Please accept my resignation. I don't care to belong to any club that will have me as a member.

Groucho Marx via the Groucho Letters, New York Simon & Schuster 1967.

Club Internet

⊕
The Club Internet logo, in glorious Gifnicolour.

Internet Éireann had come and gone, but there were a lot of people, myself included, who owed their careers, however indirectly, to Steve O’ Hara-Smith and his vision.For example, the founders of directski.com, of whom more at some later stage. My first taste of the harsh reality of startup life came from my experiences at Club Internet, founded out of the ashes of Mary O’Hara-Smith’s celestial intransigence. As noted above, Tom Kelly and Stuart Fogarty started it with a bunch of folks from the Internet Éireann usergroup, and the largest resemblance it bore to Internet Éireann was the almost hilarious lack of cash.Club Internet was incorporated with £30,000 of share capital on the 12th of June 1995. Stuart Fogarty’s background might have suggested otherwise, given his control over the well-known and respected family advertising business Aubrey Fogarty Associates, but it was, at least from my point of view, another righteously penny-pinching startup.

⊕
The rendered logo, in glorious Gifnicolour.

I was drafted in by Alex French as general technical dogsbody, systems administrator, network operator, and phone support provider.Similar to my experience in the Irish Domain Registry later on, it was clear that I was not intended to be one of nature’s customer service agents. One of my abiding memories is having to explain how to move a window around to a customer who had called us in shock when the Netscape browser “had its bottom half chopped off by the screen.” Well, at least no-one got hurt. Alex had crucial and rare experience in this realm: not only had he worked for Internet Éireann and brokered the failed meeting between Tom Kelly and Mary, but he’d had a lot of other experience too. For example, he had founded a web design company called Virtual Impact with a guy called Donogh O’ Donovan, an unusual thing for the time:

It was the first dedicated web design company in the country. There were lots of little designers, for example graphic designers, who did web pages. There were a few ad agencies who did web pages. But I think we were the only company who did only web design. Remember that at the time it was an engineering feat to create a web page – it was programming! It wasn’t considered a design thing or, god forbid, a UI thing. It was an engineering thing: Can you make web pages? […] I made a deal with Steve where we could use a room, rent free, and in return we would get people to host stuff with them and basically pimp their service to people who were going to be our customers. So we ran a co-ax cable down the corridor and drilled a hole in the wall to connect it into the main server that gave us internet connectivity. Again at the time, having a permanent internet connection to a backboneTechnically speaking, Internet Éireann was a leaf node and not a backbone, but a permanent Internet connection was still a very rare thing. was a fairly good feat. After an exhausting first big customer called Monaghan Mouldings, burned in my memory because we sold them on the idea that we could do them an online brochure, gave them a price for it and then it turned out that they had a hundred page brochure of different ceiling mouldings which we had to scan in manually. Anyway, we did that and then about six months later, because of college pressures really, Donogh and I decided to go our separate ways. I failed some exams.A common theme in early Irish Internet stories. The Irish Internet of the future was mostly built in the college access labs of the past. I think he did as well.

Many people who know Alex would say that his negotiation skills are second to none – if I can put it that way – and his achievements here were an early demonstration of the persuasiveness he was to bring to bear on Club Internet’s behalf. Having convinced Tom that Club Internet could be built rather than outsourcing it to Internet Éireann, Alex also applied those famous rhetorical techniques to two people he knew: myself and Ross Chandler. Alex recalls:

Ross had just graduated from Theoretical Physics. He had just finished up so this must have been the summer of ’95. He had applied for a research post in the UK, a PhD perhaps, but it hadn’t come through so he was at a loose end. He had actually been doing a little bit of tech work in Internet Éireann as well, but was mostly known for his cinema guide at the time, a manually maintained list of cinema showings which people used to figure out what to see. So I brought you to my mother’s house and talked to you guys and basically said, “Look, here’s the opportunity.” You guys were more technical. “Can we make this work?” Tom was willing to pay for us to build this ISP if we could make it work.

