Growing Pains

We told the world in our new, professional ad campaign that WordPerfect 4.0 was the most perfect WordPerfect ever, and the product was worthy of the hype. 4.0 was fast, reliable, and significantly better than our competition. Now all we had to do was convince the computer buyers of the world we were right.

This was not an easy job, because word processing users had a religious-like zeal for their favorite products. For most people, switching from one product to another was almost unthinkable. WordStar was especially hard to learn and master, and fans of the product defended it with an irrational fervor. Their loyalty was similar to that of a mother who has given birth to a very ugly baby. It was almost impossible to get expert WordStar users to admit their product had any flaws.

The diehard WordStar users pushed their word processor on unsuspecting friends and co-workers, and this very vocal group represented our biggest obstacle to convincing new customers to buy our product. Luckily, Micropro made our job easy by attempting to replace WordStar with WordStar 2000. Micropro tried to fix all the problems in the original WordStar, and rather than making their loyalists happy, they alienated them. Their strongest supporters had bonded to the program’s quirks and problems and had come to believe that the difficulties were desirable. Worst of all, WordStar 2000 was very slow compared to the old WordStar. Even a mother could not have loved the new product.

As far as I can remember, only one reviewer, Ronni Marshak, liked WordStar 2000. In spite of her comment that WordStar 2000 had “soundly thrashed” WordPerfect 4.0, Micropro’s sales figures headed downward. After reaching a high of $67 million in 1984, the company’s sales declined to $43 million in 1985. Eventually Micropro would realize its mistake and try to push the classic WordStar, but its sales figures would never improve.

Though the InfoWorld review of WordPerfect 3.0 in 1985 had been somewhat schizophrenic, extending compliments while giving us a less than positive overall rating, the 4.0 review had only one personality. It was positive in the extreme. The final sentence of the review read, “We believe WordPerfect 4.0 represents a new standard of excellence for microcomputer word processors.” They gave us a perfect rating, with a grade of excellent in every category.

The InfoWorld rating was worth millions of dollars to us. The perfect score became the theme for all our advertising. We reprinted the review and sent copies everywhere. When we called on customers, we made sure they saw copies of it. When we visited our dealers, we made sure they had plenty of copies to hand out. We had been given the best possible endorsement by the most influential computer publication, and we did our best to make the most of it.

Our January sales topped one million dollars–our first million dollar month. Bruce and Alan took the whole company out to lunch to celebrate. Over the course of the year, sales would only get better. Bruce and Alan would pay for lunch again in August, when we would have our first two million dollar month, and again in October, when we would have our first three million dollar month. Although we started the year with our product in only 15% of the computer retail stores, we would end the year with WordPerfect in practically every store. Everywhere we went dealers wanted to carry our product.

About one third of our sales were in the Washington, DC area. Although I am inclined to criticize the decisionmaking abilities of the US government, to their credit, they discovered WordPerfect when it was practically unknown. Forced by federal purchasing guidelines to evaluate most products, they gave us the highest rating in most of their evaluations. Although individuals would buy more WordStar in 1985, and large businesses would buy more Multimate, government offices and agencies would buy more WordPerfect.

Keeping up with the growth was a big problem. Our informal management style worked well when eighty people were focused on getting 4.0 out the door, but it was breaking down under the weight of all the new people we had to hire. Crisis growth worked fine for office space and computers, and management by luncheons helped improve the communications between departments, but our biggest problems had to do with how they operated. Each department had the freedom to design its own organization and set its own policies. With everyone solving their problems by trial and error, we had all kinds of inefficiencies creeping into the organization. The orders department, for example, always seemed to keep a copier running full time. When I looked into the reasons, it turned out they were making a lot of extra copies of invoices so they could file them by invoice number, by customer, and by product. When asked why, they were not sure, except that it had always been done that way. To make a point, I threw one full set in the trash and told them to call me if they ever missed the extra copies.

As we added people to the company, we generally tried to keep our organization flat, but the number of managers we used varied considerably from department to department. Our developers, for example, had a very flat organization–all of them answered directly to Alan Ashton. At the other extreme, the customer support department had seventeen employees, of which more than half were managers or assistant managers.

