Chapter 7 - Data Communications: Market Order 1973-1979
7.9 Modems, Multiplexers and Networks 1976-1978
In the mid-1970’s, the modem and multiplexer markets were revitalized, not stalling at some maximum size as forecast by Frost & Sullivan in 1972. LSI semiconductors and microprocessors made possible new product innovations that in turn enabled customers to construct data communications networks. The evolution from point-to-point data communications to far-flung multi-point, distributed data processing networks, with management and diagnostics, fitted perfectly with the needs of customers installing more and more computers.
In March 1977, the headline in one of the largest information technology weeklies read: “Data Com, Distributed EDP Push Modems Toward $200 M Year.”53 Sales of modems for 1976 were reported as $184 million, far surpassing the $67 million forecast in 1972. Sales of high-speed modems alone reached $91.7 million. Little wonder Milgo and Codex were so successful. The article notes growth for the coming year would come from networking/distributed processing. (For an analysis of the competitive mix see Exhibit 5.2 Datapro Multiplexer and Modem Surveys 1976 and 1980.) Each of the leading firms in data communication responded differently to these market dynamics. Cumulatively the firm decisions and actions constructed a social network, or population, known as Data Communications. Within Data Communications were more tightly connected networks such as modems or multiplexers, and even within these networks there were sub-populations or sectors such as high-speed modems versus dial-up modems.
Intertel
In early 1973, Jerry Holsinger started asking how Intertel, the company he founded in 1970, was going to be successful against competitors such as Codex, his former employer, or even against Milgo. Doubting he could continue to grow the company on OEM sales as originally envisioned, he began asking some large end-users if they had problems that were not being solved. From discussions with companies such as Manufacturers Hanover Trust Company and Eastern Airlines, one rather significant problem emerged. Holsinger remembers:
In the old days of point-to-point, there were skilled people at both ends of the line. If you have a problem, you call the guy up at the other end and he gets out his patch panels and his test equipment and you work on it. In a multi-point network, you have operators at the remote end on terminals. They don’t know nothing from nothing. So you were now operating in an environment where the only technically skilled people were at the computer site, and in the meantime, these were on-line networks where, if it was down, it wasn’t like: ‘Oh well, what the hell, we don’t run our batch today, we’ll run it tomorrow.’ It was like: ‘Hey, they’re not working.’The bank wasn’t doing transactions or the airline wasn’t making reservations, and that’s really how it was. People were starting to understand they’ve got problems.
Working with Gunther Kempun, their contact at Manufacturers Hanover Trust Company, Intertel engineers designed a diagnostic card that would go into their modems; modems that would also have dial back-up capability to restore malfunctioning lines. Holsinger recalls.
As I recall, we sold off xeroxed spec sheets to good old Gunther Kempun, because he was a real pioneer. You know, all this leading edge stuff you sell to pioneers, the guys that want to do the leading edge kind of thing, and they get excited because they are influencing your product, so if you get the right guy, then it is a dynamic interaction. and then he’s really hooked. So in 1973, we sold our first stuff to him.
At first selling network control products proved very difficult. In 1974 network product sales could not offset the loss of a couple of big OEM’s, and sales remained flat with 1973 at $3.4 million. In 1975, network products began having market traction and 1976 and1977 were great years, with sales approaching $20 million. Intertel had established itself as leader in the shift to networking modems. In 1978, they ran an ad that came out just before the TCA trade show challenging customers to ask their Codex, or Milgo or Paradyne salesmen if they could match the features of Intertel modems. Matt Kinney of Milgo remembers thanking Holsinger:
When we got to this show, I made a bee line to go and thank him for legitimizing my company in the network management business, because the leader of the network management business said that we were in it. I chased down Art Carr and the two of us went off and we had a beer together and roared.
