don sued SSI and those of us on the Board as a result of the sale. Throughout the fall and winter, depositions were taken, and settlement offers and counter offers went back and forth. I came to view my first deposition as a rite of passage. It was time to start growing up a little.

By January, we were very anxious to settle the lawsuit. I asked the attorneys to do whatever they could to negotiate a settlement right away, before we were required to announce our software sales. By now we were offering Don $100,000, and he was asking for $150,000. We increased our offer to $139,000, and for that amount Don agreed to sell his stock back to SSI and drop his suit. $139,000 was all the money we had in the bank, but it was certainly a bargain considering what WordPerfect Corporation would be worth in a few years.

After the settlement, Bruce and Alan each owned 50% of the SSI stock. Not too much later they let me buy .2% of the company. They talked about offering me more stock, perhaps as much as ten or twenty percent, but I was not interested. I was worried that one of my in-laws might someday claim I had taken advantage of Bruce. I wanted only enough stock to be a tie breaker, in case Alan and Bruce were to disagree or one of them were to die. SSI loaned me the money to buy my .2% of the shares, and I wrote and signed a buy-sell agreement, which gave the company the right to buy back my shares at the price I had purchased them. I did this because I did not want anyone to question my motives. Actually, I never even used my vote to break a tie. In the years we were on the Board together, I cannot remember an occasion when Bruce and Alan disagreed. Decisions were always made by a unanimous vote.

Sales for the first quarter of 1983 were slightly higher than the previous quarter, proving that our record fourth quarter of 1982 was not a fluke. Almost overnight our PC product had doubled the size of our business. While we did not have a large share of the PC word processing market, our expectations were so low that we thought we were wildly successful. Success is, after all, measured by expectations. Had we known more about playing the software game, we could have done far better that year.

We learned a lot from watching Lotus do it the right way. They spent about half a million dollars developing 1-2-3, which was approximately the same amount of money we spent in developing the DG and PC versions of WordPerfect. They spent about two million dollars on their 1-2-3 roll-out; their ads, brochures, packaging, distribution, and public relations were all very professionally done. We, however, spent only $100,000 on our roll-out and generally looked like amateurs at everything we did. 1-2-3 would become the most popular spreadsheet as soon as it was released. We would need five years to become the most popular word processor.

Watching another company, Softword Systems, Inc. (another SSI), also taught us some valuable lessons. They developed a PC word processor for an insurance company back east and than released that product to the public early in 1983. Their word processor, first called WordMate and later MultiMate, was designed to work like a Wang word processor. Wang apparently did not object to the copycat product, probably because it had a multi-billion dollar hardware business and was not interested in bothering with a multi-million dollar software business. This other SSI would do ten million dollars in sales in 1983, outselling us almost three to one with what I thought was an inferior and outdated product.

Most of the blame for our comparatively slow start belonged with my marketing department, but part of the problem was our software. Our biggest software shortcoming was our lack of printer support. While our Diablo support was very good, our Epson driver was a kludge (a “kludge” is a makeshift solution which is usually poorly designed and unreliable), and our support for all other printers was minimal. Early reviewers also complained about our paper template and unprofessional looking manual. We were able to correct the problems with the template and the manual early in the year, but the printer support would not be fixed until the release of version 3.0 the following year.

One other serious problem was our growing reputation for buggy software. Any complex software program has a number of bugs which evade the testing process. We had ours, and as quickly as we found them, we fixed them. Every couple of months we issued improved software with new release numbers. By the spring of 1983, we had already sent out versions 2.20, 2.21, and 2.23 (2.22 was not good enough to make it out the door). Unfortunately, shipping these new versions with new numbers was taken as evidence by the press and by our dealers that we were shipping bad software. Ironically, our reputation was being destroyed because we were efficient at fixing our bugs.

Our profits were penalized as well. Every time we changed a version number on the outside of the box, dealers wanted to exchange their old software for new. We did not like exchanging their stock, because the costs of remanufacturing the software and shipping it back and forth were steep. This seemed like a waste of money, since the bug fixes were minor and did not affect most users.

