Death by Doppelganger Micro Pro

Rubinstein's initial goal in founding MicroPro was to develop and publish a high-end database management system (DBMS) designed to compete with Ashton-Tate's dBASE and similar products, but during his stint at IMSAI, he learned the CP/M market needed a good programmer's text editor. Because developing one would take less time than a full-blown DBMS system and provide the company with a revenue stream until the database product was ready, Rubinstein hired Rob Barnaby, a top-notch assembly language programmer, to build the product. Barnaby, in an inspired burst of creativity, wrote 137,000 lines of code in 4 months and produced both the editor and a high-speed sorting program intended to be the first component in the forthcoming database program, Supersort.

A Star Is Born

Barnaby's text editor was christened WordMaster and upon its release in 1976 sold so well that Rubinstein decided to take the next step and release a full-featured word-processing program based on WordMaster. The new product, named WordStar, hit the market in 1978 and quickly became the dominant product in the CP/M market. The product was so highly regarded that it even became popular on the Apple II, as people bought CP/M computers on a board and slipped them into their Apples so they could use WordStar.

There were several reasons for WordStar's early success. The first was power: For its day, the product was feature packed. The second was what came to be known as WordStar's Control-key interface. Rubinstein had deliberately designed WordStar to meet the needs of touch typists. To enter commands in the program, you held down the Control key (most CP/M systems of the time had one) and pressed a key. WordStar's layout was not mnemonic; instead, in the interest of fast typing, Rubinstein designed the interface so that all cursor movements were performed with the left hand while less common operations fell to the right hand. WordStar users came to swear by this system, and today diehards still retrofit Microsoft Word and other products with add-ins and utilities that resurrect the WordStar keyboard system.

The third and most important factor was that WordStar was the first What You See Is What You Get (WYSIWYG) word processor. Prior to WordStar, formatting text with a software product meant sprinkling formatting commands amongst blocks of text, printing the document to see the results, and then sprinkling in more commands and reprinting until you were satisfied with the results (a process very similar to working with raw HTML and an editor today). WYSIWYG, a term coined by Rubinstein, meant something far different in 1978 than it means today. WordStar, like all early CP/M and IBM software, ran on character-driven screens that couldn't display different fonts or combine graphics with text a la the Mac or a Windows machine. Nonetheless, the software accurately displayed line lengths and paragraph breaks (assuming you were willing to concede everything you would print was set in 10 pitch) and allowed you to set margins and tab stops onscreen. Soon most word processors were emulating this new approach to editing.

By 1983, WordStar's success had made MicroPro International the largest microcomputer software company in the world, with sales peaking that year at close to $70 million. During this period, MicroPro attempted to diversify into other markets, publishing InfoStar, Rubinstein's long-dreamed-of database product; ChartStar, a business graphics product; and even an unfortunate spreadsheet called CalcStar. (It was unfortunate because the product was infamous for its bugs. Until the product went to its well-deserved and unheralded demise, an entire row of the CalcStar workspace was nonfunctional, and internally the product was known by such nicknames as "WoofCalc" and "DogSheet.") MicroPro even briefly attempted to manufacture its own CP/M computer, the PBM1 (supposed to remind you of IBM) until someone came to his senses and shut the project down. However, none of these other software products sold particularly well, and WordStar remained the pillar on which the company's fortunes rested.

Version 3.3 of WordStar for both IBM and CP/M computers was released in 1983 and sold briskly, and all seemed right with MicroPro's world. Unfortunately, the situation soon changed. Rubinstein had gotten into a contretemps with his WordStar development team, and they, depending on who is telling the story, either a) quit or b) were fired. (The departing programmers promptly set up shop in an office not far from MicroPro headquarters and proceeded to found a new company called NewStar, which published a WordStar clone called NewWord. Their fate and MicroPro's would become closely intertwined.)

