I worry that there is something within Irish culture that is deeply anti-entrepreneurial. I believe it is evident even within Ireland’s own business community, the very community and sometimes the very individuals who say they want a Silicon Valley-like entrepreneurial culture here. So much so that most who achieve success do it outside Ireland.

Karlin Lillington "The Irish Times: Net Results"

Ireland On-Line

And now to another entrepeneur, luckier than many others, but unluckier than some.

⊕
Barry in days gone by.

In the honourable tradition of many self-made men, Barry Flanagan departed the school system in the North at 16, thoroughly disgusted by what he saw as the system’s willingness to lie to him for the sake of an easy time. He recalled doing well in Maths, Chemistry, Biology and other scientific subjects:

… I went into the A-Level Chemistry class and the first thing the teacher said was, “Right lads. All that stuff we taught you like covalent bonds and all that? Well, that’s not actually the way it is.” And I just said, “Hang on a minute! I don’t have time to be sitting around and you telling me what you think I can handle.”

It was a familiar story to me, for I had had much the same experience of intellectual betrayal when the same thing was done in my school in Dublin. ⊕I, however, did not respond by leaving school and becoming a serial entrepreneur, more’s the pity.

Barry’s introduction to computers was in the first wave of commercial home computer deployments that washed over Ireland in the 80s:

“‘Right. Computers. I should know how they work.’ I went and bought a Dragon 32, with a pretty basic keyboard and a tape recorder that went ciw-wakzch and figuring it out I said to myself, ‘Ok. I get it. I can use that.’”

I partook in the same revolution as Barry did, albeit with an Amstrad 464, but at this stage it was still mostly about computers in the home, not particularly talking to other computers. However, fast forward a few years, and it was in London during the late eighties and early nineties that Barry had his first experience of the “magic” of a modem. Working for his sculptor uncle, he discovered that NatWest had brought out dialup banking; this meshed well with the large amount of travelling he was doing at the time, and meant he could “still do an important banking thing wherever [he] was” – as long as the international phone bills could be paid. So he followed up by buying

“another modem and stuck that into the machine that was in London. So I was also able to dial into that and I could upload a document to the secretary who worked there and she’d have it.” It was another moment of revelation: “That’s fecking big! I can be anywhere in the world…”

The emigrant experience that Barry was living then became the foundation of his vision. What if, he wondered, you were able to create a service that allowed Irish people to connect to one another, but also to connect to external people around the world? And for people around the world to connect back?

But he was held back by the practicalities. This was 1991: Barry’s idea was “building an online service, but one of the primary things that had to be there was that you couldn’t be dialling long-distance. You had to be able to get to a local number. I didn’t know how that would happen.” Oldtimers may recognise the CompuServe model.Newtimers may require some background. At the time, CompuServe was the first (and for a long time, the largest) commercial online service in the United States. It was started in 1969 (!) and was most powerful and most popular during the 80s and 90s, operating much like a giant BBS. See later for more on BBSes.

According to Barry, he did indeed have the CompuServe model in his head at this time. “Compuserve was the big daddy. It was never going to disappear.” He paused ironically: “It was too big to fail, you know?”. But in common with the others in this book who went down this road, the problem was actually getting people connected. Barry bought John Quarterman’s “The Matrix” to try to get to grips with how to do this, and started wrestling with the complicated technicalities underpinning this stuff. Then, in October 1991, Telecom Éireann did something surprising – it launched Minitel in Ireland, on the French model.Minitel was very successful and very French. In that great Gallic spirit of not-invented-here, it was a proprietary text-based system which allowed access to a multitude of databases, the ability to book tickets, chat online, so on and so forth – essentially a mini-Internet as most people experience it today. The business model was to give away the terminals and charge a small monthly fee for access. It was so popular and so wide-spread (in France) that it was only turned off in 2012. (By “launch”, I do of course mean “announce on the Late Late show”, but the final product seems not to have shipped until 1992.)

For Barry, this was the final tipping point.

I watched that and I said, ’Right. That’s it now. I’m doing it.’ My big concern had been who is going to provide the network and how might you get paid for this. Minitel were offering that they would provide the network and get the terminals out there, […] and they collected the money for you and sent you a cheque at the end of the month. Great!

Barry rang them up. Perhaps unsurprisingly, he ended up speaking to a bunch of French people, who were stationed opposite the Volkswagen car sales forecourt in Ballsbridge. He records his shock when, having run his proposal by them, they quoted him a price for implementation of about two hundred and fifty thousand Irish pounds. It was neither the first nor the last of many Irish initiatives to be scuppered by the “charge what the market can bear” model. Barry reacted instinctively, asking for them to send him copies of the Minitel protocol so he could investigate in detail what was required. His absorption of “The Matrix” had led him to thinking of communications along the Internet model – cheap and cheerful, decentralized, and developed from a small seed. Minitel were talking about needing a fax machine, X25 PADs, and other infrastructure which felt like pure overhead to him. In the meantime, while thinking over what else to do, he started running a BBS.

⊕BBSes, standing for Bulletin Board System, were a precursor to many Internet services we know today. Special pieces of software ran on copmuters which allowed other computers to dial into them and access chat, download files, and so on. The total number of users might be high, but the total number of simultaneous users was limited by the number of phone lines that the BBS operator wanted to pay for so that people could connect. Most BBSes didn’t have very many (<= 3). See, for example, this boards.ie post.