I knew Alex through the Internet Éireann user-group and some shared circles in college. Even at the time it was clear that it was fantastic to be presented with this opportunity, and I jumped at the chance. I had no time for the standard model of success in business: graduate, find a stable company, slowly work your way up the ladder. It was the culture of national telecommunications companies, the culture of law firms, and indeed the business culture of Ireland, broadly speaking; deference to authority was a prized attribute, and experience was valued above all else. Or at least that’s how it seemed to a pair of twenty-somethings on their first unsupervised outing.

For those of you in the audience who think that this deference is a potentially harmful trait for business, the good news is that in no way did Alex possess it; we, and Alex in particular, wanted to do something different. We wanted to do our own thing: wanted to be on the edge of the new frontier, however uncomfortable it was, and relished the prospect of cocking a snook at anyone who dared to try to slow us down. Alex relates his rationale in more detail:

I thought the whole thing was fantastic. I would have done it for free because it was a chance to play with all this stuff and do it for real. It was just a fantastic idea to be able to build an actual ISP from the ground up, from Day Zero. Buy servers. Install the stuff on them, decide what equipment you want, configure everything – it was a dream come true that somebody would actually bankroll this and even pay us a little bit of money - which I think at the time was a very very very little bit of money to do it! But speaking for myself, I would have absolutely done it for free. So we said yes. I remember you came over to my house, and you insisted on using this new-fangled operating system called Linux, which I had never used or encountered. I didn’t ever think it would catch on.Dry, witty comment relating to how that all worked out.

Club Internet found headquarters on 2 Lower Mount Street, in a Georgian building where we initially occupied a front room on the ground level. It was a crowded space for a bunch of people with strong personalities, and most of them college age who were not cautious about the environment they worked in. There was not just physical detritus – a bike stolen from the hallway, a tuna and sweetcorn baguette hidden behind a desk which could never be satisfactorily linked to me – but emotional detritus as well. Never afraid of giving as good as he got, Alex partook in many arguments with almost everyone else, but particularly Tom.

I, on the other hand, did not handle conflict well. I had come from mostly failing a long series of exams in college and so had the twin pressures of regarding myself as a failure, while also needing money. So I stayed and tried to cope as best I could with the situation. Meanwhile, the company continued to develop. Tom Kelly found a connection via Galileo, the travel provider, who had a data-centre in Fenian Street for airline ticket-booking purposes, and he authorised a 64k connection from them, which I set up.Alex was absent on holidays for the first delivery of that connection, so I oversaw that process and enjoyed it greatly. I was there to see the first few packets trickle in, saw the lights flicker on the modems, watched the UseNet goodness trickle down our 64k line, and generally had experiences that few other 21-year olds were having. I imagine on some level it was parallel to the first telegraph operators on Valentia Island as the Atlantic cable came online.

In fact, Club Internet did a number of things differently from other ISPs of the time, structurally speaking. Alex French:

We did something which nobody had ever done before in the ISP business in Ireland, which is we said, “We’re not going to build our own data-centre room”, which at the time was crazy. Everyone else said, “Come in and I’ll show you the size of my data-centre and my cooling system,” but we couldn’t afford to do any of that. Instead we rented a rack, or a partial rack, in Galileo. This made perfect sense for us because we were getting what at the time was an excellent data-centre, we didn’t have to build it for ourselves, we didn’t have to maintain it. Of course now that’s what everybody does, but at the time most Internet companies ran their own data-centres; all the other ISPs poo-poohed us for it, y’know: “They don’t even have their own data-centre.” But ours was actually better than anything a lot of them had because it was outsourced.

Alex may be exaggerating a little here. It certainly involved less work, and certainly made financial sense for a startup company to outsource at the time, but it’s not clear to me that the quality of what we were getting was necessarily better. (It was, however, better able to survive rain than Internet Éireann’s facilities.)IT’s Monday from 22 January 1996 had a slightly different take on it: “MEDIANET - The first new Internet service provider to step forward this year is Club Internet […] The company has taken the unusual route of locating its main servers and communications hardware at the node of its international partner, Sprint. […] Medianet managing director Tom Kelly says this arrangement will allow it to obtain extra bandwidth at will, eliminating the lead times associated with leased line delivery from Telecom Eireann. He also guarantees that customers who seek a 64 Kbps leased connection to the Internet will get just that, and not a 64 Kbps connection to the service provider’s premises. […] Club Internet is also offering free Web space to its subscribers, who automatically become club members. It undertakes to distribute twenty per cent of profits among members, who will have a representative on the company’s board.” (I don’t know if any of those things happened.) Thanks to John Sterne for the reference.