In any event, at just about the exact moment MicroPro needed to ship an update to WordStar, it had lost the ability to do so.2 No update to WordStar would appear in 1984 or even in 1985. A 12- to 18-month upgrade cycle had become the norm in the software industry, and competitors were busy building new products that matched, then began to surpass, WordStar's capabilities. Things looked bleak until an unexpected savior appeared on the scene.

This white knight was brought to MicroPro courtesy of AT&T. The phone company was about to begin a disastrous foray into microcomputing by introducing a line of new desktop-based UNIX computers that

1 These computers were assembled in San Rafael, California, on the checkout counters of a former A&P supermarket. They sported dual Z-80 processors, a 5MB hard drive, a quad-density single-sided 5-inch floppy, and Televideo terminals, and they were preloaded with MicroPro software. Only about 100 were ever built, and they were sold to the company's employees. The units were originally supposed to be called "SyStars" (for Seymour Rubinstein). Rumor had it that the reason MicroPro went briefly into the hardware business was that Seymour was jealous that his friend Adam Osborne had a computer named after him.

2 One of the difficulties in upgrading WordStar lay in the practice, common at the time, of implementing bug fixes by directly modifying the binary executable rather than updating the source code and reassembling the program (reassembling was a lengthy process). When the development teams working on WordStar examined the original 8080 source code, they found it didn't match the WordStar.exe files being shipped in the latest product. The lack of documentation on what fixes had been implemented made working with the WordStar code base very difficult.

would fail to sell in any significant quantities. AT&T decreed that some choice software fodder needed to be produced for its forthcoming line of white elephants, and the company proposed that MicroPro port WordStar to its UNIX operating system and the C language in return for some cold, hard cash.

MicroPro actually lacked the capability to do this, but a seeming bit of serendipity intervened. Rubinstein got wind of a new software product developed by a programmer outside the company that was written in C, ran under UNIX, and cloned WordStar's functionality and design. Seymour took a look at the embryonic word processor, bought it, hired the programmer who had written it, and told him to hire a small team of coders and port WordStar to UNIX.

Operating outside MicroPro's normal corporate structure, the team worked busily for several months at their task. When they were done, the results of their work weren't what Seymour had originally envisioned. The new "WordStar port" used a mnemonic set of Control key-based commands, possessed some features that WordStar lacked, lacked some features that WordStar had, and sported a new file format completely incompatible with the original product. It was written in C, and it did run on PCs and the AT&T UNIX boxes. And it was clearly not WordStar.

But by this time MicroPro was desperate. It was now more than a year and a half since the release of WordStar 3.3, and the program was growing very long in the tooth indeed. MicroPro decided to make the new product the focus of its future sales and marketing efforts. The new product was named WordStar 2000 (the idea for the "2000" was lifted from the logo of a local furniture store). WordStar 2000 was priced at $495.00, then the median price for a high-end word processor, and rolled out in 1985. The original WordStar remained on the shelves (it was still selling strongly, though sales were slowly declining) at its suggested retail price (SRP) of $495.00.

All hell promptly broke loose. With its release of WordStar 2000, MicroPro had just committed a fundamental positioning mistake. The company would pay dearly for this mistake, ultimately with its very existence.

The Doctrine of Positioning

Positioning as a marketing concept became all the rage in the 1970s and 1980s, and a great deal of time and ink has been dedicated to the topic. The Orthodox Creed of Product Positioning, as decreed by one of the great cardinals of high-technology consulting, Regis McKenna, is that positioning is a

. . . psychological location in the consumer's mind, pertaining to the relative qualities a company, product, or service may have with respect to its competition.

The "relative" qualities a company, product, or service may aspire to in the buyer's mental geography include the following:

• Best quality

• Most popular

And so forth.

The virtual locations most desirable for your product or service depend on its particular characteristics, your market, and your competition. For example, in the case of Joe Whitebox's local computer company, it can't credibly claim that it's the leading manufacturer of desktop computers; Dell, Compaq, Hewlett-Packard, or another company owns that "location" in the market's mind. But Joe Whitebox might seize the "service" terrain because he runs a local business and has a shot at making that claim stick.