Our hiring methods were also inconsistent. We had no interviewing standards, nor any policies for hiring or not hiring friends or relatives, and we rarely checked references. Almost everyone who was hiring made promises about raises and future opportunities without writing anything down, and we all tended to describe job opportunities as more glamorous than they were. The unkept promises made for a lot of unhappy people.

To make matters worse, new employees were given very little training, very little supervision, and, in general, did not understand their jobs or how the company worked. Although we told people our basic management philosophy was to “teach people correct principles and let them govern themselves,” we did very little teaching. We were much better at turning people loose than we were at helping them learn what it was we needed them to do.

Although I was not in charge of the entire organization, I spent a lot of time trying to figure out how to fix these problems. I knew we needed better training and more supervision, but I was not exactly sure how to achieve these results. The problem of who should do the supervising was one that particularly bothered me. I did not see how to make everyone happy, if some were asked to supervise and others were not.

Whenever I got stuck on a problem, I called my dad in California. He had been a vice president of marketing for one of the divisions of Harris Corporation and had a lot more experience than I did at running a large organization. When I explained the problems we were having, he told me a story he had heard years before at a management seminar. I would think about this story many times over the next few years.

The tale concerned a good king, his knights, and his prime minister. The knights of the kingdom were brave and loyal. They loved to jump on white chargers and rush off to slay dragons or save damsels. The knights performed their feats with very little planning and rarely did their post-feat paperwork, but they made up for these shortcomings with their fearlessness, their strength, and their energy. The king was also lucky to have a very good prime minister. The minister saw that taxes were collected, that the majority of the revenues were wisely spent to improve the kingdom, and that the excess monies were well managed.

One day the prime minister died, and the king asked his bravest and most fearless knight to be the new minister. Unfortunately, the knight chosen for the job liked jumping on his white charger more than he liked caring for the affairs of the kingdom. He continued to gallop off to slay dragons whenever he had the opportunity and neglected tax collections, investments, improvements, and repairs. After only a short time, the king ran out of money, the peasants ran out of patience, and the kingdom collapsed.

My dad compared the brave knights to the very best marketing and development employees. These people loved to jump on airplanes or into impossible situations to save the day by closing big sales or inventing something new. Their first response was to act, and results were their primary concern. They used a Ready, Fire, Aim approach and in many cases were successful.

Unfortunately, this approach failed if a problem required careful planning or attention to small details. Brave knights generally did poorly when assigned to management positions. They were always too busy and too involved in the current crisis to spend much time supervising others or planning for the future. They lacked the patience to stick around long enough to train and support their people.

Good prime ministers made better managers than did knights. Their first course of action was to think, and their primary concern was to come up with a good plan before they wasted a lot of energy. They stayed late to finish the paperwork and always knew how much money they were spending.

This story helped me begin to understand the magnitude of the problem we were facing. A company needs its superstars, but it also needs good managers, who make careful decisions, who pay close attention to details, and who care enough about the people they supervise to dedicate some time to them. If we were to be successful over the long term, our superstars would either have to take the time and trouble required to become good supervisors or let others take over.

In our case, we were a company filled with superstars. We had all worked on impossible problems and had all made significant contributions to the company’s success. We had enjoyed the freedom to run our own areas of the company as we wanted, and few of us felt we needed any more supervision. If there was to be a new structure to the organization, we all wanted to be at the top. My dad suggested we consider limiting the size of the company and try to avoid the problem altogether. In all his years of experience, he had yet to see a knight turn into a good prime minister.

I was discouraged enough with the problem to ask Alan if perhaps we should be more cautious and put a cap on the number of employees in the company. Neither of us had dreamed the company would grow so quickly. Though our sales were increasing, our expenses were increasing as well. The responsibility of meeting these expenses was an enormous weight on our shoulders. As a result, I think we were ready to slow things down a little.

On the night after our conversation I had a dream. In it I was in the reception area of our offices standing knee-deep in money. There were three or four other people in the room with me, and together we were trying to pick up the money and stuff it into bags as quickly as possible. I remember wishing we had more people to help us. The money was ours to take if only we would pick it up.

