Historical notes on the application software business. Related subjects include:
I recently wrote a long post on the premise that enterprise analytic applications are not like the other (operational) kind. That begs the question(s): What are operational enterprise applications like?
Historically, the essence of enterprise applications has been data management — they capture business information, then show it to you. User interfaces are typically straightforward in the UI technology of the era — forms, reports, menus, and the like. The hard part of building enterprise applications is getting the data structures right. That was all true in the 1970s; it’s all still true today.
Indeed, for many years, the essence of an application software acquisition was the database design. Maintenance streams were often unimportant; code would get thrown out and rewritten. But the application’s specific database structure would be adapted into an extension to the acquirer’s own.
Examples that come to mind from the pre-relational era include: Read more
|Categories: Application software, Cullinet, McCormack & Dodge, MSA, Pre-relational era, SAP||1 Comment|
In the process of researching my recent post on Management Horizons Data Systems, I came across an excerpt from a 1972 marketing brochure (quoted in the “History of Management Horizons” piece cited there). General notes include:
- The brochure quote basically pitches business intelligence/ performance management, with unconstrained drilldown.
- Pitching BI/analytics benefits for what start out as being transactional applications has been going on for pretty much the whole history of the applications industry. This just happens to be a great proof point.
- The unconstrained drilldown part could almost be taken for granted today, in the relational era. In 1972, however, it was a rather bold (and for all I know exaggerated) claim.
The exact verbiage is: Read more
I started drafting this post along with others around the time of my parents’ deaths, then put it aside. However, I have been informed that my father’s old colleague Alton Doody has cancer himself, and if we are ever to get his input, it would be best to solicit it REALLY SOON. So I’m finishing this up now as best I can.
Here’s the part I know from my own memories as
- The son of a Management Horizons employee (namely my Dad).
- A software industry stock analyst (in particular, one who followed Informatics General).
My father moved to the Columbus area in 1973 to join Management Horizons, a consulting firm serving retailers. Management Horizons had its own spin-out already, a time-sharing company called Management Horizons Data Services (MHDS), with which it still shared a building on what is now Old Henderson Road in Upper Arlington. And, this being a world full of coincidences, MHDS is very on-topic for the primary focus of this blog (software industry history).
MHDS’ main business was a full suite of what we might now call ERP for distributors and/or retailers. That never amounted to much. But its secondary business was an electronic interchange for direct placement of orders, called Ordernet. Ordernet turned into Sterling Commerce, a > $1/2 billion company that has been acquired for >$1 billion more than once.
The chain of events, roughly, is: Read more
|Categories: Application software, Companies and products, Computer services, Pre-relational era||8 Comments|
The approach of April Fool’s Day has me thinking of software industry pranks and other hijinks. Most of what comes to mind is verbal jousting of various sorts that doesn’t really fit the theme. But there was one case in which ongoing business competition got pretty prankish: mainframe-era accounting software leaders MSA vs. McCormack & Dodge. Read more
Wikipedia’s current article on Cullinet is long, detail-laden, and slanted. The difficulties are not of the sort to be fixed with my usual pinpoint Wikipedia edits. So I’ll just reproduce it here, commenting as I go. As for copyright — this particular post is as GPLed as it needs to be to comply with Wikipedia’s copyleft rules. All other rights remain reserved.
The company was originally started by John Cullinane and Larry English in 1968 as Cullinane Corporation. Their idea was to sell pre-packaged software to mainframe users, which was at that time a new concept in an era when enterprises only used internally developed applications or the software that came bundled with the hardware.
|Categories: Application software, Companies and products, Computer Associates, Cullinet, Database management systems, Industry sectors, Pre-relational era||7 Comments|
Computerworld got software industry history a bit wrong by implying that John Cullinane innovated packaged software (specifically, they said “packaged application”). Here’s what really happened, as I learned soon after becoming an analyst in the early 1980s:
- Most early packaged software companies were hybrids, offering both packaged products and professional services (including services unrelated to the packaged products).
- Applied Data Research, led by Martin “Marty” Goetz, is the clear innovator in third-party packaged software. Not only is ADR’s Autoflow the generally acknowledged first packaged software product from an independent company (“independent” as opposed to, say, IBM), but ADR was a leader in legal and political anti-trust action to gain market space to sell against IBM.
- If you use the term “application” narrowly — so that anything whose main function was to help manage IT shops and activities is “system software” rather than “application” — there’s no way Cullinane was an early leader. Think instead of American Software, MSA, McCormack & Dodge, or several specialists in regulated verticals such as banking and insurance. But if you use the term “application” loosely, ADR gets priority as noted above.
- The credit Cullinane usually gets for leading the way in software company success (e.g., first IPO of a product company) is absolutely justified.
|Categories: Application software, Applied Data Research, Cullinet, IBM, McCormack & Dodge, MSA, Pre-relational era, System software||6 Comments|
MSA (Management Science America). This section got so long I’m breaking it out as a separate post just about MSA.