The Orthodox Creed has, however, often proved inadequate to the needs of software companies. This is because software, by its nature, is an abstraction. The Reformed Creed of Product Positioning for software states that positioning begins with describing a product in such a way that the purchaser can tie it to a real-world process or object. On the face of it, this seems like an easy, straightforward thing to do, and sometimes it is. For example, when word-processing software for desktop computers was first introduced, most people quickly grasped the idea that these products put "a typewriter in your computer." The benefits of fast revisions, spelling correction, and flexible formatting of documents were immediately apparent.

But for other categories of software, positioning has proved to be far more difficult. One of the most famous examples is Lotus Notes. If you're in the software industry, you've certainly heard of Notes and you may even use it. But when the product was first introduced in 1989, Lotus seemed unable or unwilling to explain what the heck the product did. The Lotus Notes 4.0 documentation, rather pathetically, highlights the problem best:

What Is Notes Anyway?

People have been asking that question since the beginning of time (or at least since Notes first came onto the market). It has been hard for people to define Notes because you can use it to do so many things.

— From the Notes 4.0 Beginner's Guide, published in 1996

Actually, this documentation never does tell you exactly what Notes "is." As you can imagine, the Lotus sales force had a great deal of trouble explaining why someone should buy Notes when the company that published it couldn't explain what it was.

(Oh, what does Notes do? Well, its most popular function is as an e-mail management program, a post-office system for your computer network. The most obvious feature that differentiated the product from its competition was that the electronic letters you sent back and forth could be annotated with notes and comments. Other people could see your comments and add their own "under" yours. It's not that hard to explain and tie to the real world, but Lotus somehow could never bring itself to do so.)

Positioning Wars

MicroPro's positioning mistake was of a different nature than Lotus's and far more difficult to manage. The release of WordStar 2000 created an irreconcilable positioning conflict that pitted MicroPro against itself. After the product's release, the WordStar user base took one look at WordStar 2000 and decided no thanks, opting instead to sit on its hands until MicroPro released an upgrade to its favorite product. As a result, MicroPro now found itself selling two high-end word processors called WordStar for $495.00 to people using IBM PCs. Precious marketing resources had to be expended in creating collaterals, ads, and promotions for the two products while attempting to provide a convincing rationale for the existence of both.

It was an impossible task. A day selling WordStar 2000 to the market went something like this:

MicroPro: Hi there! We're here to tell you about WordStar 2000, our new word processor!

The Market: Great to see you! But, we have to tell you that although WordStar is a wonderful product, it's hardly new. You must mean this is the new upgrade. Great! I'm so excited! Let's take a look!

MicroPro: No, no, this is WordStar 2000! It's really totally new!

The Market: Oh. (Long pause.) When are you releasing the upgrade to WordStar? In 2000?

MicroPro: No, no, the upgrade to WordStar will be released real soon now!

The Market: Oh. (Longer pause.) Well, why will you release a new product in 2000 when you haven't released the upgrade to WordStar?

MicroPro: No, no, WordStar 2000 is available right now! We just call it "2000" because it's new and powerful and easy to use! But you don't have to wait until 2000 to enjoy all those benefits! The Market: Oh. (Dead silence.) In other words, WordStar, which won't be upgraded until 2000, is old and not powerful and hard to use?

MicroPro: No, no, no. WordStar is a classic and is powerful and has a wonderful interface for touch typists!

The Market: Oh. Does that mean you can't touch-type with WordStar 2000?

MicroPro: Don't be silly! Of course you can! It's easy to type with WordStar 2000's new mnemonic commands!

The Market: Then is it hard to type with WordStar's regular commands? And don't call me "silly."

MicroPro: Sorry about that! No, you can type quickly with WordStar!