Barry had the same difficulty of many people in start-up businesses: he couldn’t really explain it. Nobody had seen it before – no-one in Ireland, anyway, and at that time, Barry had no credibility because he wasn’t old enough:

When I was setting up IOL I was 29. A 29 year old had no credibility. One of the biggest barriers that I had was that people looked at me and said, “You’re too young. You gotta be forty and have a couch-belly to do this game.” Whereas now it’s completely the opposite. I’m 49 and I’m way over the hill.From one misguided extreme to another!

As well as the basic credibility barrier, a conversation explaining the business case might well have gone something like this:

“Look, I want to set up this thing that you would be able to connect to, and you would be able to do email or FTP on the computer…”

“But most people don’t have computers.” “What’s an FTP?” “Do you need a phone line? Wouldn’t that be terrible expensive?”

As an illustration of the state of play when Barry moved to Ireland in 1976, his family had a wind up phone. Even acquiring a phone-line had a lead time of months and it usually required a word in the ear of a monsignor to actually make it happen. Telecommunications was not foremost on the agenda then. But as the Department of Posts and Telegraphs changed into Telecom Éireann in 1984 and more modern equipment was installed, this gradually changed.

So Barry was running a BBS partially because it was difficult to explain what he was doing, and partially because he was still searching for something he could do which could maybe plug into Minitel at a later stage. But then there was a sea-change: at the end of 1991See e.g. Wikipedia. the news came through that the NSF had lifted the restriction on commercial use of the internet:

I dug out my Matrix book, and I found the TCP/IP protocol. Having studied the Minitel specs and the X25 specs – X25 was the proverbial dog with fleas, designed by committee. Whereas TCP/IP: I looked at it and I said, “This was made for what I want to do.” As of now, you have this global-ish, ready-made network that’s out there, that I can conduct commercial traffic on.

Now the plan was coming together:

I’ve got my network. I’ve got this TCP/IP thing with which I know I can do something. Now, at the time, on the internet you had basically FTP and Telnet. There were a few public libraries around. There was Gopher. It wasn’t that diverse.

So Barry had moved from London back to Ireland by this stage, to Galway in fact, again on foot of his vision: he felt ideologically or idealistically that this new technology should allow anyone to live anywhere in Ireland. Clearly the name for what he was doing should be Ireland On-Line, but Barry didn’t actually use that for a while; instead, using a credit card from when he had lived in England with a ten grand sterling limit, he ran the BBS as Galway Online. It was an inspired move of brand separation. It could fail and not damage the main name and idea. It also opened the doors to acquiring regional funding, and in the white heat of the startup game, Barry was anxious to acquire whatever funding he could.

For the moment, however, Barry had to feed the kids and keep everyone going, so his main business was modems. Their lack of availability was one of his major problems – if people weren’t able to get their computer connected, he didn’t have a business. At this time they were ferociously expensive; for a 2400 baud modem, approximately the speed that makes smartphone users stare at their device in frustration and meaninglessly hit reload, it was a wallet-hollowing 400 Irish punts.Approximately 500 EUR at the time of switchover, 2001.

Of course, they all had to be licensed by Telecom Éireann, and Telecom Éireann were the only people who sold them.Good old incumbent culture. So Barry:

started importing modems from California. I’d buy from, well 9600 baud initially, which I bought for $99, and sold them for £199. So I was making a hundred percent margin. I was actually selling twenty a month. I had a little thing going where I bought a big roll of brown paper and tape and labels and stuff. The kids were wrapping them all up.

As well as providing a profit margin sufficient to make ruthless capitalists rueful, Barry’s fulfillment business was providing employment for the whole family. The guarantee was excellent too; if it didn’t work, Barry just sent another one in the post, and binned the old one because it was easier than shipping back to California. It all helped to keep the wolf from the door. By May of 1992 Galway On-line had been launched and Ireland On-line was incorporated. It was originally called that, at least, but when Barry obtained funding from the regional development agency Údarás na Gaelteachta, the company had to be renamed An Nasc Teoranta, which could be translated as, among other things, “The Link Limited”.

In various ways, Ireland likes to think of itself as being different from other countries – celtic exceptionalism, if you like – but in this way it is very similar: the regions hate the capital, and the capital couldn’t care less about the regions. And this is the story behind Ireland On-Line’s first funding, and a story of politicians doing correctly and honestly what they are supposed to: connect people.

Barry’s major problem with raising capital was not knowing anyone. One day, he saw an ad in the newspaper for The Western Alliance, a lobby group set up by Pól Ó Foighil,Anglicised as Paul O’ Friel. a man from Spiddal who was a senator at the time. His big thing was trying to get the regions to go directly to Europe for funding rather than having it routed through Dublin where, as previously mentioned, it was perceived that money had a suspicious habit of staying, whatever its intended destination. Barry volunteered to help out with Pól’s Western Alliance project and eventually the subject of what he was working on came up.In fact, Barry had other plans if this whole Internet thing didn’t work out – one of them was a computer game about Cú Chullainn, and another was on electronic music distribution, one and a half decades before it actually happened. A distance learning effort called Coláiste Lectronica was another one. Barry told him about Ireland On-Line and Pól was enthusiastic:

So Pól said ‘Listen, you can’t be working in an environment like that with kids running around and all that and tugging at you. I have an office up in Spiddal. We’ve got computers, printers, faxes. I could use a hand. I don’t know anything about all this computer stuff. If you want to come out there and work out there and use all the facilities and you can help me out.’