Our solution was not to stop releasing the fixes, but to stop changing the version numbers. We changed the date of the software on the diskettes inside the box, but we left the outside of the box the same, a practice known in the industry as slipstreaming. This was a controversial solution, but our bad reputation disappeared. We learned that perception was more important than reality. Our software was no better or worse than it had been before, but in the absence of the new version numbers, it was perceived as being much better.

Because we had such a small market share, we missed out on the biggest opportunity of 1983, the panic buying by hardware companies. IBM’s personal computer success led to reckless efforts by other hardware companies, like DEC, Hewlett-Packard, Tandy and Data General, to come out with their own PC’s. Since these companies were afraid of copying IBM’s PC exactly, and perhaps because they felt they could make better PC’s, their new computers were not completely compatible with the IBM PC. The incompatibilities meant that IBM PC software applications would not run on the non-IBM machines. Because software is essential for a computer to be successful, the hardware companies were eager, if not desperate, to pay the most popular software companies to port their products to the new machines.

Hardware company representatives hurried from one successful software company to another trying to hand over their money. The lobbies of the software companies were like crowded butcher shops, where customers had to take a number and wait their turn. VisiCorp, the publisher of VisiCalc, was so busy or so arrogant that it forced the hardware representatives to wait weeks for appointments. When Lotus announced it did not intend to port 1-2-3 to other machines, but would instead introduce and port a new product (which would be called Symphony), the rumor was the hardware companies together were paying a total up front fee of $32 million for rights to sell the unfinished and unproven product.

No one waited in our lobby, however. Data General came by to purchase a few copies of our IBM PC software for its machine, but it was the only hardware company to do so. We were at best a second tier company with unknown products. There was some money available to second tier companies from a number of the lesser known hardware companies. Many wanted to include software with the purchase of their computers, but the going rate for this bundled software was at most 5% of the retail price for a complete package. A $25 price seemed like an insult to us. We were neither hungry enough nor desperate enough to give our product away.

Although we did not receive a windfall from all the crazy porting and bundling activities of 1983, we were porting our software to many of the non-IBM MS/DOS machines anyway. We made versions for the Victor 9000, the DEC Rainbow, the Tandy 2000, the TI Professional, and a few others, while other companies were demanding large fees or declining to port their products to these machines. We did not make a lot of money from these versions, but we did make a lot of friends. Eventually we were given a lot of help to promote our product from various hardware companies, namely because they had so few products to promote.

None of these incompatible computers did very well, however. The best selling incompatible machine was the DEC Rainbow, which had the advantage of either working as a terminal for a DEC computer or running MS/DOS or CP/M. Even with this and other technical advantages and the strength of DEC behind it, the computer never had a chance. Customers did not want to wait around for their favorite software products to come to the incompatibles, and software companies quickly tired of the porting. Soon, out of desperation, the hardware companies began to copy IBM’s machine exactly, producing what were called clones. With the clones running the software made for the IBM PC, the huge fees for porting software vanished.

At first we sold most of our software directly to dealers rather than distributors. A dealer to us was any company which sold its products to end users–the people who actually used the products. A distributor was a company which sold products only to dealers. The reason we sold through dealers was that we could not find any distributors willing to carry our product, even though we offered them an attractive discount. The large distributors like Softsel and Micro D sent rejection letters telling us they already had plenty of word processors.

Our first dealer to order more than one or two copies at a time was 47th Street Photo. 47th Street was a perfect account–they ordered in quantities of 10 or 20 and paid COD. When I offered to let them take 30 days to pay, they declined the credit. Louis Schwartz, the man I dealt with, said he knew we were small and needed the money. At the time we knew nothing about 47th Street Photo, except that they were a great account from the New York City area. Later we learned they were disliked by almost all of our other dealers because of their very low prices. In spite of the pleas which later came from these dealers, I could never agree to cut 47th Street off. Their help in those early months had been far too valuable.