⊕
It seems they also thought different about logos, as you can see here.

Another difference was the decision to go with multiple brands from the beginning. Alex French:

Tom Kelly’s idea was that a business ISP and a home ISP were different things, which really only in the last couple of years has become common practice. But at the time, an ISP was an ISP. The idea of splitting your brand was unheard of. That was something which was probably never fully exploited, but there were a lot of services which were sold under the MediaNet brand and especially to media companies like advertising agencies and things like that, what would now be called “managed services”. Whereas the Club Internet service, they were trying to keep it as a domestic service provider. To this day, people still think that there were two separate ISPs.

The flip-side of being afforded this fantastic opportunity coupled with the existential threat of the launch of Indigo’s free dialup product was feeling that almost everything was an existential threat to the company; the stakes seemed extraordinarily high. Alex confirms my characterisation of it as a war:

It was a very interesting time and I think people probably don’t remember this now, but there was such a cut-throat competitiveness between the ISPs […] that you just don’t see now. Which is obviously a good thing. We all hated each other!For some value of “we” and “all”. You would cross the street to avoid somebody who worked for the other company. You don’t see that in any sort of a mature industry now. Even the mobile phone companies, people move back and forth between them. If somebody left Club Internet to work for Indigo, it would have been like they had slapped your wife.As opposed to slapping your husband, presumably. –Ed It would have just been unacceptable in every shape and form. I think it was probably a symptom of how it was really kind of a frontier industry. There was no precedent for what anyone was doing. Nobody really knew how the industry was going to evolve. Nobody knew what was going to happen. Nobody knew if you were, literally, going to survive a couple of months or if there was going to be somebody new who was going to launch into the market. You don’t see that with a lot of things now. I can’t think of another industry in Ireland where it’s like that. Obviously, thank god, over time, everybody became friends.

One of the things that happened on foot of this existential angst and environment of a frontier industry was the “finger daemon” incident.

Let me explain. In the UNIX systems of the time, but very rarely today, the finger program allowed you to see who was logged onto a system when supplied with the name of a machine. ⊕A program had to be running on the machine to allow this information to be retrieved, and any such long-running program is generally called a daemon in UNIX parlance. On the other hand, if you just gave it the name of an individual user, it would allow you see some simple, mostly user-controlled, information about them.

Ireland On-Line had one of these daemons, and it allowed you to see who was connected to their service. This had two major effects. The first one was you could do the 1990s equivalent of starting an Instant Messaging sessionE.g. WhatsApp, Google Hangouts, SnapChat, etc. with anyone that you saw online, provided they were running the right software at the time.

The second effect was that it indirectly allowed you to measure what the company’s customer base was like, because you could write a program which would regularly “scrape” (i.e. get the results and interpret them) the user-list and build a picture of who was online and therefore what their relative capacity was. Alex greatly enjoyed this window into Ireland On-Line’s operations and thought about the equivalent for Club Internet:

The finger-daemon was really a d**k-measuring exercise. I think it started with IOL, which was the first one. They’d run this finger-daemon to allow you to see everybody who was online. So when we started we had to have one too. But the problem was that we didn’t have anybody online or we didn’t have very many people online [so it didn’t look good]. The only people who were going to check us were the people who worked in the other ISPs. No member of the public was going to go, “I wonder how many people Club Internet have logged on right now?!” So we wrote this increasingly elaborate program, which created fake people and logged them in and out. I think initially we were going to show how many people we have logged on. OK, we can’t do that – we have too few people. Then we said, “OK, we’re just going to multiply by a certain factor!” Then we realised we couldn’t do that because obviously you can’t have ten Niall Murphys logged on – so then we had a random name generator with first and last names, [based off the existing customer base]. There were lots of Niall Frenchs and Alex Murphys! But then we realised that we couldn’t do that either, because if your factor was a multiplier of ten you would have ten people logging on and logging off every time somebody logged on and off. So we extended the program and it became an increasingly elaborate attempt to try and simulate [a larger userbase] statistically, depending on the time of day! The amount of effort that was put into what was effectively making us look good only in the eyes of our competitors!Additionally, see e.g. VW emissions scandal, ISP speed test result distortions, and so on.