Barry agreed, and traded office space and facilities for an early implementation of the state of the art of computer-aided politics: a database for the constituency, complete with mail-merge functionality. Barry finished his business plan and went looking for money. The capital markets were quite poorly developed at that stage: there wasn’t the large cargo-cult Silicon Valley infrastructure that operates today, or at least if there was, the connections weren’t in place in Ireland. So Barry went to the banks, and perhaps it would not be stretching the reader’s credence too much to report that the banks didn’t believe in his business plan, didn’t believe in him, and told him he’d never get any money from them. Barry was discouraged, but Pól stepped in and said that he knew a guy in the leasing business, and so Barry ended up applying to this company for the price of a serious new computer at the time – three thousand Irish pounds.

Unbeknownst to Barry, Pól had privately undertaken to act as personal guarantor on the lease, and it came through. With this act of selflessness, Barry could now develop on one machine and use the other as an actual server.For historical context, this was on a 386SX with 1 megabyte of RAM, and a 40Mb hard-drive; in many dimensions less than a hundred times powerful than today’s phones. But Pól’s helpfulness did not end there: he had a suggestion for someone else to take the full funding proposal to. He was a business partner of Pól’s in Coláiste Lurgan in Inver and he worked in Digital Equipment Corporation (DEC), then headquartered in Galway, so he had the background in I.T. that Pól did not. His name was Michael Grealy.Later HR director of Aer Lingus and previously Head of Group Human Resources at Bank of Ireland. That surname would turn out to be significant.

Barry met Michael in a hotel on Eyre Square, and Michael, having read the business plan, said that he’d been thinking about this kind of thing for a long while: “But you’re the first person that has ever made it make sense. So I want to put some money in.” He ended up putting in twenty thousand pounds. The plan which had attracted this money, a very significant amount of money at the time – particularly for an individual – was what was to become the classic “selling access” model, expressed as a subscription. There were parts of the plan that were based on the extremely early versions of the e-commerce world, but:

… you couldn’t put that in the business plan; that this was going to be turning over a million in three years time and that you’d have a hundred staff or anything like that. They would just laugh at you. So I was looking at something that would provide a living for four or five people and maybe we’re talking about subscribers in the thousands – that sort of thing. You know business plans; they’re models. The idea was the important thing and showing that there was a model.

The next step was to get some actual Internet connectivity as opposed to simple BBS access, although that was enough to get up to three users on at once – Barry recalls sitting up in the middle of the night watching people use this thing that he’d built, and even though it was not as exciting as what he had planned, it was still intoxicating.

As with all other matters Internet related, Barry turned to Quarterman’s “The Matrix” and turned to the reference on Irish Internet access – under that was a phone number in Trinity College. Barry rang it, and Michael Nowlan answered. Barry introduced himself and said:

“I have this online service here in Galway.” I wanted to offer [the customers] purely Internet services. And Michael said, “Sorry, let me get this straight. You want to take what we have and you think that you can sell it to other people?” “Yeah. That pretty much sums it up.” And he said, “Oh. Nobody’s ever suggested that before. Sure. Why not? Come on down!” So I went down to O’ Reilly Institute and there’s Michael. Now, they were all full time lecturers. Bear in mind that I’d never even seen the Internet at this stage – I’d literally only read the book. And Mike says to me, “So, what bits do you want?” Barry had to confess that he didn’t know. So he reached into his bag of acronyms: “Show me the FTP bits.” And so Mike showed him downloading a file, probably something that looked a little like this:

csh% ftp gdead.berkeley.edu
Connected to nemesis.cs.berkeley.edu.
220 gdead FTP server ready. Complaints and questions should be sent to
  <kraitch@eecs.Berkeley.EDU>.
Name (gdead.berkeley.edu:mnowlan): ftp
331 Guest login ok, send your complete e-mail address as password.
Password:
230 Guest login ok, access restrictions apply.
Remote system type is UNIX.
Using binary mode to transfer files.
ftp>

It was the beginning of a beautiful relationship.

Of course, with Irish telecommunications being in the state that it was, Barry couldn’t immediately get a leased line down to Galway for interactive Internet access – i.e. access that you could use to do something right there and then. Instead the way it worked was that computers would dial each other up, immediately begin transferring email and files, and then hang up when they were done. This was done using a protocol called UUCP – standing for Unix-to-Unix Copy, and it was just what the doctor ordered for third world countries with poor telecommunications systems unable to support cheap connections… and Ireland. Later on, in late 1992 or early 1993, Barry was actually able to connect things up so that he had local modems in Dublin, connected to the Internet via a 64k leased line through Furbo, where he now had an office at Údarás na Gaelteachta. This meant – as far as I can tell, uniquely for the time – that Barry could service people dialling into his BBS in Galway, and people connecting to his BBS directly over the Internet. It was the first moment he had global access.