By now, however, Holsinger had decided he had had enough. He remembers:
I said, after a lot of soul searching: “I want to go out and get a president, a guy that’s a good marketer, experienced manager and all that good stuff, and let me go back and dream up new things.” And we went through a search; spent about 18 months looking for a guy, brought in a fellow by the name of Sy Rosen, IBM background in marketing, Harvard MBA. On paper, he had run a division of General Instruments, he even had a physics undergraduate degree, which I found to be valuable to make people feel comfortable in the technical world. English majors don’t feel real comfortable with technical people. So this guy on paper was perfect, smooth, polished, the whole thing.
Rosen took over in May 1978 and within nine months Holsinger, still Chairman and CEO, knew he had made a mistake. Holsinger remembers:
It was clear that either he went or the company was going under. So I did my thing and he left and I stepped back in. It was painful because I didn’t really want to go back in at that time.
General DataComm
In 1973, General DataComm began exploring if they could OEM a 201 modem from Intertel. Bob Smith remembers:
For various and sundry reasons the deal didn’t work out. However, they were aware of and had access to our diagnostic card. At the ICA show in New Orleans in 1974, you look at the front panel of their diagnostic card, it looked that same as the diagnostic card we had here. So Intertel’s out there in the marketplace selling diagnostic systems, and the first one that they sold successfully was Manufacturers Hanover Trust Company. We bid that job against them and lost by a big number of dollars and I was later told that they bought that job because they wanted the contract. The history is that Intertel became very successful on Wall Street. They owned the Wall Street community in terms of data communications.
Recognizing the importance of diagnostics from their earliest days, management invested in new products. Yet as is common, engineers move from company to company, often taking ideas with them, or companies incrementally innovate on the success of others. Smith remembers:
Around 1976, there was a guy working here who left and went to Paradyne. He took all of these diagnostic ideas to Paradyne. One trade show, they show up and they’ve got a little tape playing in the back of the booth with all these crazy displays that come up on the screen, and that was the beginning of their Analysis system. Now, some time later, I interviewed a guy who was working at Bell Laboratories. The story as he tells it is: ‘We look out in the world and we see this company, Intertel, selling all these diagnostic systems, and we did a study. We decided that the market really could use a good diagnostic networking modem.’ So they proceeded to develop Dataphone II. It is my conviction that the Paradyne Analysis, AT&T’s Dataphone II, the Intertel DMS and then the competitors like Milgo and Codex, seeing this coming along, all started here at General DataComm.
Paradyne
In 1972, Paradyne management began a process that would transform their company from one struggling to grow sales beyond a few million dollars into one of the world’s leading modem suppliers. They won a contract from Datran to jointly develop a 9600 bps LSI modem – months before Codex started working with Rockwell.54 By the end of 1973, management recognized the need to find a seasoned President. In hiring Robert Wiggins in April 1974, they made a brilliant decision. Wiggins would lead Paradyne for over a decade. With thirteen years of marketing and product development experience with IBM, having worked for a timesharing company, and most recently having headed the Information Systems division of GTE, a division attempting to agglomerate a series of data communication companies and a minicomputer company into a presumed working whole, Wiggins fully understood the importance of a 9600 bps LSI modem. In 1975, Paradyne began shipping modems to Datran and, in early 1976, won a follow-on contract worth $5,2 million. To take advantage of the emerging opportunities, Wiggins hired a seasoned Vice President of Corporate Marketing, Jay Hill, who shared his views as to how to build a successful company. In early August 1976, Hill, who was “captivated” by Wiggins and had been with IBM before building a successful career in sales and marketing at Inforex, a very successful data processing firm, came aboard full of a newcomer’s enthusiasm. Within days he received a rude awakening. Hill remembers:
I didn’t quite understand the significance of the problem when they were running around the halls saying: ‘Oh my God, Datran just filed for bankruptcy. How much inventory do we have? How much do they owe us? It’s been locked up by sheriffs,’ and things like this, but that wasn’t the important part. It was whether we could survive the loss of the top line and the attached profit that went with it. Bob just succinctly came into my office and said: ‘We’ve had a very unfortunate thing happen to us. Our largest customer is gone and you have to make up for it. Get going.’