I didn’t particularly approve of the finger program, in keeping with my usual dourly unhumourous demeanour, but it was difficult not to get caught up in the general atmosphere of enthusiasm. Alex describes an incident of great amusement in the Club Internet office:

⊕
Ross Chandler and Alex French of Club Internet allegedly discovering a computer security problem.

I still have a picture of me and Ross in the office, leaping up in joy when we found a big hole in – I can’t remember who it was – I think it was Indigo or somebody’s web server software. It was taken by a webcam in the office and we were both just yelling “YEAAAH!”. There was a lot of that kind of stuff going on, that kind of messing, rather than what you would consider nowadays to be good business practice and acceptable competition. At the time again, there were no real rules either, written or unwritten, about how you ran an internet business. Even at the time there was no legislation that governed hacking. So, some of the stuff we did which probably now would have gotten us into a lot of trouble, at the time was very much a grey area.

Alex is not exaggerating the extent of the semi-competent flailing around by people who had relatively little clue, whatever about the extent of the law-bending. As he says, there was simply very little in the way of guidelines or clear models to copy. I recall that at one stage in 1995 we figured out that we needed things called IP addresses.The equivalent of telephone numbers for the Internet – a unique number you give to every customer who’s connected. This is a simplification, but the analogy holds to the extent that you can’t just make the numbers up, they actually have to come from some organisation with authority to give them out. I somewhat/mostly knew what these mysterious things were, but not how to get them, so we did some research and determined that we had to send a request to get some of these addresses to a “registry” of some kind. Essentially at random, we emailed some likely looking contacts in some text file we found somewhere. Of course we later found out that we had done the equivalent of emailing the business plan of the company to our nearest competitors.If I recall correctly, we first emailed Mike Norris, who at this stage was involved with HEAnet and so wasn’t really a competitor, and Michael Nowlan in IEunet, who was. When we found out what had happened I recall Alex just shrugging and saying, “Oops!” It was that kind of time.

On the slightly more successful side, I’d been able to do a lot for Club Internet with respect to server architecture, setup, and general technical work. In many ways, I’d had no choice but to be successful – startups are where everyone weighs in, and fixed working hours are a fiction. (I recall fainting once or twice, perhaps due to poor eating or sleeping habits, perhaps due to general stress.) I was also having a generally intense time outside of work. Around summer 1996 when I was repeating second yearSee here for why. in an increasingly desperate attempt to obtain what sometimes appeared to be an increasingly irrelevant academic qualification, I left ClubI, because I literally could not do everything. Or at least I felt the inadequacy to be on my side, rather than the fault of Club Internet, which you could also argue was paying beneath market rates for people to perform technical work for which large amounts of education, determination, and even skill were required.

In any event, the third time was the charm and I passed those exams, but I decided I needed an on-campus company to work with rather than a commercial hothouse, struggling to survive. It also didn’t help that I found Alex very difficult to work with, although in many ways I respected his acumen more than I did almost anyone elses’.

The Club Internet story became more complicated after I left. There was a wave of consolidation which struck in the 1998-1999 timeframe and in 1999, VIA .NET Works, an American conglomerate, came knocking on the door.Currently unreachable ENN story here – trying to find replacement. Their strategy was to buy smaller ISPs around the world and firstly, take advantage of bulk discounts to buy cheap bandwidth, which at the time was essentially the major cost to being an ISP, and secondly, find out which subsidiary did which thing the best and roll that particular best practice out across the rest of them. As Alex confirms, the economies of scale were massive, so it looked like a sensible thing to do. At the dawn of the noughties, VIA NET.WORKS ultimately ended up purchasing 27 ISPs in 14 countries including the US, Mexico, Ireland, the UK, and so on, one of which was Club Internet.For a reported purchase price of IEP 10-12 million in the trade press, although this may be incorrect.