The Market: Then you have to type slowly with WordStar 2000? MicroPro: Uh, no. You can type really well with both of them! The Market: Oh. (Long, long pause, dead silence.) Now, what's the difference between WordStar and WordStar 2000 again? And why do I have to wait until 2000 for an upgrade to WordStar?

And so it went. Endlessly. Instead of answering why prospective users should buy WordStar, the MicroPro sales force for years tied itself in knots attempting to explain the difference between two products named WordStar.

The confusion within MicroPro was just as pernicious, as the company began to split internally along WordStar/WordStar 2000 fault lines. Within MicroPro there were WordStar aficionados and WordStar 2000 mavens, and each side wondered what the other saw in its choice of a word processor. At one point, the head of the WordStar product development team forbade team members from talking with WordStar 2000 programmers. (A neat trick, because both programming teams worked in the same building.) By 1987, as MicroPro wrestled itself to the mat, it had ceded its leadership of the word-processing market to Microsoft Word and Corel WordPerfect.

Yet when things were darkest, MicroPro seemed to come to its senses. A new president, Leon Williams, and new product management, including myself, were brought in to try to sort out the mess. Two things needed to be done immediately. An upgrade for WordStar had to be released ASAP, and something had to be done about the conflict between WordStar and WordStar 2000. I was given the task of figuring out the positioning strategy.

The first thing Williams did was trot down the street to NewStar software, buy the company, and use its NewWord product as the foundation for an upgrade for the long-suffering WordStar user base. The upgrade, called WordStar version 4.0, sold well into the WordStar installed base, though its feature set wasn't truly competitive with other products of the time. MicroPro even released a new CP/M3 version, which did surprisingly good business and garnered the company much favorable PR. The gloom surrounding MicroPro started to lift.

Repositioning WordStar 2000 was a more difficult task. The logical thing to do would have been to simply shoot the product. Unfortunately, this wasn't practical. Since WordStar 2000's introduction, a fair number of people had bought the program, and its sales represented an important revenue stream. Despite MicroPro's fervent wishes, WordStar 2000 was going to stick around for a while.

My short-term answer to the positioning conflict was an approach I came to call "façade." This strategy consists of taking a look at two products in conflict, deciding what key features differentiate them, and repositioning one product "away" from the other. In The Product Marketing Handbook for Software I describe the goal of a façade program as an attempt to

. . . Buy time . . . to maneuver yourself out of having to explain the differences between the two products so that you can talk about what the products are and why the buyer wants them.

By its nature, a façade approach to a positioning conflict is a transitional strategy. When done correctly and with finesse, it can provide a company with the opportunity to decide whether it's possible to kill one of the conflicting products, either via a migration strategy or via a merger, or perhaps relaunch it into a completely different market.

After a quick analysis of the options, WordStar 2000 was rechristened a "word publisher" (the actual phrase was coined by one of MicroPro's top salesman, Jim Welch, who, along with the rest of the MicroPro sales force, was slowly going insane attempting to explain the differences between the two products). And what, you may ask, is a "word publisher"? Well, a word publisher is a word processor with exceptional laser-printing capabilities, a particular strength of WordStar 2000 at that juncture. Of course, this claim couldn't withstand market scrutiny over time; in reality, there was no such thing as a word publisher. The claim to differentiation was credible only as long as WordStar 2000 was

3 WordStar 4.0 for CP/M was the last major commercial release of software for this OS.

I have a copy on 8-inch disks.

superior to its competition in this particular aspect of the product. But in the short term, the campaign worked as intended and bought MicroPro some time and maneuvering room. Sales and market share of both MicroPro word processors increased, and the price of MicroPro stock rose.