M & D (McCormack & Dodge). M & D was MSA’s archrival in mainframe financial software. They had various claims to product superiority, based on having “more CPAs on staff” than MSA and also on being first to market with realtime applications. However, M & D sold out early to Dun & Bradstreet, and lost its edge as key managers left.
M & D seems to have been a lively company. Many stories about drugs or sex emerged (I don’t actually recall any drugs-and-sex-combined stories, for whatever reasons). Key players included: Frank Dodge, a former schoolteacher who founded another not terribly successful apps company (The Dodge Group) afterwards; Jim McCormack, who happily retired from software into real estate, but sadly died a few years later; development chief John Landry, who’s been a prominent industry figure ever since, and sales/marketing chief Bob Weiler, ditto. Landry and Weiler went together to Distribution Management Systems, Cullinet (after it bought DMS), and Lotus, before going their separate ways.
M & D’s venture into manufacturing applications seemed later and more half-hearted than Comserv’s or Cullinet’s. But they eventually did wind up with a version of (and may even have bought control of) the Rath & Strong technology.
Cullinet. Cullinet was better known as a DBMS vendor. But in a precursor of what became the Oracle strategy, it pursued financial and manufacturing applications as well. The financial applications were originally licensed from M & D. The manufacturing apps were originally licensed from Rath & Strong, as were M & D’s.
One negative consequence was that the industry teamed up against Cullinet. For example, ADR in DBMS and MSA in apps formed a close marketing relationship. To general industry agreement at the time, I dubbed this the ABC (Anybody But Cullinet) strategy.
Cincom. DBMS vendor Cincom pursued a Cullinet-like apps strategy. Not many people cared.
J. D. Edwards. If I recall correctly, JDE’s main platform was the IBM System 38, the predecessor to the AS/400. Anyhow, JDE was a Denver-based financial software company. Its main claim to fame, other than the platform that it ran on, was a superb order entry system. Rob Kelley, referring to his days at Arthur Andersen, once told me that Andersen’s order entry system had had 45,000 lines of code, JDE’s had had 5,000 lines, and JDE’s had been better.
SAP. I’ve already written up what I recall about SAP in the 1980s.
This initial list leaves a lot of companies out, of course. Other than the MRP companies — ASK, NCA, XCS, and so on — the biggest omission may be Walker Interactive. But also missing are Global Software, Data Design, a whole lot of human resources specialists and so on.
Also missing are other vertical market groups, most notably in banking software, which is where general ledger products (the first major financial application) first succeeded in a big way.
I hope to get around to writing about these subjects before too long.
When I became a software analyst in 1981, MSA (Management Science America) was generally regarded as the leading cross-industry financial software vendor. Its CEO was the colorful John Imlay, best known for a variety of showman stunts, such as bringing animals to sales meetings. (He also was known as “the man who killed the keypunch” from his hardware days, when he took a sledgehammer on stage to a keypunch machine in a presentation introducing key-to-disk technology.) The president was Bill Graves, the most agile 300 poundish guy I’ve ever seen off of a football field, and still the only person at whose house I’ve held hands during the saying of Grace.
MSA software ran only on IBM mainframes. There were a limited number of modules. I specifically recall an ad campaign for the “Big Eight,” because they had eight modules, and the “Big Eight” were the public accounting firms in those days. The eight included payroll, human resources, and six financial modules, which were general ledger, accounts payable, accounts receivable, purchasing, fixed assets, and probably inventory. That’s all, versus the hundreds of modules successor companies have today.
MSA obviously modeled its “persona” on IBM. Indeed, the MSA logo consisted of the three letters in a font that consisted of thin parallel horizontal lines, exactly like IBM’s of that day did. Another major slogan was “People are the key,” with little key lapel pins given to five- and ten-year employees. Read more
Until the past couple of years, I didn’t have a lot of dealings with SAP. (That has now changed significantly.) But it seems that the things I do recall aren’t that widely known anymore.
I first heard of SAP in the 1980s. It was a smaller company than the then-leading mainframe application software vendors. Peter Zencke told me earlier this week that when he joined in 1983, the company had around 100 employees. From memory about MSA’s figures, I’d guess SAP’s revenue was somewhere in the $150-250,000 per employee range. Also from memory, I’d guess that MSA and M&D (McCormack & Dodge) were meaningfully bigger than SAP at that time. I also think that SAP combined financial and manufacturing applications earlier than the other mainframe vendors did, and hence probably got more revenue per client from a small number of clients. (MSA didn’t get into manufacturing apps until they bought Comserv, which if I recall correctly never broke the $20 million revenue mark on its own.)