Hill buried himself in work, recalling:
When I saw all my direct reports I was frightened to death. As a matter of fact I told my wife: ‘Don’t unpack. This may not work. These bozos are from a different world, and I’m scared to death.’ I called on a friend of mine who worked with me at both IBM and Inforex and was just a really solid, good guy. His name is Jerry Kendall, and I said: ‘Jerry, I need one guy I can trust. I’ve been here 30 days, and nobody speaks my language, and I need one person to tell me what the customer is saying so that I can understand it. I’ll treat you well. Please consider coming over.’ It took me a month to get him.55
The ever-composed Wiggins never lost his confidence. He remembers:
We had an LSI technology where we could build a 9600 bit per second modem, I believe, less expensively than anyone else, and we were very aggressive in pricing those products, because technology permitted it, and we needed to gain market share, because we had to have business.
Over the next two years, Hill began building an “IBM-like” salesforce largely on the back of a rapidly growing market for high-speed point-to-point modems.56 Then in late 1977, engineering delivered a new product that allowed Paradyne to sell networks of modems with network management and diagnostics. The Analysis Network Management System catapulted Paradyne into the role of supplier to then emerging need of large corporations to tie together their far-flung computers and peripherals into on-line networks. Hill remembers Analysis as being “the breakthrough and one of the things in Paradyne’s history that really gave it a big lift.” Hill elaborates:
After 1978, we would sell one Analysis to a major insurance company like Hancock, and 100, 200, 300 modems would go out, and at that time, pricing of modems played a very key role in Paradyne’s success. The list price on a 4800 modem back then was $3,000, and we would sell them in any reasonable quantity, at $2,400 per copy and an aggressive bid, back then, and this is ‘78, ‘79, we would just get below $2,000, and our gross margins never on a sale like that, got down to 50%.
Success in modems did not translate to mastering statistical multiplexing however. Wiggins recalls the relationship with Case, the firm Evans of Micom once worked for and whose multiplexer technology had begun with ADS:
What we had tried to do, realizing that it was going to be difficult to fund development programs in every area of technology, we set about trying to forge a working relationship. Case had a good multiplexer technology. They were not selling in the US, and we were furnishing them our modems, so we said what we’ll do is forge an agreement in principal that will be a working relationship so that we’ll keep you competitive in the high-speed modem marketplace, and you keep us competitive in the statistical multiplexing marketplace. That went along well for a while, and we were moving multiplexers pretty well, and then Case decided to come into the US marketplace, and they bought the old Rixon company, and became competitors.
In July 1978, Paradyne became a public company; selling stock for $10.00 per share, raising approximately $5.9 million and having a market valuation of $30 million.
Milgo
Throughout the 1970’s, Milgo remained market leader in modems due to early entrant advantages and not making mistakes. But there was more, as Kinney explains:
I would attribute some of Codex’s growth, and frankly ours too, to the fact that the market was so damn big, and there was so much potential that there was no way that one single organization could grow fast enough to accommodate the entire thing. You could not move fast enough to dominate it. There were literally a few years when Codex’s sales group and our sales group never saw one another at all. They were exploiting one set of opportunities. We were exploiting another set of opportunities. They were strained to the maximum to be able to grow at the rate they were growing, as were we. I don’t think competition as such became a real factor in this industry until the mid ’70s.
But then came Paradyne and market competition. Kinney remembers:
Paradyne considered itself a latecomer into the marketplace and did what most latecomers do, and started slashing, cutting, burning the prices. So that opened up the users’ awareness that they can get equivalent products for a lot less money So they started playing us one against the other as you would expect, and competition became a real factor in the industry.