Stupid Printing Tricks

As a "reward" for my efforts, I was "promoted" to group product manager and given responsibility for the product management of the resurgent WordStar. A new version, WordStar 5.0, designed to build upon the momentum built by the successful 4.0 release, was being hurried along to market. If MicroPro could launch it in a timely fashion with a competitive feature set, there was a chance the company could regain its lost market leadership, or at the very least generate enough revenue to branch out to new and more lucrative opportunities in other software categories. The product was slated for release in early 1988.

The first thing a product manager does when he or she is assigned responsibility for a new product is take a look at it, and I was soon handed a fistful of disks that contained the latest version of WordStar. Like any upgrade, it had a raft of new features and capabilities, but to my annoyance you couldn't print with it. A quick look at the files that made up the program revealed why: The newest version of WordStar lacked a printer database.

Now this was odd, because if there was one thing MicroPro had learned to do over the years it was to support printers. In the pre-Windows era it was the responsibility of software developers to obtain, test, and debug printers and their drivers to ensure they worked with their particular products. As of 1987, MicroPro had built a quality database of more than 300 printer drivers. The information in this database represented years of careful debugging, testing, and implementing capabilities specific to each printer. When you installed a printer in WordStar and told the program to print, you could be fairly confident your text wouldn't appear upside down or in a character set that resembled Sanskrit.

What made the omission of the database even more puzzling was that in 1985 a decision had been made at MicroPro to base all future printing code for other products on the WordStar 2000 printer database.

It was tested and debugged, and it was extensive. MicroPro had introduced a low-end word processor, Easy, that utilized the 2000 database. Why wasn't it in WordStar 5.0?

Several inquiries made by me to the development group elicited vague responses about "new support" issues and "implementation questions." A sense of dread began to haunt my soul. A heavy weight seemed to descend upon my shoulders. More inquiries elicited even vaguer answers. The weight pressing down on me grew heavier. It was time to find out what was going on.

As a product manager I had developed the habit of periodically stopping by the MicroPro development center to schmooze with the programmers about product features and problems while providing them with feedback on what our customers liked and disliked about our programs. One fateful day I headed to the center and floated by the section occupied by the WordStar programming team. While skulking about, I saw a group of agitated programmers pointing at a screen and arguing heatedly.

Sidling closer, I listened to their conversation with growing horror, and then I heard a word that confirmed the bad news I'd been overhearing. The impact of this word on me was stunningly physical. On hearing it, a bright light burst upon my eyes and filled them with a dazzling clarity, one that let me see the future. Simultaneously, the great weight was lifted from my shoulders. This wasn't because I was feeling better; rather, it was because I no longer had any shoulders as I underwent a miraculous transformation from product manager to small gray rat desperate to abandon a ship I knew would soon be sinking.

That word was "pointer."

As in a hierarchical pointer. As in a hierarchical database pointer. As in the development group had decided to discard the WordStar 2000 database and replace it with a new one based on hierarchical database technology. It was an incredibly foolish thing to do, and it sealed MicroPro's fate.

To understand why this was a disastrous course, you need to have an understanding of database technology, something that I, having once worked as a DBMS programmer, possessed (and something the previous product manager had not). WordStar 2000's printer database was basically a flat relational table. When you installed, say, an HP LaserJet printer, the WordStar install program looked up the driver information for this unit from a row in the printer database. Specific printer functions, such as boldfacing and italicizing letters, were stored in columns within this row.

The new hierarchical database being built for WordStar discarded this paradigm. Printer information was stored in something that resembled a tree. Pointers were used to locate specific information about printer functions within the tree.

In all fairness, there were some minor technical advantages to this new printer structure. For instance, it would be smaller than the 2000 database. MicroPro might save the cost of a floppy in the WordStar cost of goods. But I also knew that hierarchical systems had fallen into disfavor after the introduction of relational technology. No commercially available programming tools or utilities were available on our desktop development platforms to convert the current flat table structure to a hierarchical one. Porting the printer database to the new model would first require building a series of custom programs to accomplish the task. This would take months. Then the tools themselves would have to be tested for proper operations, which would take more time. Of course, once the database had been ported to the new structure, all printer operations would have to be retested to ensure the accuracy of the process, which would take even more months. There was no way WordStar 5.0 was going to meet its projected ship date or even come close to it.

