What Happened to CompuServe?

A friend of mine on a consulting assignment asked me about the sequence of events which transformed CompuServe from a subsidiary of H&R Block to a part of both AOL and Worldcom. I figured it was worth posting in my blog...

The process started with HRB coming to the conclusion that they wanted to sell their 80% interest in CompuServe (soon after the 20% was sold on the open market). Through their investment bankers, they shopped it to a number of potential buyers, including AOL, LBO specialists and even AT&T. None of the offers were acceptable to HRB, as most required HRB to take a stock they didn't trust as currency (e.g. AOL), or required seller debt financing, or the price was just not what they were looking for.

Eventually John Sidgmore of Worldcom figured out a structure that made everyone happy (we had known Sidgmore for a number of years, from the days when he joined UUNET, whose first network was an IP-over-X.25 implementation running over our network).

At the time UUNET was the major dialup network provider to both AOL and MSN, and Sidgmore knew Steve Case well. The imaginative deal he proposed was accepted by all parties, and was executed like this:

Step 1: WCOM bought all of the outstanding shares of CSRV using WCOM stock as the currency. The value of the WCOM stock was approx $1.2 billion, and HRB sold all of it within a day or two, pocketing the cash (Had HRB held the stock for another 18 months, it would have turned into $3.5 billion. But then if they had held it for 3 years, it would have been worthless...) At this point, HRB is completely out of CompuServe, and Worldcom owns 100% of CompuServe.

Step 2: WCOM sells CompuServe Interactive Services and various pieces of the intellectual property and infrastructure to AOL. WCOM actually has to throw some cash into the deal -- about $100 million as I recall -- but gets back a long-term dial services agreement from AOL for zillions of hours. AOL gives their network subsidiary, ANS, to WCOM in this transaction. WCOM and AOL agree to a complex infrastructure cooperation agreement which defines who owns what (e.g. AOL owned the CompuServe mainframes, but WCOM owned the source code and use licenses for the CompuServe operating system -- my idea by the way, to make sure neither party could hold the other hostage). At this point, CompuServe Network Services (renamed Worldcom Advanced Networks) and UUNET are sister companies within Worldcom, both reporting to Sidgmore. ANS became part of UUNET.

Step 3: However, during one their acquisitions, WCOM had picked up a little IP networking company called GridNet, based in Atlanta. Management of GridNet was assigned to WAN (the former CNS). I found out that GridNet had been given responsibility for managing another little network company in Memphis, and I don't even remember their name, but great folks. I made the recommendation that this company be shut down, but instead it was absorbed into a group managed by Robert Hudspeth, who I believe had been responsible for Memphis guys in the company that WCOM had acquired to bring the Memphis guys into WCOM (beginning to understand what a complex world WCOM was?). Anyway, it was at this point that we began using the GridNet technology to consolidate dial points on the old CompuServe network. We made it so that you could call their modems using a CNS phone number, and by reading the DNIS (Caller ID), their modem system would route the call to a "reverse gateway" which would present the CompuServe LOGOUT interface.

Step 4: Sidgmore decided to take a Darwinian approach to determining whether CNS should continue to survive: he directed us to give our entire customer list to UUNET Sales, and they immediately began going to our customers and cherry picking the folks whose usage profiles matched UUNET technology (IP orientation, coverage matchup with the UUNET network, no fancy billing). The trend was clear, UUNET was going to drive the price of dialup network down to a point where CNS would be losing money, because we were still saddled with about $250 million/year in non-WCOM telephone expense, primarily via AT&T and MCI. Ultimately, the CNS management team decides that the only sensible thing to do was roll in under UUNET and stop the war. The CNS leadership reported to Mark Spagnolo, the UUNET CEO. The CNS integration with GridNet ceases, and GridNet is shut down.

Step 5: WCOM buys MCI. The SEC and FTC require MCI to divest their IP network, but interestingly allows them to keep Tymnet. Vint Cerf, to whom Tymnet reported, initiated conversations with us about merging the CNS and Tymnet networks. What we found was that the Tymnet network was a generation of technology behind the CNS network, and that besides, the CNS network had features DESIGNED to allow users to move from Tymnet to CNS with a minimum of hassle, but not visa versa. That conversion was just starting when I left.

Step 6. Meanwhile, the technology we developed to facilitate reaching the CNS network over GridNet access was applied to begin moving the CNS network to UUNET. This conversion is still going on -- there are still CompuServe nodes running in the middle of the MCI network. One of the reasons is the POS authorization network service. It is my understanding that MCI now provides 100% of the dialup POS authorization service in this country (one of the significant competitors was Tymnet!).

Step 7: WCOM self-destructs. Bernie is fired, Sidgmore dies, Spagnolo leaves, and the former MCI executive team rises to the top (other than Mike Cappellas, who is hired after Compaq is acquired by HP). UUNET disappears as an organization, and ultimately as a brand as what had been competing and redundant internal divisions (e.g. both MCI and WCOM had substantial technology headquarters) were reorganized under an MCI-like structure with MCI leadership.

Gradually, the old CompuServe team is disbanding and moving on to a next life. It was the ride of a lifetime.

The Columbus Dispatch did a story on Sept 6, 2009 about the history of CompuServe. It has been captured here if you wish you read it.