ComputerLand gave us our first big break when they picked us up in the spring of 1983. They were the largest chain of retail computer stores and the most important part of IBM’s dealer network. A decision by ComputerLand’s corporate headquarters to sell a product was an important seal of approval. In addition to the industry-wide recognition that came with the headquarters’ endorsement, the various ComputerLand stores were much more likely to sell a recommended product over an unrecommended one.

At the time ComputerLand would accept any product for review, as long as the software vendor sent in nine free copies of its product. Eight of these were sent to the eight ComputerLand stores which made up the evaluation committee. One copy stayed at CL headquarters. The stores on the committee were kept secret, so vendors could not lobby directly with committee members for approval. Periodically the committee met by telephone to discuss the merits of each product and to decide which products the chain would carry. Around the first of the year we submitted nine copies of WordPerfect and waited impatiently for the verdict.

Although I’m not sure we knew it at the time, the Baltimore ComputerLand store was on the committee and liked our product a lot. Will Fastie, who lived near the Baltimore store and worked for General Instrument, used our DG version. He also wrote a column devoted to the IBM PC for the magazine Creative Computing. Because he liked our DG version, he encouraged his friends at the Baltimore ComputerLand to try WordPerfect, and they became enthusiastic supporters.

The ComputerLand of Fox River in Osh Kosh, Wisconsin was the only other committee member to recommend our product, and their recommendation came almost by accident. I had traveled to Wisconsin at the request of Kimberly-Clark (the Kleenex company), which was in the middle of a word processing evaluation and was soliciting demonstrations from different vendors. I arrived in Wisconsin on the afternoon before my appointment and went calling on computer stores to fill up my afternoon. I happened to stop at the ComputerLand of Fox River, and having nothing better to do, I spent a few hours looking at competing word processing products while waiting for a chance to talk with the store owners. At the end of the day, I finally had an opportunity to tell Mr. and Mrs. John Teska about WordPerfect’s great features and mentioned my visit to Kimberly-Clark. When John told me KC was his client, I invited him to the demonstration.

As luck would have it, the ComputerLand evaluation committee was meeting the evening after my appointment with Kimberly-Clark, and because of KC’s positive reaction to WordPerfect, John went ahead and recommended it to the committee. We had only two committee stores out of eight that wanted to carry our product, but two was enough.

Getting picked up by the chain was only half of the battle, however. We still had to convince the rest of the stores to sell WordPerfect. We asked two college kids, Scott Worthington and John Lee, to take some time off from school to visit the ComputerLand stores across the country. Scott went east and John went west, and they spent the next nine months collecting Holiday Inn frequent sleeper points. They went from city to city, dropping off a demonstration copy at each store and showing as many salespeople as would watch how to use WordPerfect.

Our smartest marketing move that year was to offer ComputerLand employees the chance to buy one copy of WordPerfect for $10 if they would take a simple test. WordStar was so popular at the time that it was almost impossible to get salespeople to try any other word processor. Although most of them were WordStar fanatics through and through, the $10 offer was too good for many to pass up. The questions on the test were very easy. We asked things like, “Does WordPerfect or Word Star have the largest dictionary?” WordPerfect was the correct answer to every question, but the salesperson had to do some comparisons to verify the answers. Quite a few salespeople started using WordPerfect at home as a result of the low price we offered them.

By the time ComputerLand held its annual owners’ conference that summer, we were the chain’s second best selling word processor. Although we were second, we trailed far behind WordStar. During the conference at the Fairmont Hotel in San Francisco, we booked a suite and invited the owners to drop by. Only six people came by, and only two actually came into our suite. We did not understand that the hospitality suites were for having fun, not for doing business. Our shrimp and egg rolls could not compete with the mixed drinks served in other suites. The four or five of us there from SSI sat all alone eating our expensive shrimp in our expensive suite until we were sick.