Milgo was not averse to pressing its own advantages, whether real or perceived. As early as July 1971 Milgo had attempted to forestall competition by filing a patent lawsuit against Rixon Electronics, et al., claiming three of its patents had been infringed upon, including those on its 2400 bps and 4800 bps modem. Proving others could play the same game, in December 1974, Western Electric, a division of AT&T, sued Milgo, claiming ten modem patent infringements.57 After conferring with the Justice Department, that the prior month had filed an antitrust suit against AT&T, Milgo, in February 1975, counter sued Western Electric charging violation of the 1956 Consent Decree involving patent royalties on the manufacture of modems.58
Codex unintentionally added to the confusion after receiving their QAM signal structure patent in 1975.59 Forney remembers:
The most dangerous time in a company’s existence is when it gets issued, a patent. It looks at this thing and says: “This must be worth something,” and you’re very tempted to go out and assert the damn thing. So we got this patent. Milgo clearly had been using the technology. They weren’t the only ones, but what we finally wound up doing, not really knowing what we should do, we wrote Milgo a letter and said: ‘We have this issued patent, and it seems you’ve been using it, and we ought to have some discussions about it.’ Well, what we were told later by people at Milgo is that they got this letter and what made a profound impression on them was that it came by certified mail, and so they took a paranoid view of this letter, that it was a prelude to a lawsuit.
After years of slow motion, legal confrontations shifted into high gear. In January 1976, the court awarded Milgo a victory in its patent suit against Rixon et al.60 Not leaving others long to wonder what they would do, ten days later in February 1976, Milgo sued AT&T, asking for injunctions stopping Western Electric from manufacturing and AT&T from distributing its 4800 bps and 9600 bps modems. Milgo also sued Yellow Freight Systems of Kansas, a user of Codex 9600 bps modems, claiming they had violated its patents. Thirdly, Milgo sued Codex seeking “declaratory relief” since Codex patents violated Milgo’s patents.61 Codex responded almost immediately by filing lawsuits against Milgo in Miami and Boston in March. In reflecting on Milgo’s behavior Forney offers:
I think it was very much an afterthought that they swept us into this litigation. Their concern was almost entirely for AT&T, but secondarily, they had a problem with Codex’s patent on the signal structure. So, it is interesting how these things start. In retrospect, it was a great mistake for Milgo to sue us. They had lots of hooks in AT&T and they hassled them to death. We were, it turned out, a much more formidable foe for them.
Later that same March, the Federal Court sided with Milgo and ordered Western Electric to defend itself against Milgo’s claims of violations under the 1956 Consent Decree before it would hear Western Electric’s suit against Milgo.62 In April, Milgo asked the Federal Court for $44 million in damages it claimed “to have suffered as a result of alleged antitrust violations and unfair competition by Western Electric, AT&T, and the Bell System in the licensing of modem patents and the sale and rental of modems.”63
Then, as if Milgo did not have enough legal concerns, on November 8, 1976, Applied Digital Data Systems (ADDS), a computer terminal manufacturer with sales of $17.8 million, announced a take-over attempt of Milgo, offering a swap of stock valued at roughly $50 million – a $15 million premium over Milgo’s public valuation.64 The much larger Milgo, seen as struggling after a 56% drop in earnings for fiscal year 1975, albeit sales still totaled $40.4 million, rejected the offer two days later claiming it was “not in the interests of the shareholders and does not reflect the long-term prospects of the company.”65 After weeks of jockeying, on December 1, Milgo, in a maneuver to stave off ADDS, announced its intentions to sell 312,000 shares of stock to its UK partner, Racal Electronics. ADDS immediately filed a restraining order to block the sale that the court issued on December 9. Jockeying continued and by February 7, 1977, Racal and ADDS each claimed 45% ownership of Milgo’s stock. By the 18th, Racal appeared to control over 50% of Milgo’s stock, and on the 23rd, ADDS tendered its stock to Racal, ending the hostile action, and giving Racal subsequent ownership of Milgo for $62 million.66 (Milgo’s other legal battles ended with an out-of-court settlement with AT&T and, in June 1981. they lost their lawsuits with Codex.)
Multiplexer firms faced different challenges as they responded to customer needs for networking, competition and their own organizational dynamics.
Timeplex
At the end of 1973, Timeplex became one the last technology companies to go public. Even so, the cash raised did not ensure Timeplex’s success. Botwinick remembers:
The company essentially was insolvent at the end of fiscal year June of ‘76. What had happened was that the auditors had come in and finally figured out that the assets were significantly over-valued. All of the profits were sitting in overstated assets, and there was no management. It was a simple as that. Anyway, they struggled through 1976 with Allen & Co. pounding on Chemical Bank to keep the doors open.