At that time, Club Internet had drifted a little from its original aspiration to commoditise connectivity and get into the value-added services market – the market needed to be expanded before value-added services made sense, and that kind of high-effort infrastructure rollout always takes time and money. It had, however, started an in-house web design department, which was quite a high-margin business.As Alex says, “At the time, ISPs got web design business. It seems crazy. Who would do that now? It doesn’t make any sense.” Alan Judge recalls that Indigo and Eircom Multimedia had them as well. However, Club Internet’s non-core businesses were theoretically in the firing line with Via’s strategy of amalgamation. As it turned out however, none of this was destined to make much difference. As part of executing on the “buy-in-bulk” strategy, Via:

[…] bought a big-ass IRU for trans-Atlantic connectivity – IRU stands for an indefeasible right to use – it was basically a lease on trans-Atlantic fibre. They bought a lease on a big chunk of trans-Atlantic fibre, or what seemed like a big chunk at the time – I can’t remember how much it was. It was probably 50 megs or something like that!Again, seems very small now, but the practical minimum for an ISP at that stage was probably around the 1 meg mark. But they bought it just before bandwidth pricing started to implode, just before Global Crossing completed their global roll out. So almost as soon as they bought it, you could buy bandwidth cheaper on the open market than it was costing them to buy it. That part looked really good when they did it but unfortunately didn’t stand the test of time. It felt great to have the resources of an international company behind us, but [that play] ended up being their downfall.

Their downfall wasn’t quite so apparent in 1999, however. Alex ruefully recalls how great things looked immediately prior to that:

In 1999, VIA .NET Works went public on the NASDAQ. They raised a third of a billion dollars, which at the time was the largest tech flotation that the NASDAQ had done. Again, it seems crazy now; a third of a billion dollars is nothing today.I dunno. . . I’d take if off your hands if it was annoying you. –Ed I got some very nice share options, We all got share options, which we weren’t allowed to exercise until just after the crash, just after the dot-com bubble burst [2000/2001]. I know one person who made a little bit of money out of his Via options because he sold on the day after our lock-up ended and it was after the crash had happened and they were still worth a little bit. I remember telling him, “You’re crazy! They are going to go right back up!” But you know, at the time, I was picking out houses in my head, because it would have been worth a significant amount of money if the market had held on six months. It would have been six figures. Not seven figures. But it would have been worth six figures in terms of the stock options. But yeah, I never got anything. I never saw a penny out of that.

There were consolations, however. Alex went from working with a purely local organisation with about 15 people to being:

… a global product architect for a company that was trying to rationalise platforms across however many – 15 countries, was it? Suddenly you were asking, “How do you deliver web hosting all over the world? How do you deliver DSL? How do you deliver dial- up?”

It was a great time to be working and Alex got to work with some great people, perhaps ultimately the best thing you can hope for from any job. But since Via was really a company that was built for floating without a clear idea of how to execute on the amalgamation idea, unsurprisingly they pulled back when the cash ran out and it proved very difficult to make the argument that particular members of the VIA NET.WORKS “family” should stop doing things that they’d been doing for years, now that the prospect of centralisation was receding. Interestingly enough, the tradeoffs of the value-added services versus bare connectivity play became clear in the recession immediately following the dot-com collapse of 2001; while companies were generally easily able to cut (for example) their web design costs, very few of them were prepared to actually cut their Internet connectivity. So the connectivity supplier side of the business remained relatively stable throughout the recession, with continual cost improvements as the bandwidth glut brought down the crucial upstream bandwidth prices. In Autumn 2002, Alex left Via on foot of their de-consolidation of the whole Head Office and Technology function, which moved the company back to the situation that pertained prior to the take-over: allowing each local office to do what it wanted. It was time for him to do something new.