The next day of the conference went much better. We had a small booth at the conference vendor fair, a small trade show where store owners and managers could wander around talking to vendors and watching demonstrations. As people walked by, we literally grabbed them and dragged them into our booth. We were too eager and naive to be shy or to believe we might have been bothering anybody.

Microsoft was about to release its word processor, which we saw for the first time at the CL show. We had been a little worried about the product, because we knew the company had the resources to do a good job. One look at Word, however, erased all our worries. The product was no more innovative than its name. Microsoft had designed Word to work like its spreadsheet, Multiplan. It was clumsy to use and came with no new or interesting features. You could tell the Microsoft programmers had not talked to many secretaries while developing their product. It would take Microsoft a little more work before they would be a significant factor in the word processing market.

In the summer of 1983, Terry Brown and I went to Syntopican, a trade show for word processing managers. At the time almost all word processing work was done by a company’s word processing department. The manager of the department, typically a woman, commanded a lot of power and influence inside her company, because she decided whose work would get finished first. When the word processing managers met together at Syntopican, they were treated like queens and kings by the word processing computer vendors. IBM, CPT, Wang, Dictaphone, DEC, Xerox, and NBI were just a few of the companies spending millions of dollars trying to win favor with the wp managers. As Terry and I toured the trade show, we were amazed to find so many companies spending so much money on a dying industry. The dedicated word processing computer was quickly giving way to the personal computer, but not many people at Syntopican seemed to see it coming. The word processing managers had trouble believing that executives would soon be typing their own letters and that within three or four years their jobs would disappear.

Although we were not doing as well as Lotus, our sales were growing steadily each quarter. After doing $427,000 for the first quarter, we reached $645,000 in the second and $843,000 in the third. In September we were actually mentioned as a best seller in Softalk, a trade magazine which has since gone out of business. They listed us as number 22 on their list of top-selling software products, which made us the fifth word processor behind WordStar, Word Plus, MultiMate, and PFS:Write. Although ComputerLand’s list was not published, we discovered we were number 12 on their hot list–the second best selling word processor in their stores.

As sales grew, we were hiring as many people as our budget would allow. We had no formal method of advertising our job openings, so everyone we hired was either a friend, a neighbor, a relative, a passerby, or a student in one of Alan’s classes. Doug Lloyd, for example, was a salesman from IBM’s Salt Lake office who dropped by to see if we needed a large computer. We did not buy a computer, but we did offer him a job. Doug became our chief salesperson for large accounts.

Bruce hired my brother André and told me after the fact. He knew Dan and I needed some help and that I was reluctant to hire more people for marketing. Although I was happy to see my brother get out of the drapery business, I was depressed about sharing my duties with someone else. I was going through the same kind of discouragement I would watch others experience over the years. As an organization adds people, some responsibilities must be taken away from the older employees and given to the newer ones. It is a process which is usually painful, but is necessary if everyone in a company is to have a meaningful job. André’s job was to keep track of our dealers and attempt to sell our products to distributors.

We increased our customer support department to four people. Until then, support calls were unofficially answered by everyone in the company. This method worked fairly well for the Data General support, because we did not get very many calls. It did not work well for the PC product, however.

If I remember correctly, our toll free support was somewhat of an accident, rather than a conscious marketing strategy. When we first released WordPerfect, we had a toll free number for orders, and a regular, non-toll free number for our other calls. Our customers outsmarted us, however, and most called us on the free number whenever they needed help. To keep the order lines clear for orders, we decided to add a toll free number for support. On two or three occasions during 1984 and 1985, we considered dropping the toll free support to save money, but each time we convinced ourselves that the service was worth the cost. After a few years, we had received so much good press for the service that it was impossible to curtail it. The service was very consistent with our “good guys in the white hats” image.