SAP was almost unique among significant software vendors in being based outside the US, Software AG being the other obvious big example. There was no Business Objects then, of course. I don’t think that any of the UK companies that eventually made a modest impact — MicroFocus, LBMS, and much more recently Autonomy — were even active then. So it was pretty much off of people’s radar screens …
Indeed, at one point in the early 1990s I wrote to the effect of “Hey! There really are some important European software companies!” And spurred by that, my clients at Fidelity Investments invested in SAP. Too bad they were perennially stingy about compensation for good investment ideas …
Anyhow, the word on SAP from its competitors was that in the US at least, SAP focused tremendous sales effort on a small number of prospects, and in those accounts they were very hard to beat. These accounts seemed to be centered on the chemical and pharmaceutical industries, presumably because those industries were particularly strong in SAP’s home German market. Not coincidentally, SAP’s US operations were headquartered in Pennsylvania, near the New Jersey stronghold of those industries in the US. It’s natural to conjecture that SAP had superior functionality for process manufacturing industries, something that was pretty primitive in those days, but I don’t recall any direct mentions of this.
I learned more in the early 1990s when Jeremy Coote called up and introduced himself. He was the CFO of SAP’s US operations (he later went on to a big job at Siebel). It turned out that SAP had some contractual reason only to invest limited resources in the US. But that would change soon; one of the directors was coming over to run things in the US personally; and so on. Obviously, they lived up to that much more than I could possibly have envisioned at the time.
The story of how SAP’s rise dovetailed with the growth of the public accounting firms’ consulting practices is better known; I’ll leave the telling of that to another time.
The news about Ingres being spun off by Computer Associates brings back a lot of memories. First of all, Ingres (then called Relational Technology Inc.) was one of the centerpieces of my first-ever research trip to the West Coast in April, 1982. Second, the day CA’s acquisition of Ingres closed, Charles Wang (CA’s CEO, of course), called me personally and asked me to consult to CA about their forthcoming product strategy. It was an intense, month-long project, perhaps still the single largest one I’ve ever done.
So with no further ado, here some observations of and about Ingres through the years.
- Ingres was of course the first of several DBMS companies spun off from UC Berkeley’s INGRES research project, and one of several started with Mike Stonebraker’s involvement. I wrote about that history briefly in my now-defunct Computerworld blog.
- Ingres (then called RTI) and Oracle (then called RSI, for Relational Software Inc.) were of course arch-rivals. As a general rule, Ingres was first to market with new features such as a 4GL or a truly distributed DBMS. Oracle, however, was the first to market with the features customers most cared about, at a level of completeness they found acceptable. Eventually, when Sybase was a factor too, Ingres was always betwixt and between — everybody’s second choice, but not the first choice of enough buyers to keep on prospering. (Later on in the 1990s, Gupta took over the Ingres role in the low-end market — the product was broader than Powersoft, but who cared?)
- Ingres was eventually merged into ASK Computer Systems. While surely a distraction, that’s not what killed it. Each predecessor company had its own problems, and they pretty much stayed out of each other’s way, at least in product strategy. What killed them is that neither side of the business managed to stay fully competitive in product.
- Ingres’s fatal technological mistake was whiffing on parallelism. And it did so in the most painful of ways. Ingres had a joint development project going in the Portland, OR area with Sequent, to develop a parallelized version of their DBMS. They pulled out due to expense, and Informix stepped in. And that’s how Informix managed to be competitive with Oracle in parallel processing, while lack of competitiveness in that area is what doomed Sybase and Ingres. Ouch!!!
- A second Ingres failing probably wasn’t as big as I thought at the time. This was an inability to offer abstract datatypes, aka object/relational, aka UDBMS (where the “U” is for “universal”). I thought this feature would be hugely important, and my opinion on that score probably was a big part of influencing Informix to overpay for Illustra. But Microsoft has never had the feature, and it doesn’t seem to have suffered all that much in the marketplace for its lack.
- ASK was doing even worse on the product side than Ingres — it never came out with a decent GUI version of the product, although ASK did get a license to resell Baan’s code — and the whole sorry mess was eventually sold to CA. CA has a well-deserved reputation for slashing development costs and profiting from slowly-dying software products. But I watched this acqusition from the inside, and to this day I think they really wanted to make the product competitive. But there was one not-so-little problem …
- … CA ran off all of Ingres’s engineers right after the acquisition. CA’s policy upon acquiring companies was requiring employees who wanted to keep their jobs to sign non-compete agreements. In Ingres’s case, however, that policy was a spectacular failure. Oracle, Informix, Sybase, and much of IBM’s DBMS development were all located in the Bay Area. Finding another local job for these guys (and gals) was EASY. Competitors went into a feeding frenzy hiring Ingres engineers, and there was essentially NOBODY left. In my judgment there was a reasonable chance CA could revitalize development with an aggressive investment strategy, but they ultimately blinked. And with very limited ongoing development, the product obviously faded quickly as a mainstream competitor.
I think I’ll go write about the rest of the story over in the DBMS 2 blog.