This was way before the web as we knew it, so the interface that Barry supplied to his customers looked at little like this:

Employment opportunities in this new field were a little … ad-hoc. Barry got talking to one of his subscribers at 2am one night – of course, the phrase “talking to” actually meant “sending text to” via the Internet – and it turned out the person on the other end was doing Computer Science in University College Dublin.A few years before I did. Ronan Mullally was his name. A few months later there was a knock on the door in Galway, Barry answered the door and there was a stranger there. He said “Hey Barry! It’s Ronan Mullally!” Barry brought him in, made him a cup of tea, and showed him the infrastructure; a common social ritual amongst geeks. In the manner of a street dealer in logic, Ronan reached into his coat and pulled out two floppy disks. (Young ones amongst you probably know this as “the save icon”, but I guarantee you, they were real.) He said: “Here, you might like these. You can stick them up for people to download. This one’s a thing called Minix and this one here … this, is a thing called Linux.”

(Remember that name, Linux – it turns out to be important later.)

So Barry installed a Linux system by compiling his own kernels,A complicated and technically fiddly procedure which mostly consists of staring at the screen, waiting, and swearing. in time-honoured tradition. In this case Barry didn’t have enough memory to do it in RAM, and the supplied kernel didn’t support his hard-disk controller, so he had to use another floppy disk as swap space. It took about a day and a half. But it worked! Within a year, the previous software infrastructure, which relied on SCO Unix, was unceremoniously dumped. As Barry confessed, “We were one of the first companies in the world using entirely an infrastructure that relied on Linux. It was really early. It was really dodgy! It was all that we could afford. It just worked. It has never let me down since.” In due course, Barry’s “dealer” became head techie of IOL.

Round about this time (approximately winter in 1993 or early 1994) the early stirrings of the web were beginning to make themselves felt in Ireland. Barry recalls with glee the time he first encountered the first of many graphical web browsers, Mosaic:

The first time I loaded up that I just laughed my balls off. I just felt like the cleverest fucker on Earth because everything I had built was centred around that technology - HTML, et cetera - and suddenly there was no more text-interface. I could literally point Mosaic at the thing I already had and it was suddenly graphical. And you could click it with a mouse! I put up a picture and I laughed and laughed and thought, “This is great now. Now we’re getting somewhere!”

It was a salutary lesson in skating to where the puck would be.Wikipedia has the famous quote.

The disadvantage to skating where the puck will be, of course, is that the one place it definitely isn’t, is where you are now. This was particularly acute in the case of Ireland On-Line, because the market had to be able to use the product, and it was the era of Windows 3; there was no “native” TCP/IPInternet protocol software. for Windows - no concept of it in the base operating system at all. That would be literally unthinkable today.

Barry knew that the problem had to be solved, and most of all, it had to be kept simple. A simple solution could be handed out to people on a single disk, it would control modems via the “AT command set”Basically a specific set of characters that would compel the modem to do specific things, see e.g. here. and dial in, doing as much as possible for the user. While the first browser had been released, it was far from ready for any serious use; what was needed was an interface which could be put into people’s hands right away. To this end, Barry enlisted the help of a man called James Gleick, at that time famous for a popular science book called Chaos, who also happened to be into this trendy Internet thing. It turned out that Gleick had “… developed a Windows program, basically, that talked to gopher servers, as-was, but formed a little Window on the screen and you had the menus for mail and all that; [simple] Visual Basic sort of stuff.”

Barry knew that the interface was key, however – it had to be something that people could use and would even want to use. So Barry flew to New York, met Gleick, and licensed his software. Suddenly he had a disk:

… that we could give to somebody who was interested and they could load it up and it would log in and they could use the service. It made all the difference in the world. So we opened with that for a couple of months and it went really well. Certainly it got a lot of people’s attention.

In particular, IEunet were distinctly unhappy; they were apparently ready to shut down Ireland On-Line because they interpreted what was being supplied as a violation of the commercial agreement that was in place. Barry’s claim is that in fact, it did not do this; the client did not get an IP address, it just connected over serial and the characters of the text display flew back and forth. It took some explaining, but it was a useful stop-gap while the graphical browser situation developed. Barry worked solidly on that too:

[and] spent many hours and hours putting together a single floppy disk that basically had Netscape 0.9-something, Trumpet Winsock and Eudora Mail. All on a single 1.4MB floppy disk. I wrote an installer around it. Basically you put it in, you installed it, it asked you for your username and password and – bang! – you’re done.

He phased out Gleick’s software and was settling down to rolling out his new installer when Netscape came calling; they came to see Barry in Furbo, saw his disk, nodded politely and asked for a copy to take away with them. Shortly thereafter they attempted to get Barry to “license this […] at $10 a copy, and for a minimum of $10,000.” For a small startup ISP in Ireland, always strapped for cash, that would have been the end of everything.