We were also hiring programmers. Although it may not seem logical at first, once a product goes out the door, a software developer always needs to add more programmers to the project. Releasing a product is a lot like having an older child move out of the house and a new baby move in. The product that ships, like the child who moves out, still has problems and needs some attention. The new version of the product, like a new baby, also demands a lot of attention. Producing a better version inevitably means a bigger and more complex program, which many times requires even more work than the first version. Fixing bugs is only a part of the maintenance required to support software once it ships. As new printers, monitors, graphics cards, keyboards and hard drives are introduced, more programming work is necessary to support these new peripherals (just about anything that plugs into the periphery or any other part of a computer is called a peripheral) and keep a product up to date.

In addition to meeting programming needs for our existing products, we still wanted more programmers for new products. When we first released WordPerfect 2.0, many large corporations told us they were not interested in our word processor unless we had a complete family of products. As it turned out, this excuse quickly disappeared with the success of 1-2-3, because practically every large company bought the Lotus spreadsheet, even though it was the only product the company offered. With this excuse eliminated and companies looking for the best product in each category, it would have been logical for us to abandon the family of products strategy, but we did not. Much like the mountain climber who climbed the mountain because it was there, we were intent on writing a new software product whenever we had the opportunity.

Dave Moon, another of Alan’s bright students, headed up a project to develop a spreadsheet, which we would first call MathPlan and later PlanPerfect. Lew Bastian, Bruce’s oldest brother, came from IBM to write a legal time and billing system called SSI*Legal and a version of the computer language Forth. Eventually he would expand the scope of his work and produce a database called SSI*Data, which would later be called DataPerfect. We were also working on a communications program called TranSSIt.

In spite of all these other projects, most of the programmers were working on WordPerfect 3.0. There were a number of new features, but the most important improvement was the new printer code. In the 2.2x versions (2.2x stands for 2.20, 2.21, and 2.23), we treated all printers except the Diablo and Epson printers as if they were dumb printers. A “dumb” printer is one which cannot do anything but print a standard set of characters (a character is a letter, a number, or a symbol on the keyboard) in a fixed pitch (fixed pitch characters all have the same width, while proportionally spaced characters can vary in width). A dumb printer cannot print proportionally spaced characters or different fonts, or do any of the more attractive things printers with smarts can do. If you had a smart printer that was not one of the two we supported, you could not use all your printer’s features with WordPerfect 2.20. Even with a Diablo printer, which had our best support, you could not print proportionally spaced text.

Our less than great printer support was a big drawback, because back then it was important to have a computer document look like it was professionally typewritten. For some reason, the early computer users wanted to hide the fact that a computer was doing much of the work. It would take a couple of years before people would consider it a status symbol to show off the fact that they were using a computer. The professional typewritten look was called “letter quality” in the industry, and one of our goals for WordPerfect 3.0 was to print a letter that looked as good as one typed on an IBM Executive typewriter.

Our biggest problem in supporting the “smarts” in many printers was that all of our printer instructions were hard-coded into WordPerfect. Having hard-coded printer instructions meant that the instructions for the printer were written inside of the WordPerfect program. This meant that whenever we wanted to add a new printer, we had to add more code to WordPerfect. Since there were hundreds of printers on the market, adding code for each of them would have made the program far too big, not to mention the nightmare it would have been to release a new version of WordPerfect every time another printer was supported.

The solution was to put the printer instructions in tables outside of the program. A table is nothing more than a list of values. A box score in the sports section of the newspaper is an example of a table. In a box score for a basketball game, for example, each player’s name is listed with values for minutes played, points scored, rebounds, turnovers, etc. In a printer table, the printer features are listed along with the instructions that are sent to the printer for each feature. A printer table would hold the instructions for things like turning bold on, turning bold off, turning underlining on, etc. For version 3.0 of WordPerfect, the printer code was entirely rewritten to use look up tables for all printer instructions.