Once I confirmed what was going on,4 I went squeaking to my boss, the vice president of sales and marketing, and warned him of our impending shipwreck. VP to VP, the head of MicroPro's development assured the head of sales and marketing I was exaggerating the situation. Officially, WordStar 5.0 was still on track in development and would ship on time.

As the weeks went by and WordStar still refused to print, I prepared to move myself and my skinny pink tail to what I hoped would be a more favorable clime (I was wrong, by the way). I accepted a senior

4 Final confirmation, from my point of view, came during an impromptu basketball game near MicroPro's headquarters. Steve Evangelou, a talented programmer at the company, and

I had gotten into the habit of driving to a nearby court to shoot some hoops and discuss company gossip. During a lull in our game he informed me that the WordStar 5.0 project faced some "issues." Before he went further, I interrupted him and said, "Let me guess. You guys have decided to discard the WordStar 2000 database in favor of some hierarchical system, and you have no idea of how to port the data. And we're not meeting our ship dates." He looked at me and said, "I guess you've got a handle on this after all."

product management position at Ashton-Tate (the company had a lousy word processor, but I was pretty sure it would print) and handed in my resignation with a final warning to all that financial projections based upon WordStar 5.0 revenues needed to be revamped. On my last day at MicroPro, as I left corporate headquarters and walked through the parking lot to my car, the company's director of direct promotions bustled over to say good-bye. Before I pulled out of the lot for the last time, he informed me that that morning the vice president of development had finally confessed that WordStar 5.0 wasn't going to meet its ship date. Nor could he provide a firm estimate of when it would ship. He quoted my boss as saying, "Rick Chapman told me this was going to happen."

The WordStar development group's decision to discard the company's existing printer technology delayed the critical 5.0 release for more than half a year. When the release did ship in late 1988, the new version was widely criticized for having a printer database about one-third the size of that of previous WordStar products. Upgrade sales, as well as sales to first-time buyers, were disappointing. Time spent re-creating the printer database was also time not spent on adding new features to the product that would have made it more competitive. The cumulative effects of three blown financial quarters and disappointing sales led to MicroPro's upper management, including Leon Williams and my former boss, being marched out and treated to a summary executive execution. MicroPro had lost its last chance to regain its footing in the market, though the company staggered on in zombie-like fashion for several more years, living off its steadily decreasing installed base of WordStar users. WordStar finally faded away in the early 1990s, subsumed in a merger with a flock of similarly unsuccessful and second-rate software companies. It was an ignominious end to the career of a great piece of software.

The question that remains, of course, is, why? What had possessed the development group to embark down such a destructive path? What were their motivations? The technical case for their actions was never strong. That this was the wrong thing to do from a business standpoint was even clearer.

The answer lay in the positioning conflict unleashed within the company. While MicroPro worked hard to placate a confused market, within the company the WordStar versus WordStar 2000 struggle raged on. The WordStar programming team hated WordStar 2000 with a passion and wanted nothing from that product to pollute "its" WordStar.

Its decision to rip out the existing printer technology was based on emotion, not a rational cost-benefit analysis of the consequences of such a course.

Positioning problems constantly plague high-technology companies, particularly software ones, because of the industry's rapid pace of change, the malleable nature of software, and acquisitions. In 1991, Borland International split itself along Paradox versus dBASE lines via its purchase of Ashton-Tate. Novell, like MicroPro, would shoot itself in the foot by creating two competing product lines with its purchase of UNIX from AT&T. Today, Sun wrestles with the issue of Solaris vs. Linux. And in 1993, Microsoft demonstrated with the release of Windows NT that previous success doesn't necessarily provide protection against future stupidity.

Continue reading here: Who Killed OS2

Was this article helpful?

0 0