In order to keep the bank happy, an effort was made to sell the company. Codex came to look the company over, as did Milgo. Neither saw enough value to make a bid. However, neither Timeplex’s poor financial performance nor lack of interest on the part of potential buyers prevented a power struggle for corporate control from developing between the investment bankers, Allen & Co., company management and eventually a group led by Botwinick. By the spring of 1977, Botwinick had exhausted his financial resources in buying enough shares of Timeplex’ public stock to end the stand-off. On June 27, 1977, he became Chairman of the Board and Chief Executive Officer. Botwinick remembers:
The second week I was here, I went out to lunch with the guys from Chemical and United Jersey banks, and at the end of the lunch they said to me: “This is all very interesting. You’re the first guy we’ve ever talked to in this company that seemed to know what he’s talking about, but we’ve had it. You have 60 days to tell us how we’re going to get our money back or else.” I don’t know what the ‘or else ‘ would have been, because there was nothing there if they wanted to shut it down. So I didn’t take that terribly seriously. On the other hand, I was in imminent danger of going broke personally.
Botwinick immediately began aggressively collecting over due receivables to raise desperately needed cash. In doing so, he learned the company had shipped defective product in order to record sales, defective product that would need to be replaced. He then began selling whatever assets had any value. Combined, Botwinick recovered enough cash to pay the banks and create some breathing room. He then took a hard look at the recently introduced Dynaplexer, a statistical multiplexer hub that worked with its TDMs but had no terminal or computer ports. While offering their large installed base the option of upgrading to the latest multiplexer technology, Botwinick remembers:
I got a close look at the DynaPlexer and I realized that it was conceptually flawed. First of all, they had no capabilities to develop it. It was too complex. They had no software skills. They didn’t understand microprocessor software development – it was an old icon of a hardware company. Second was that the concept was flawed, because the old TDMs that were in the remote sites were all hard wired. There was no way to remote configure them. It wasn’t going to sell. Also, they were never going to get it to work. We pushed the delivery schedule out considerably. We said we were redesigning the product, putting in all kinds of hot new features.
Realizing the Dynaplexer was not salable, and might never be, Botwinick turned his attention to developing a microprocessor-based stand-alone statistical multiplexer. Investing long hours with a focus only desperation can inspire, Botwinick and his new engineers innovated their Series I Microplexers that was introduced in 1978. Then in one of those strokes of good fortune, Western Union, literally up the road, decided to end their efforts to develop a statistical multiplexer to compete with Codex. Botwinick leaped on the opportunity and hired a trained team that immediately started developing a product to replace the feature-weak Series I Microplexer. The Series II Microplexer was introduced in 1979. Botwinick remembers:
Series I came to market and it held the doors open. In fact, it stimulated sales of our old TDMs, which were still the bulk of our business. At that point, we did get some people from Western Union because they had gone out of the StatMux business. They came in and did the Series II. Now, the Series II was the first networking unit that we did, and that’s what we built the company on.
Infotron
When James Hahn, founder and President of Infotron, saw Codex’s Series 600 statistical multiplexer literature in early 1976, he instantly knew they needed to develop one as well. To do so meant hiring an engineering manager with microprocessor experience, because no one on staff knew the least bit about these miniature computers. In discussing where he might find someone with his wife, Leigh, who still worked for Ultronic Systems, the company Hahn had left to found Infotron, she recommended they interview Bill Dambrackas, a young engineer who had just developed a specialized terminal to display stock quotes for brokerage firms using Motorola 6800 microprocessors. Hahn’s overture couldn’t have been better timed. GTE, under the leadership of Wiggins, was consolidating Ultronic Systems and a number of other companies to form an Information Systems division to be homed in the facilities of the minicomputer company they had bought in California. Dambrackas did not want to relocate. The idea of working just a few blocks down the street, with the challenge of managing a high profile statistical mulitplexer project, made for an easy decision. Dambrackas joined Infotron in March charged with having product to ship by year-end.