The result of the change was that our biggest weakness in 2.20 became our biggest strength in 3.0. We bought and borrowed all kinds of printers, and by the time we released 3.0 we supported more than fifty of them. Within a year after the release, we would support about two hundred printers, more than any other word processor. We could also add support for new printers faster than our competition. We even went so far as to include a special program with WordPerfect which let users make their own tables or drivers for printers our software did not support. The program could also be used to make changes to our printer definitions should a user want to improve on our work.

The only disadvantage to the table driven method was that printing was a little slower. Unfortunately, we underestimated how vocal and angry a few of our customers would get because of the printing slowdown. We quickly learned that no matter how many improvements we might make to a new version, if we took something away, we would have many unhappy customers. Some customers went so far as to demand a partial refund for the printing delays, however, I don’t think we gave them any money back.

We introduced the new version at Fall COMDEX in a small 10 foot by 10 foot booth located at the far end of the convention hall facing the back wall. COMDEX (which is short for COMputer Dealer EXposition) was the biggest computer trade show in the United States. It was expensive, noisy, crowded, tiring, and essential to attend for almost everyone in the computer industry. The fall version of COMDEX was always held in Las Vegas, which was fitting because a lot of companies spent their last dollars to come to COMDEX, gambling that they could find enough interest in their products to stay in business a little longer. Ours was the smallest booth space you could rent. The largest booth spaces were sometimes as large as 10,000 square feet or more.

In addition to paying for booth space, companies had to pay for the actual booths they used to fill their spaces. A booth could easily cost $100,000, and a few companies spent more than $1,000,000 on their booths. Once a booth was paid for, a company had to pay shipping charges to send it to and from the show. The company then had to pay set-up charges, charges to run power to the booth and put computers in the booth, and the costs of sending employees, models, and actors to the show to work the booth. (We did not use models, because we all looked so good.)

Our first booth was not very extravagant. Dan was in charge, and he managed to cram four computers and six workers into our small area. Those of us who worked the show dragged as many people as we could into the booth for a demonstration. Though our booth was minuscule compared to Micropro’s, the WordStar company, we were proud of ourselves nonetheless. Just making it there and getting the booth set up seemed like a great accomplishment.

One man who came by our booth was from one of the big three auto makers. He offered me $250,000 for a site license to run WordPerfect on all their Victor 9000 machines. That was a lot of money, but I was very idealistic and turned him down. Even though our company had little money in savings, I did not want to sell them a license to run WP on all their current and future Victor machines for a one time fee. I wanted us to make a little money on each new machine they brought into their company. Later the auto company dropped the Victor 9000 and replaced all their machines with IBM PC’s. We could have taken their money for the Victor software and then sold them new copies for the IBM PC.

We were all a little in awe of companies like Micropro. They had a very large, beautiful booth and what seemed like thousands of people jamming into it to look at their products. At the time, I do not think any of us dreamed we would ever pass them up. We were a little frustrated because we knew we had a better product, but we did not know how to use this advantage to overtake WordStar. Micropro was outselling us by about twenty to one.

With only $60,000 of savings in the bank as a war chest, we did not have the money to fight Micropro aggressively. Unless we changed our minds about going into debt or selling off part of the company, we had to be content to continue our course of growing as we had the money. This was probably a good thing, because we were not yet smart enough to spend a lot of money wisely. We were no longer like innocent little children, but we were only beginning to learn about things like distribution strategies, product branding, product life cycles, and test marketing. Our advertising was still homemade, and our public relations was not very good. We still did not know how to makes friends with the writers from the press.

In spite our frustration and our inexperience, we were happy to finish 1983 with $3.5 million in sales, up from about $1 million in 1982. Because about half of the sales came in the last quarter, after the release of version 3.0, we were confident we could at least double sales again in 1984. We were barely coping with the problems that came with the sales increases, but we were surviving nonetheless. We were starting to get favorable reports in the press. The November issue of Softalk announced that their dealer survey showed us to be the best selling word processor for the month of October. While this was far from the truth, it did show that there were at least a few dealers who liked our product. More importantly, our customers loved us and were saying good things about us to their friends.

Go to Chapter 5