In the end, however, they didn’t charge for the ones which had already made it out into the field, so the request for a commercial relationship effectively became a cease-and-desist letter. Barry reacted by noting that while Netscape 1.x might have a license precluding what he wanted to do, 0.9 did not, and was not restricted. Netscape attempted to argue that this was not correct, and eventually Barry was forced to pay, but he was not happy about the terms:

I said, “Right. I’ll tell you what. I have to pay you this but before too long, there is going to be an alternative. Someone’s going to come along and when that day happens, you’re out of here. And you’re dead as far as I’m concerned. And I’d say the rest of the world as well.” Which duly happened. Microsoft came along and said, “We’ve got a good browser!” (It was actually a good browser for the time.) “Not only are we not going to charge you. But we’re going to pay you to distribute this.”It was the Irish arena for the browser wars.

Ireland On-Line had its day in the sun with Netscape and thereafter was able to tell them over the course of a month that Microsoft’s share of the browser market in Ireland had gone from approximately 0% to 73% mostly on foot of that install disk. It was a convenient alliance, but it lacked, as they say in the North, parity of esteem; Barry’s native inclinations being anti-establishment, he always had an uneasy relationship with such authorities as already existed in the computer and networking worlds, even as he was working very hard to make Ireland On-Line one of those authorities itself.

Nowhere is this clearer than in the relationship with Microsoft. Taking into account the above, one must also take, for example, the trick that Barry played on Microsoft for April Fool’s Day in approximately 1997:

So anyway, for April Fool’s day, I asked Garrett Flynn, I said, “Garrett, just recompile the HTTP daemons so they report they’re IIS.”Microsoft’s web server software, then competitive with the open source Apache. And Garrett said, “Why do you want to do that?” And I said: “They’ll find it later. It will take a month but it will be good.” So Garrett did it and that was grand and the next thing, in May, came a really excited call from Microsoft: “Barry! Barry! We can’t believe it!” I said, “What, what?” And the rep said, “IIS server market share in Ireland has gone up to 65%. You switched over!” I just started laughing at him and said, “Alan, come here. I wouldn’t go spreading that around internally too much.” He says, “Why? It’s great news.” I say, “No. Look at it tomorrow again.” He says, “What do you mean? What do you mean?” I says, “I just changed our servers to report they’re IIS as an April Fool’s joke.” He says, “How can you do that?! You can’t do that!” And I said, “Ah yes, the power of open source!” The next month the IIS figures were down again.

This was the kind of thing that keeps multinationals on their toes, jumping at sudden noises, and generally remarking querulously that things aren’t what they used to be.

Go Grealy Go

Back to Michael Grealy, the initial investor, who after examining the original Pipeline-based service complimented Barry on it, but told him that he needed sales people to hit the road and actually sell it. So Michael suggested: “My younger brother Colm Grealy, he’s a school teacher, but he can sell anything. He loves technology. He plays games and does all that stuff. If we gave this to him, he could sell this like nobody’s business.” Barry agreed to meet him not only on the basis that an additional pair of hands would also be useful, but also on the basis that, as Colm already had a full-time teaching job, he therefore wasn’t going to cost anything.Ah, startups. One meeting in the Ashling Hotel later, and Colm was on-board as sales manager, balancing this with being a teacher in a special needs primary school in Jobstown. Barry recalls that:

Colm used to get changed in the toilets at school into his suit and then after school at three o’clock he’d go and do a couple of meetings. He started in the summer, but in the school year he was doing both these jobs for about probably six months.

In 1994, early 1995, things really started to solidify and intensify; Barry gritted his teeth about the logistical necessity of setting up in the big smoke; opened up an office in Amiens Street, Dublin, where Colm was based; took on a few more people, and gawped at the momentum this Internet thing was acquiring. One way of measuring it was by the activity at trade shows: in those days Frank Quinn and Mark Egan of Computerscope used to hold shows in the RDS (and elsewhere) hosting stands selling or advertising various companies’ products. At that time, IOL had started pressing CDs as well as handing out floppy disks, so they had free stuff to hand out at these shows - always popular. Consequently, most of his job was crowd control – the queue at his table used to be:

“… twelve-deep, twenty-deep, everyone crowding around. The next show that Computerscope had, all the exhibitors were asking if they could get a stand next to IOL’s!” People were definitely engaged, and becoming more so: “We had a great user-group. We used to have user-group meet-ups and 300 people would show up for a day down in Abbey Street and different places. We’d have workshops and talks and different things.”Of course, as we gradually commoditised over the years, that engagement was lost – today, we no longer see similar meetings for O2 or Magnet, say, although we do see it for the likes of Facebook and Google developer meetings.

Barry was bowled over by the loyalty of his customer base, and would travel to Cork, Galway, and to different regional meetings, as a way of showing loyalty in return. The buzz amongst the customer base was, however, in sad contrast to the buzz amongst potential investors. Having exhausted the family-and-friends route, Colm’s onboarding allowed them to make time to go to Dermot Desmond and other “high-net-worth individuals”; venture capital was virtually unheard of, and Delta Partners and the like were just setting up. Barry found he was spending more time educating the people he was asking for money about the Internet than the time he spent actually asking for money. The problem with having to educate the investors of course, is that they nod agreeably, go away, take that knowledge, and apply it elsewhere.