Dambrackas knew he did not have the luxury of generating an entirely new design and that he had to use as much existing TDM circuitry and packaging as possible. If their existing TDM’s were field upgradable into statistical multiplexers, Infotron might just pull the carpet out from under Codex’s grand design. The marching orders: keep it simple and get it done on time.
Dambrackas strengthened the Infotron team with some hires from Ultronic Systems and set to work. In October they announced the Supermux 780, for a base price of roughly $6,000, subject to having Infotron’s TDMs. Arguably no where near the product of Codex’s 6030, it nevertheless gave an interested buyer a much cheaper alternative with which to test this new technology. Dambrackas remembers:
While Codex did a revolutionary design, where they had a whole fresh start on the whole box, they had a lot of headaches and a lot of start-up problems. So even though they announced the product, we ended up bringing our product to market and beating Codex pretty good. And it was a big hit for Infotron.
Indicative of the impact, over the next three years, over 3,000 Supermuxes were sold compared to 950 Codex’s 6030s and 6040s.67 In dollar terms, however, Codex’s sales exceeded Infotron’s. Dambrackas comments on what happened:
Codex quickly made a 6010 which moved down into the lower price range. It was cheaper and simpler. But they started off, I guess, saying that statistical multiplexers would be sold like a very special thing, and they didn’t realize it would drop down into the dirt so quick and end up being the bricks and mortar of getting data around. You know, it’s funny how these things start off to be very exquisite, special things, and then they drop down to being something that people throw around like modems.
Under Dambrackas’s leadership Infotron would introduce a range of statistical multiplexers over the next few years. Then in early 1979, headhunters called on him with another opportunity. In June 1979, Dambrackas would leave Infotron to head Milgo’s efforts to develop their own statistical multiplexer as Director of Data Network Products.
- [53]:
“Data Com, Distributed EDP Push Modems Toward $200 M Year,” Electronic News, March 14, 1977, p. 1
- [54]:
“Datran LSI Modem Contract,” Mini-Micro Systems, May 1976, p.29
- [55]:
Kendall would later become President of Paradyne.
- [56]:
Paradyne had a second modem-type product named PIX that accounted for roughly 50% of their sales.
- [57]:
“Milgo Charges Violation of Court Ruling by WE,” Electronic News, Feb., 3, 1975, p. ; the Western Electric suit was probably prompted by the fact that AT&T had OEMed Milgo’s 4800 bps modem before coming out with its own.
- [58]:
“Milgo Confers With Justice.” Electronic News, Jan, 6, 1975, p.
- [59]:
The patent was for its Double Side Band-Quadrature Carrier modulation.
- [60]:
“Milgo Wins Patent Suit Covering Data Modems,” Electronic News, Jan. 26, 1976, p. 2
- [61]:
“Milgo Sues AT&T, Users on Modems,” Electronic News, Feb. 2, 1976, pp.1 and 16
- [62]:
“WE to Go on Trial on Milgo Patent Charges,” Electronic News, March 29, 1976, p. 2e.
- [63]:
“Milgo Seeks $44M in Suit Against WE/Bell System,” Electronic News, April 12, 1976, p. 1 and 14
- [64]:
“Applied Digital Sets A Stock-Swap Offer for Milgo Electronic,” The Wall Street Journal, Mon. Nov. 8, 1976. The data to follow all comes from the Wall Street Journal articles over the period of the transaction.
- [65]:
WSJ, Nov. 10, 1976
- [66]:
“Applied Digital Tenders Its 47.8% Stake in Milgo to UK’s Racal for $29.7 Million,” Wall Street Journal, Feb., 23, 1977, p. 8. (Authors note: Based on prior fiscal year revenue and net income, Codex sold for 2.9 times sales and 24 times net income and Milgo sold for 1.5 times sales and 26 times net income.
- [67]:
Datapro, “Multiplexers - Basic Characteristics and Equipment Specifications, October 1979. (In dollars, the author estimated comparison might be: Infotron - $18 million; Codex - $22 million.)