Word came recently, via one of those catch-up obituaries in The New York Times, of the passing of Bob Allen, who had been the chairman and CEO of AT&T during a time of merciless transition in the telecom industry during the 1980s and 1990s.
I have a reminiscence to offer, not as a criticism of Allen, who by all accounts was a decent man and strove mightily to reinvent AT&T in the wake of the forced 1984 divestiture of the regional operating companies, but to underline an example of “innovator’s dilemma” with which I had direct experience.
In the early 1990s I was serving as interim executive director of the then-New Jersey Commission on Science and Technology, a public agency charged with spurring science- and technology-based economic development in the Garden State (which in our propaganda we were pleased to redub “The Invention State.”)
Our commission was a public-private board, led by prominent citizens with sci/tech experience, appointed by the state’s leading politicos. Our board chair was a retired R&D executive of Exxon (yes, the same unit now in the news over its global warming research in the ’80s), and our vice chair was a retired director of AT&T Bell Labs, perhaps even then still the most prestigious corporate R&D lab in the nation, though it fell subsequently into relative obscurity.
Alongside the National Science Foundation, our commission had invested state funds in one of the first five national supercomputer centers, this one based in Princeton. We also partnered with NSF on of one of the world’s first non-military, high-speed, Internet Protocol (IP)-based data networks designed to connect that center with its users at academic institutions throughout the Northeast. That network, called the JvNCNet and originally hosted by Princeton University though it was soon spun off, would go on to become one of the first regional feeders to the NSFNet, which in turn became the backbone of the commercial Internet. (For some additional history, see a nice lecture given at Stevens Institute of Technology by the network creator, Sergio Heker, now the CEO of GLESEC, an Internet security firm.)
At that time, AT&T was struggling, as the Allen obituary indicates, to come to terms with its new role as a provider of “long-distance services,” competing with upstarts like MCI and Sprint, and severed from the revenue and profit of the regulated regional Bell operating companies (RBOCs) that connected local users to the telephone network. AT&T’s two main assets were its status as the default provider of long-distance services to local phone customers, who had to chose actively to switch to a lower-cost competitor, and its ownership of lots and lots of physical inter-exchange “trunk” capacity, much of it having been upgraded from copper to fiber. While the RBOCs were carrying analog signal over copper to their old-fashioned, electro-mechanical switches, AT&T was already happily moving those conversations cross country in bits via fiber.
Even as early as 1991, it was obvious to any informed observer that the long-distance cash cow could not be maintained forever, and to close observers of telecommunications, that AT&T was already in the business of moving voice conversations by bits! Could they not see that bits are bits, and that the Internet would evolve into voice protocols that would eat their source of profit, unless they leveraged those physical assets of high-capacity fiber?
Our friends at the JvNCNet were unable to engage AT&T as a partner or even a vendor, even though the latter was a leading corporate citizen of the state and had a direct strategic interest in the rapidly evolving Internet. (Despite the AT&T tower in Manhattan, AT&T was very much a New Jersey entity, with both Bell Labs and corporate offices then based in Union County, and the very heritage of telephony and telegraphy tracing to the Vail family of Morristown.)
So, as good, aggressive agency staff are wont to do, we put our principals up to a set-up job: I drafted a letter for our chairman (remember: a former Exxon executive) to send to Bob Allen, placing the issue squarely on the table, and we copied our vice chair (former Bell Labs head William O. Baker), and for good measure I sent a covering note to then-current leadership at Bell Labs. Recently I recalled this correspondence, regretting that I had not taken a copy with me when I left state service. Luckily, Dr. Baker’s very able assistant at Bell Labs kept his copies on file, so while the New Jersey state archives had not held a copy, everything ended up in his papers, which he donated to his graduate alma mater, Princeton University. So the accompanying image was retrieved by and is used thanks to the courtesy of the William O. Baker Papers, Public Policy Papers, Department of Rare Books and Special Collections, Princeton University Library.
You’d think that with all these corporate connections, some dialogue would have ensued. But it never did. I would love to know what transpired as that high-level correspondence kicked around Bob Allen’s office, but I’m assuming that leadership of AT&T simply froze like deer in headlights at the prospect that a “free” service like the Internet would eat their cash cow. They could not mobilize themselves to coöpt these developments, and watched MCI and Michigan’s MERIT network walk away with nearly all the benefit of federal investment (much shepherded by Al Gore, to be sure.)
And of course that’s pretty much what happened. While not everyone yet uses voice over IP (VOIP) explicitly, interexchange fiber carrying digital signals drove the marginal cost of a minute of telephone time nearly to zero and erased the distinction between local and long-distance calls. And now many regional phone companies are transitioning their remaining landline customers (wireless is a story for another day) to fiber, bringing us ever closer, with fewer and fewer gaps, to a voice network that is defined less by circuit switching and increasingly by packet switching exactly as in data networking, just as we all foresaw.
The bottom line is that even faced with an obvious strategic threat, and one that was explicitly pointed out to them by folks like us who didn’t know half as much as they did internally, the old AT&T could not innovate its way to salvation and ended up someone else’s lunch. The modern AT&T is really a rebranding of Southwest Bell, which acquired the old AT&T a decade ago in part for that national network of high-capacity fiber trunks, which turn out to be useful even in the circuit-switched wireless world for “backhaul“; in part for the wireless assets AT&T had acquired in the interim; and in part for that grand name.
UPDATE/CODA – Replying to an early draft of this post, my friend Sergio Heker (see above) adds this recollection and gives me permission to post:
“Bob Allen did assign John Petrillo (VP) to work with me on evaluating their involvement with the Internet and I did spend six months going back and forth from Princeton to AT&T headquarters to work with Wood Kerkeslager who was assigned by John Petrillo to work with me. There was much exchange of information and the conclusion of this from my recollection was that AT&T did not see the Internet as a business of interest to them… I don’t know exactly what was their decision making process at the time but it was clear to me that they missed an incredible opportunity of buying JvNCnet which was one of the biggest ISPs and the largest international Internet operator.”