Mostly in the meetings, however, they would say they were concerned about CompuServe and Microsoft as possible competitors, and then Barry would point out that it made no sense for CompuServe to come over here, it made much more sense for them to partner if they were going to do anything at all, and the same for Microsoft – the launch of MSN scared the market, but in the end MSN in Ireland amounted to a default placement for IOL on Windows ’98, thanks to some skillful negotiation on Colm’s part. Literally, clicking on the Internet icon brought a user to the Ireland On-Line sign-up page. But all the time Barry and Colm were doing this, they didn’t actually want to do it; someone had to get the customer connected, someone had to develop the market – but IOL was, despite trying very hard, never in the best position to do this. The plan had always been to be an online service, in common with Club Internet and so many others. Barry felt very strongly that if anyone should be doing this, it was Telecom Éireann – after all, they had the national network, and stood to gain from its use.

But the difficulties of dealing with The Incumbent Telco were considerable. Barry was wondering about how to deal with this, when he got a call one day (this was when they were still primarily in Galway.)

It was from the regional manager of Telecom Éireann. He says, “Hi Mr. Flanagan. I’d very much like to come out and meet you. What is it that you do?” “We’re an online service.” “Right. Well, I’d very much like to come and meet you. Is there a time when…?” And I said “Yeah, that’s great,” and I met him in Barna, just outside Galway city. We sat down and he says, “The reason I wanted to meet you was …” and he pulls out a printed-out spreadsheet and he folds it out and says, “… these are the top revenue generating phone numbers in the country. This top one here is Directory Enquiries. You are number two. Twelve and a half percent of all call revenue in this country is being generated by your numbers. What is it with you guys?” And I said,“Funnily enough, I’ve got to talk to your people in Dublin because I’d really like for Telecom Éireann to provide us with a national network so we could generate those kinds of revenues throughout the country. And you could pay us, and everyone would be happy.” So we set up a meeting in the head office, on Stephen’s Green at the time. Colm and I go down. I’ve got my laptop with me. They steered us into a big boardroom and all the head people seem to be there. We’re just chancing our arm. I said, “Look, it’s probably easier for me to show you what we do rather than to sit here and describe it. Have you got a phone line?” They all laughed, “Ha ha! Very funny!” So I just plug my laptop in. Open it up. Dial into IOL and here we are, going around the table and they’re all mesmerised. They can’t believe it. “You’re doing this over our phone network? That’s amazing! So what can we do for ye?” Colm got up and says, drawing on a whiteboard, “Well, you’ve got this national network. You’ve got Galway and Dublin and Limerick and Cork and all those other places, right? Let’s call this, “Eirnet”. If you had modems in each of those places, all in local places, then we’d connect into your network and people could dial with a local number, stay on much longer. You guys are winning and you split the revenues with us.” They said, “Yeah, but if we did that for you, we’d have to do that for anybody, because we’re a state monopoly.” We said, “That’s great! Do it for everybody! That would be great. Let anybody plug services into this. We’re going to generate higher revenue. You can sit back and laugh while other people do all the work. Everybody wins.” “Oh alright. We’ll have to have a think about that.”

The next time Barry heard from them was when they sent him a thirty-day notice for unpaid line rental.

Although it may well be highly in keeping with Telecom Éireann’s culture for them to have reacted in this way,See this article for more. there is also a significant case to be made that this was negligent on the part of the decision-makers. Nothing happened for IOL, but in fact:

… subsequently they basically [rolled out that model…] they did Tinet, which is the same principle. Which is a pity because our idea was better. Imagine where the country would be today if that network had been built and had been allowed to be used like that “for the good of all service providers” so that anybody could plug in and use that common infrastructure for everybody’s benefit. Instead, we made a big fucking hames of it.

This was far from the first, and far from the last time that a failure of vision amongst the decision-makers would disadvantage very large numbers of people.“It is a feature of the financial world, much remarked by Warren Buffett, that people would rather be wrong in a group than right on their own; the people who insist on being right on their own tend to have the psychological equipment to match.” – John Lancaster At this stage, it was still more expensive to ring Dublin from Cork than Cork from Cork, so it would have made real financial sense for local numbers to have been provided for folks to dial. However, it was not to be. Instead, Telecom Éireann’s intransigence forced Barry to go looking for a network partner, probably to TE’s ultimate loss.

Quite by chance, Barry and Colm came across a company called PostGEM, an An Post subsidiary, at a small conference in the Burlington, Barry Collins being the contact. It turned out that they had leased lines into every post office in the country, which made them de facto the second biggest network through Telecom Éireann. It was under-utilised (being primarily designed for carrying X.400 and inter-office traffic, and most idle in the evenings. It was a perfect fit. In “no time at all, we had done the deal with them that based on a demand service, they would put in the modems, they would manage it and just deliver us back IP.” This deal also led to one of the most hair-raising equipment moves in the short history of ISPs:

In early ’96 I think, Ronan Mullally and I, at one o’clock in the morning one Tuesday, shut down all the servers in Galway, put them into the boots of his and my car, high-tailed it down to Dublin and into the data centre in Dublin and booted them all up again. By six o’clock that morning we were up and running.

The PostGEM deal meant that the network could offer reasonable cost endpoints more or less everywhere with any significant population presence. It was essentially what they had wanted from Telecom Éireann except by another route, and was almost certainly an easier relationship to manage. But IOL’s success, and growth, was attracting attention from others. Barry’s deal with IEunet for a 64k line was increasingly under threat, since they were still “charging eight punts an hour for connectivity, and once IOL switched from the Pipeline software over to a pure web-based approach, our current arrangement with them wasn’t going to work.”