This dream was not too far from what the software business was like for us. It reminded me that the opportunity was there for the taking, if only we did not lose our nerve and back off. The next morning I told Alan about my dream and that I no longer wanted to put a limit on the growth of the company. Opportunity was not just knocking, it was breaking down our door. All we had to do was get out of the way.

From that time on I felt like WordPerfect had a life of its own and that I was merely a spectator to an amazing phenomenon. I realized that I had no way of predicting or controlling how big the company would get in the next few years. Though I could not control its size, I was going to try my best to control its shape and structure. I saw myself as an architect, with the responsibility to keep the organization flat and efficient, with the fewest number of management layers possible.

I was not sure how much Bruce and Alan wanted me to take on this responsibility, but I did it anyway. As a college student, I had worked two summers for government contractors and observed real bureaucracies first hand. Remembering my experiences, I was determined that SSI would not fall into the same unproductive traps as had my early employers. These corporations were full of organizational charts, formal lines of communication and authority, and a lot of wasted time and effort. My getting to work on time and getting off to a fast start was discouraged by most of the other employees. The work was done in bits and pieces, when it did not interfere with taking breaks, eating lunch, playing bridge, talking sports, participating in betting pools, bringing each other up to date on our personal lives, celebrating birthdays, or watching and commenting on the girls in the office. As a young idealist, I had hated working at a slow pace in order to fit in with the rest of the guys.

My time spent working for Don had a lot to with the way I felt about supervisors and helped to shape my beliefs as to how supervision should be handled in an organization. During my first year at SSI, my financial duties took up only a small amount of my time. I spent the majority of my time working as Don’s assistant. When he was out of the office, I served as a messenger between him and our clients. When he was in the office, I did whatever he told me to do for the day. When I wrote a letter, I had to sign his name at the bottom. Any training I got came in the form of lectures given when I messed up or did too much on my own. I felt like a little kid who could not go out to play because his mom wanted him at her side in case she had an errand for him to run. Once Don had left, I enjoyed working without someone looking over my shoulder. I came to the conclusion that working without a supervisor was better than working for an overbearing one.

Taking on the role of company architect did not help my image much. I was already the designated bad guy on the Board. Bruce spent much of his time in Europe, and when he was in the office, he was either working on international issues or helping with development problems. Alan was always nice and agreeable and preferred to avoid contention. He was much better at smoothing ruffled feathers and settling disputes than he was at handing out bad news. That left me to say no, if no had to be said.

Because I watched over the finances and jealously guarded the distribution of our money, I was in a position where anyone wanting to try a new program or hire a new employee had to come to me. I played the role of the sheriff in a Wild West town, where most of the town’s citizens were unsure they wanted a sheriff. Bringing some sort of financial law and order to the company while attempting to maintain a flat organization meant disappointing a lot of people.

A flat structure by its nature is bound to discourage some, because it limits management opportunities. If you have an organization which requires two hundred people to do its work, and if the supervisors in the company each have no more than five people reporting to them, then you will need approximately 50 supervisors. This tall type of structure will likely have a president, two to four vice presidents, five to ten directors, and forty managers. Counting the two hundred employees who do the actual work, you will have five layers in the company.

If you were to change the number of people supervised by one person from five to twenty, you would create a much flatter structure. In theory you would need only a president and ten managers. This would reduce the number of layers in the company to three and the number of managers to eleven. Not only would this flatter organization be likely to function more efficiently, but the company would also save the salaries of 39 managers.

In a real company of two hundred people, eleven in management is not likely enough. Not all departments and functions break down to exactly twenty people each, and not every manager is likely to have the experience and skill to help and support twenty people. Still, a company this size should be able to function very well with no more than twenty to twenty-five managers and no more than four layers.

Within a flat structure, the management opportunities are likely less than half those found in a tall structure, so the odds of moving up the management ladder are much lower. If the company in our example were to grow to the point where 1000 people were required to do its work, a tall organization would need to add 200 managers, while the flat one would need only 43 more. There is room for everyone to move up in the tall organization, but in the flat organization there is room for only one in five.

