Lotus 1-2-3 for mainframes (bringing a product to market)
Kevin R. Weaver · ACM SIGAPL APL Quote Quad · 1985
The number of widely-used, business oriented APL applications is few. It's difficult to pinpoint why — other than to reflect on the complexities of conceiving a marketable product and executing the development, release, sale and promotion, distribution, and ongoing support for a product. Although many companies provide APL services, and many custom applications have been developed for specific needs, the ability to bring a mainframe APL product to market and have it succeed seems cornered by IBM (e.g., ADRS, FPS, GRAPHPAK, APL DI). Parallax Systems is one of the few companies to be successful in this arena. The product is called ExecuCalc tm , and has been licensed to approximately 200 sites since January 1983. What are the elements that contribute to the success of this product in such a short period of time? Can these be applied to other APL product ideas to insure the developer is at least on the right road to producing a winner? First, there was market awareness. The buying market (corporate data processing managers and senior decision makers) should have a mental picture of the concept being explored. ExecuCalc is an electronic spreadsheet product. In the early 1980's, VisiCalc was almost a household word. The concept was clear, accepted, and in practice in the micro computer market. Missionary work (often a key process in promoting APL and APL-based applications) for spreadsheets was already done, and the masses were going to the altar. Second was market need. VisiCalc satisfied a need. Since this audience accepted the spreadsheet concept long ago, the market need does not have to be detailed. At the same time, VisiCalc lacked both obvious features (e.g., variable column width) and features which resulted from widespread use creating a demand (e.g., file combine for consolidation). ExecuCalc initially contained all of the VisiCalc features, including identical commands and their syntax as well as file formats (“.VC” and “.DIF”), plus a few of the obvious needs (determined by limited market research and personal experience, harnessed by the desire to release the product in a timely fashion). ExecuCalc paralleled the need for a VisiCalc, and added a new dimension: it ran on a mainframe using 3270 terminals. The needs satisfied included: (a) spreadsheet, (b) a widely accepted approach, (c) use of existing hardware (mainframe and 3270s), (d) a product requiring little training and support for information center users, yet (e) enabled users to be immediately productive, and (f) enabled users to process corporate data resident on their mainframe rather than down-load and proliferate segments of the corporate databases. Third, boundaries were established on the features in each release, especially release 1. By concentrating on a well-defined end of development, Parallax was able to bring ExecuCalc to the market in three man-months. The boundaries were set by VisiCalc's features and the few “obvious” add-ons. Subsequent releases over the next year were timed and delivered as maintenance releases. The subsequent releases always were bounded by a finite set of new features. Fourth, development was focused and uninterrupted. As a two-person company, Parallax had no alternative other than to focus on a timely completion of the product. We had financial incentives (the desire for income and the need to minimize expenses) as well as the concern that a competitive product would be released and substantially reduce our potential. There is an undisputed edge in being first in the market. As a result, one person spent long hours at the terminal keyboard (implementing, testing, and polishing the product) and one person spent long hours at the typewriter keyboard and telephone (developing press releases, promotional letters, prospect literature, etc.). The argument might be made that this is easy or easier (as well as a necessity) for a small, start-up firm; but a larger firm should keep in mind that fewer interruptions during product development and market planning result in a net gain in productivity. Fifth, the product was priced to sell. As a product increases in cost and complexity, so does the evaluation time and corporate authority level to sign for the purchase. ExecuCalc was (and still is) priced at $5000 per CPU license. Considerations in arriving at this price included: (a) most middle level managers and up (information center manager, DP manager, etc.) could sign-off on the purchase; (b) the product was approximately the price of one micro running VisiCalc or Lotus, yet offered access to everyone with a 3270; and (c) the price was substantially lower than most mainframe products being purchased by data processing, consequently reducing the decision process to buy. Sixth, there was a plan for the future. Once Lotus 1-2-3 overpowered VisiCalc. ExecuCalc was enhanced to include most of the popular features of Lotus. Again, we implemented these features in the same (or as close as possible) syntax as that used by Lotus. Also, a high-resolution, color business graphics product, ExecuPlot tm , was released. ExecuPlot was a look-alike for VisiPlot — meeting all of the criteria noted above. Parallax also provided customer training in ExecuCalc, ExecuPlot, VisiCalc, and Lotus 1-2-3, so that a full service could be offered to a company as the user base grew and requirements for ongoing support and service rapidly increased. Seventh, we realized sales, marketing and promotion can make or break a product. Although a traditional sales approach and marketing strategy are essential, especially for a newly formed company — and in fact we spent heavily on large frequent advertising in Computerworld — there are many variations on the standard theme which greatly helped the success of Parallax with ExecuCalc. Our sales brochure was printed on a ledger sheet (i.e., a spreadsheet) to draw attention to the nature of the product. When we exhibited the product in a trade show, we used promotional gimmicks generally affordable by larger companies (e.g., free T-shirts with our name attached to a newly-popular phrase: The Best Has Always Been Good Enough — Parallax). We also gave a T-shirt to each of our customers who responded to a product questionnaire. We pursued every opportunity with the press to gain recognition for the company and/or the product. This meant writing editors and following up with telephone or personal contact, using every contact we knew in the trades, never turning down an opportunity to be mentioned no matter how insignificant the mention or the publication seemed. We also joined ADAPSO to work with others in support of our industry and to make contacts with other executives in the industry. We always gave the appearance of being a large company with a highly successful product. Since product margins would greatly decline if we were required to make sales calls to sell every copy, we shipped the product to prospective customers for a thirty-day evaluation. In the cases where a demonstration was critical to make the sale, we would demonstrate the product (a two-hour demo) to the prospect at their site only if they would provide us with a two-hour period during which we could demo the product to a group of other companies in the local area. We also offered the product under a monthly lease plan ($500/month) with lease credits toward the perpetual license fee. The above material could relate to almost any new product launch, not necessarily an APL product. What were the advantages and disadvantages of using APL? APL was the chosen language in which to write ExecuCalc and ExecuPlot because it was the language in which we had the most expertise. At the same time it afforded us several advantages: First, we were able to bring the first version of the product to market in a relatively brief period of time. One of our main competitors with a similar product written in assembler took 18 man-months to bring the first release to market. Second, we were able to develop enhancements in a timely fashion, keeping the same order of magnitude of time difference as noted above for the initial release of the product. Third, we were able to respond to and fix problems easily. In some cases we could apply a fix over the telephone with a customer on-line with the product. If the product were in assembler we'd have to apply the fix to a master copy and distribute it en mass (although we ultimately did this after fixing any collection of bugs found in the product). Fourth, with a minor amount of guidance (and proper authorization) from us, a client may enhance the product with formulas specifically related to the company's needs or industry specialization (e.g., actuarial formulas, banking calculations, etc.). This is not possible with other mainframes products, nor is it with Lotus, VisiCalc, or other spreadsheet micro products. Naturally, the company needs an APL terminal and an APL programmer. Beyond that, the process is very easy. Fifth, IBM “promotes” APL and APL products in the information center. As a result, we fit in nicely with this environment. Sixth, we were able to use existing software in APL needed to accomplish some of the tasks required of the product. For example, we used AP124X (the full screen manager) or AP126 (part of GDDM) for ExecuCalc; we used GDDM and GRAPHPAK for ExecuPlot. APL was also a d