There had been a certain amount of friction and risks to our service, shall we say. So I got a 64k connection into PIPEX, a UK provider, who we were also talking to about possible investment in the company.

A later meeting with IEunet became quite heated until Barry revealed that Ireland On-Line had bought another connection to the Internet, and didn’t really need IEunet’s. Immediately that changed the tenor of the conversation, but the relationship’s days were numbered. Barry thought that “IEunet did have a slight feeling of owning the Internet [in Ireland]” and that perhaps that proprietariness was ultimately not going to make for a viable relationship. In any event, despite an offer of peering, it was all shut down by mutual consent. But Barry was feeling pretty good, despite all the ups-and-downs:

So coming into the end of 1995 we were feeling pretty good. We had the deal going through with the network with PostGEM. By this stage we probably had about twenty staff in Dublin in Amiens Street, in sales and technical support. I was spending much more time in Dublin – I was actually sleeping on the floor of Amiens Street in a sleeping bag under the boardroom table. I did that for six months.

In true startup-and-an-insane-growth-curve fashion, everyone did everything. The installation pack creation and the whole procedure around that needed to be changed: the current state of the art required them to be manually copied, and that wasn’t scaling well. So it was revamped, including proper graphic design, new boxes and manualsRemember when things came with manuals? , and Barry got them placed in Compustore, Xtravision, and EasonsThe first of which is out of business since 2004, and the second of which went into receivership in 2013. , to name but a few.

It was a retail extravaganza; you couldn’t go anywhere significant and not see them at the till. Retailers would take as many as they could get and sell them for a quid or two, at no charge to themselves. If they managed to sell the larger box, they would make 10 to 15 pounds off them. Things were going like gangbusters. It was coming towards the end of 1995. “[Bill] Clinton was in town on the 8th December. But on the 5th December we heard that this new crew, Indigo, were going to be launching and had five million behind them and would blow the whole market.”

This seemed to be serious competition, and for a while everyone was running scared. But Ireland On-Line had one inviolable advantage:

Indigo couldn’t get into any of the shops because the shops get their Christmas stock by the 8th of December, and on this particular 8th of December the entire town was being shut down, I think by eleven o’clock in the morning, because of Clinton’s visit. We had people who were working for forty eight hours up to that getting all the boxes ready, getting all the orders for the shops. Every staff member who had a car was loading up their boot to get these things in. We knew that whatever they were launching, they couldn’t have a retail presence in time.

Barry still thinks that was a major miscalculation on their part, albeit a highly non-obvious one, since retail packs for access had not been popular in previous years. Barry alleges that:

… they hadn’t originally intended to do this free thing. They were built for just your classic pissing in the wind advertising campaign, billboards and all the rest of it. It was a bit wasted when all the retail shops and absolutely everywhere was the green install-pack of IOL.

But it was still difficult because Indigo were giving away for free what Ireland On-Line was selling for ten, twenty quid a month.

That was a tough time. I nearly had a complete breakdown. I did get ill. I was ill for three weeks after Christmas, just with the sheer pressure of it. I thought was it all going to come unstuck, with somebody just coming in and basically buying the whole market out from under you? They reportedly had five million to spend, while we had fuck-all. We were literally running on cash-flow, always had been. It made us smart and nimble and lean and all that but it made us pretty ropey too.

It was the way business was done. But despite the difficult times, in the end, the people who were with Ireland On-Line stayed with them – the fierce loyalty that had been inspired at an earlier stage was difficult to unseat with a free product, which people were natively suspicious of. It was said that by April Indigo would have ten thousand customers and Ireland On-Line would be dead. As it turned out, Ireland On-Line gained ten thousand and Indigo got a certain amount of traction, but never enough to significantly disrupt the overall business. Then Indigo went looking for a buyer. As Barry says, “it was quite the year”.

One unanticipated consequence was that the door was definitely shut on private money; the existence of a direct competitor with known deep pockets made it effectively impossible to make non-organic growth happen. So in the first half of 1996:

… we entered into exclusive talks with An Post or PostGEM/An Post, depending on how you look at it. There was really no option at the time, which is unfortunate. If we had been able to get a million quid it would have been a totally different story. Being in the ISP business did not suit An Post one bit.

Barry paused. “Well, except for they made a hundred and twenty million out of it.” There is a lingering frustration, even today, however, that An Post never really understood what it was getting, or why it was getting it. Barry contended that they didn’t do what could have been done, and they weren’t really interested in developing the brand, which was the strongest one in the market at the time, although it was undermined by the inactivity forced upon them by corporate dogma and “inability to actually see something as it is and embrace it. Instead they wanted to smother it, insulate it.”

The question then has to be asked: why sell the company at all? Of course, Ireland On-Line had constant cash problems and the business required constant investment. Investment which was in great danger of actually being returned, seeing as how it was a growth industry, but it required a leap of faith every time because the capital expenditure was in front of the customer growth, not trailing it. As Barry says,

… while it builds up, the economies of scale get better but you really have to have the resources to see that through. And without [those resources], you are just going to end up in a big hole. So, I mean, we managed. The fact that we managed it at all on a hundred and fifty thousand in share capital is quite astonishing. The fact that we beat all-comers, that we beat the incumbent, handily. I mean, they had to buy Indigo to try to make something of it. We beat IEunet. We beat everybody else that was out there. When I left, we still had seventy percent market share.