For those who joined SSI to move up a career ladder, this flat structure was a source of tremendous frustration, and their frustration was generally directed at me. If employees needed help to get all their work done, they would usually want to hire assistants to help them. Instead of letting them hire subordinates, which usually created another layer in the company, I would ask them to give up part of their responsibilities. If we needed new employees to take those responsibilities, the new people would be hired as peers rather than subordinates to those requiring the help.

For example, Doug Lloyd, the salesman we hired away from IBM, at one time had responsibility for all large accounts, including large corporations, federal agencies, and schools. As we grew and he had more work than he could handle, we turned his government and school accounts over to two other people. Later, when he needed help with corporate accounts, we gave two thirds of his accounts to two others. In both cases, the people who were given a part of Doug’s responsibilities became his peers, reporting to the same supervisor he did.

For most people, and especially for the company superstars, losing responsibility was very discouraging. Not only were they not moving up in the organization, but they viewed their losses as demotions. Instead of having more and more assistants, they saw themselves as having more and more rivals. How could they explain to their spouses or parents that because they were doing a great job they were having to give up responsibilities?

I came to believe that I would have to live with my share of enemies inside the company if we were going to keep a flat organization. As the company grew and more supervision became necessary, I made additional enemies. Some people who had reported directly to the Board felt slighted when asked to report to a supervisor a step below the Board. Although I tried to ease the pain with a good explanation or by handing out new and better titles, there were always a few people who could not understand why the company’s growth did not give them a more important or more prestigious role.

To many it appeared as if my reason for promoting a flat structure was to exercise greater personal control. While I did not agree that my preference for a flat organization was to promote my own self-interest, I would agree that a flat organization concentrates more authority in the hands of fewer people. It also promotes better communication within an organization, because there are not as many layers to filter the information.

In spite of our problems and my efforts to solve them, 1985 was proving to be another great year for SSI. Sales of WordPerfect for DOS were going through the roof, overshadowing the fact that our other products could hardly make it out the door. Ironically, we were as determined as ever to develop more products–even if many of them were unsuccessful.

WordPerfect jr and MathPlan jr sold as poorly as IBM’s PC jr, which was a miserable failure. Not only did we fail with WordPerfect jr in English, but we went ahead and failed in Danish and Norwegian as well. Personal WordPerfect, designed to compete with PFS:Write and to fill a gap between the full version of WordPerfect and our jr version, sold very few copies. Sales of MathPlan for DOS and MathPlan for AOS went nowhere. SSI*Legal, Lew Bastian’s legal time and billing package; SSI*FORTH, our first and last programming language; and SSI*Data, our poorly received database all had disappointing sales. P-Edit was still around and selling poorly. Luckily, we came to our senses before releasing TranSSIt, deciding not to ship the low-priced communications package because the support costs were likely to be too high to show a profit.

The one new product which did sell well was our Apple II version of WordPerfect. Although we might have made a lot more money with a Macintosh version, we started out so far behind on the Mac that we decided to finish our Apple II version first. We were too small a company to get the kind of attention and help from Apple that Microsoft was getting. Bill Gates had already tilted the Macintosh software playing field so much in his favor that we were discouraged about our prospects in the new market. Unfortunately, we would not release WordPerfect for the Macintosh until 1988, giving Microsoft Word for the Mac a four year head start.

Looking back, it is easy to see we were trying to develop and sell too many products too quickly, but we were as determined as ever to be more than a one product company. Even though our 1985 sales would more than double our 1984 sales, we felt a considerable amount of embarrassment because we could not produce another best- or even good-selling product.

We could have saved ourselves a lot of headaches by concentrating on fewer products. For every product we released, we had to manufacture a standard retail version, an update version for customers who had purchased the previous release, a demonstration version for dealers, a special version to give away, a low-priced version for schools, and versions manufactured especially for large corporate and government customers. Not only did we have to prepare and keep track of these complete packages, but we had to make the templates and manuals available separately. In addition to our US English products, we had to make similar but different sets of English packages for our Canadian, British, and Australian customers. We also made two sets for our French speaking customers in France and Canada and additional sets for all the other international languages we supported. We did not make our mistakes one a time. We made them twenty at a time.