There is other evidence to consider about the Club Internet’s history and performance, however. Perhaps the clearest insight provided is via the Supreme Court action taken by VIA NET.WORKS against Stuart Fogarty and Aubrey Fogarty Associates in April 2002.See here for more details. In brief, Stuart & AFA petitioned the High Court to stop Via running the company into the ground. Via promptly countersued to dismiss the petition. (Both motions were denied by the High Court judge, and then the case ended up in the Supreme Court.)

According to the filed materials, in or around 15 March 1999, VIA NET.WORKS purchased shares from the existing shareholders and took a 60% stake in the company, while Stuart and Tom received slightly over £100k windfalls each and the other eight or so existing shareholders split £840k between them. However, what was not widely known at this time was that Club Internet was officially “in an insolvent position and in need of cash to fund its ongoing operations”, which may explain some of the urgency attached to the completion of the purchase agreement. Via also continued to support the operations of Club Internet to the tune of £2,156,549 in loans to the company for survival. After those loans it becomes less clear what happened. It seems as if there was a new shareholders agreement after the 1999 takeover, and that this allowed Via to, under certain circumstances, determine the price that would be paid for shares in the future. Via purchased the remainder of Tom Kelly’s shares in March 2000, and then told Stuart and AFA in writing in February 2001 that they would be buying the rest of the shares, but: “[the] letter also stated that as of the trans- fer date, the price calculated in accordance with the Shareholder’s Agreement would be a negative amount, but that they were agreeing to purchase the shares at a price of 0.01p per share.”From the text of the Supreme Court decision Or, in other words, the existing shares would be effectively worthless, because the price was so low.

Depending on which side you believe, Tom was replaced as managing director via a process that was either a mutually agreed settlement or a wrongful dismissal at considerable cost to the company. In an echo of the infamous bill-collecting abilities of Internet Éireann, Stuart & AFA furthermore alleged that “debts of the company have not been collected in a timely fashion, again at considerable cost to the company”. The case could be interpreted as a narrow victory for Via in the sense that Stuart and AFA’s original petition was struck out, but it could also be interpreted as a long-term win for Tom Kelly, since in 2002, Tom bought back the company which he’d previously left and turned it into a successful independent DSL provider called Netsource. According to Alex:

Via sold all their local companies – they were flogging them around 2002/2003 -– and Tom bought back what had been Club Internet […] from VIA NET.WORKS and called it Netsource. It was sold high and bought low, essentially. He bought it back from them for a song and really went into the DSL market and they were the first people who had access to the DSL platform on a wholesale basis because Eircom were being forced by ComReg to provide a wholesale platform and they did it with Netsource because they were a friendly, non-threatening company. Eircom was able to say, “Look, we’ve done a wholesale platform” because they did it with Netsource.

As Alex suggests in the quote above, NetSource went on to do quite well in the DSL market, and was later sold to Magnet Entertainment in 2006. ⊕A sad coda to the Club Internet story – many years after its disposal to Via but prior to the Netsource sale – was the discovery that the company’s other founder, Stuart Fogarty, like many other business owners before him, found it im- possible to resist borrowing money from his company, perhaps an eensy-weensy bit out- side of regulations: See for example this breaking news story for more details, although the evidence suggests the amount was fully repaid, and he later went on to pilot the merger of his advertising agency with O’ Connells to widespread acclaim. I never had much to do with Stuart – my most abiding memory of him is when he pressed a fifty-pound note into my hand at the end of some computer show or other, money unheard of to a proto-dropout like me – but I can attest to his generosity if nothing else.

In many ways Alex’s story about the dot-com boom and bust is a microcosm of what went on for many others during that time. Of course, it was many others who had been lucky enough to be born white and male, who had chosen the “right” career at the right time and so on, but there was in some ways an egalitarian spirit to the giving of share options which was not reproduced in more traditional businesses.See e.g. ABC News for a more robust defence of this practice. These options gave thousands of otherwise ordinary folks an opportunity to become rich that they wouldn’t otherwise have had. The capriciousness of capitalism aside, is it not unfortunate for society generally when wealth creation fails to happen, particularly when that wealth could be spread over a larger number of people?

Perhaps it’s appropriate to look at another entrepeneur who was, perhaps, a little less lucky.