Developing the software and the software packages was, of course, only part of the work. We had to figure out how to price the products and how to roll them out. We had come up with the brochures, advertisements, and promotions. Then we had to spend hours and months trying to figure out what we were doing wrong when the products did not sell. It was rarely appropriate to blame the developers if a product did poorly. We always had to keep searching for a marketing solution to our failures.

Fortunately, we had the foresight to save enough of our time and resources to get WordPerfect 4.1 ready for release in the fall of 1985. This version was not thrown together at the last minute like 4.0. We were now involved in what the press called a features race, a race we knew we could win. 4.1 had at least 100 new features, including an automatic table of contents, automatic indexing, footnotes which could spill over to the next page, paragraph numbering, an improved speller, and a thesaurus. All the new features were written by our programmers, unlike our competitors, who were looking to other companies for things like thesauruses and spellers. Not only did we save a lot of money by doing all the work ourselves, but all the features we built functioned similarly and worked well together.

Our key strengths had always been our programmers and our ability to keep them from leaving the company. When some of our competitors finished a product, their programmers seemed to scatter like pins after a strike in a bowling alley. It would usually take them months to line up a new programming team to work on an update version of an existing product. We kept our teams intact, however, so we lost no time between releases.

The changeover from 4.0 to 4.1 was a big manufacturing challenge, because this was the first time we were dealing with such large numbers. Sales of $2,000,000 meant that about 100,000 copies of WordPerfect were going out the door each month. That was a lot of paper, binders, and diskettes for a company doing things by trial and error.

Our biggest concern was making sure we did not end up with extra copies of version 4.0. During the mid-80’s, as soon as a new version came out, the old version became obsolete. If we ended up with 50,000 copies too many, for example, we would lose one million dollars. If we ended up with 50,000 too few, we would be out of product and out of sales for two weeks. Hitting the number right on was difficult, because the release date was a moving target.

As the 4.1 release date neared we became very nervous. With 3.0 and 4.0 we experienced a slowdown just prior to release, as word of the release leaked out from our beta test sites. We factored this slowdown into our estimates, and for the first and only time, word of the new release did not leak out. Our sales were increasing at a time when they should have been decreasing, and we were in danger of running out of WordPerfect for DOS for perhaps as long as a month.

Our security was so good that I was forced to ask someone inside our company to call Spencer Katt, the writer of the rumors column at PC Week, to spread the news of our upcoming release through the industry grapevine. Once news of 4.1 appeared in PC Week, sales of 4.0 dropped back below $2,000,000 a month, and we were out of product for less than two weeks. Our entire marketing department was so angry about the security leak, that I had to wait a few months before I could tell them I was the culprit.

Once 4.1 was released, our October sales jumped to more than three million dollars. We could have sold more product that month, but our sales estimates were low and our printing company did not have the capacity to meet an increase in our orders. For the first time we had to ration our product to distributors. WordPerfect 4.1 would be in short supply until the first of the year, when we would get more printers printing our manuals.

COMDEX that year was a lot of fun. Mark and Sherry, our professional actors from Chicago, were back doing another song and dance presentation. They added slides to their show which featured screen shots of many of the new features in 4.1. There was so much interest in our new version that we no longer had to drag people into our booth. Instead, people sought us out and told us how much they liked our product.

We ended the year with sales of $23 million, two and one-half times those of 1984. That made us the number two word processor, since Ashton-Tate had acquired Multimate and leveled the sales of that product off at $20 million. Only WordStar was ahead of us, but we were right on their tail. Their quarterly sales were down to only $10 million, and ours were up to $9 million. At Softsel, one of the two largest distributors, we were already outselling WordStar.

Our pre-tax profits were up to 25%, so we were starting to make enough money to play the software game like professionals. With InfoWorld giving us another perfect score for WordPerfect 4.1, we knew 1986 would be another great year for SSI. Although we could feel the growing pains which accompanied our increase from 84 people to almost 200, we looked like a company that knew what it was doing and where it was going.

Go to Chapter 7