I remember being in the room at Alexandra Terrace, where IOL moved to later, when Barry and Colm made their departures in January 1999.Press release announcing their departure. It was sad. I should explain that I am a veteran of some departure ceremonies where no-one gave a shit. Let me tell you that this room definitely gave a shit. They were cheered, clapped and supported. They had the decency to look embarrassed. A long litany of achievements were read out. The room was uplifted despite also being sad, since we were acknowledging in some way that vision mattered, that there was virtue to having an idea and chasing it, and that changing things was in and of itself worthwhile.

Despite the good feeling in the room, I myself was melancholy. It was all a matter of timing. So much of what Barry and Colm had done had blazed a trail, but it had all been very difficult, more difficult than it should have been. And there was always the feeling that they had been too early: that those that came after would have an easier time because the arguments that had played out in boardrooms had eventually, reluctantly come down on their side, long after the answer had ceased to be relevant. That money, never in plentiful supply in Ireland, had never been plentiful enough to do what Barry had really wanted to do, when he wanted to do it. It was all, ultimately, a day early and a dollar short.

After Barry and Colm left, it was difficult to escape the feeling that the company was not going to do anything significant, ever again. Whatever about the entrepreneurial energy that they had brought to bear, Mary Mangan, the nominated chief executive, was clearly exploring options from early on for sale in the effervescent Internet climate of the time. Despite Ireland On-Line making significant losses (accumulated ones from purchase by PostGEM apparently amounting to slightly under £3 million on revenue of £5 million) and outside observers saying that she would struggle to sell it for any reasonable sum,An Independent article dismissing a sum of £18 million as “setting sights high”. either adroit negotiation or a happy confluence of business motivations led to Denis O’ Brien purchasing PostGEM/IOL for the staggering sum of £115 millionThe Independent again. in September 1999, which was a record for an Internet company at the time. Despite much happy clapping in the boardrooms of An Post, there was widespread skepticism among the ground-level staff, which was (in their eyes) confirmed when shares in the Esat group were promised and failed to materialise. A hostile staff meeting had attendees heckling senior Esat figures, alleging that “[they] had promised to look after us”. A phrase with considerable cultural resonance in Ireland.

Well, whatever the tone, the result was the same: of all the alleged £150 millionThe Independent again. distributed in the form of share purchases, none of it went to a PostGEM/Ireland On-Line share scheme at that time. That didn’t matter to me. What did matter, however, was that Ireland On-Line post-takeover(s) was a very different place to pre-takeover. Post-takeover, and other veterans of buyouts can start nodding sagely at this point, nothing was in a position to change because no-one could find out who would authorise it; with two takeovers in a four month period, it was absolutely pellucidly clear that the whole company was just frozen, and would be for an indefinite period of time. It was not very satisfying to be in a company where nothing changed, and nothing would change for quite a while. I began making plans to escape.

Nick Hilliard’s memory of the affair has a slightly different emphasis:

PostGEM swallowed Ireland Online in an equity-for-debt swap in around 1998 or so. [It was] a salutary lesson about what happens when you don’t pay your bills on time. Part of the deal involved setting up an ESOP for the employees – this would have been exclusively Esat Telecom Group shares. A certain amount of money was put aside for this and the lawyers went off to handle the legalities of it all.

At the time that the ESOP was first promised (summer 1999), Esat’s market cap was a chunk under $1bnSee e.g. this press release. (i.e. $20 per share or $40 per ADS). There were a whole pile of things going on at around this time, but in particular Telenor was attempting to mount a hostile takeover of Esat Telecom. Denis didn’t want this at all, probably because of the Esat Digifone / Telenor difficulties, so he started courting BT. At the time, BT was flush with money and needed to do some purchasing, so a bidding war started. By the end of it all, BT [had] bought out the company for $2.46bn in March 2000, i.e. a 250% markup on the share prices of when the ESOP was originally promised.

Meanwhile the shares hadn’t actually been purchased by the ESOP, although the money was sitting there. As Esat Telecom Group no longer existed, the ESOP planned to buy BT shares instead. The problem was that the dotcom bust had just started and between March 2000 and the time that the ESOP received revenue approval, the BT share price dropped from $1300 to $800, i.e. a drop of 40%. So not only was the ESOP hit by this 40% drop in share price, it also missed out the 250% jump in the esat share price increase. All told, these two things devalued the ESOP by a factor of 75% from what it could have been.

[…] The delay was basically caused by the fact that this was only the second or third ever ESOP handled by the Revenue Commissioners, which meant that everyone was treading on eggshells with regard to getting the legalities right. This meant it all took an age.See, for example, this Irish Times article.

Anyway, all sorts of things were said at that meeting and I really don’t remember it very well, except that everyone was extremely pissed off. Latterly, the ESOP was dissolved and I believe that the people who ended up getting its remains did extremely well for themselves.

It was all a very fine story of corporate and legal incompetence and as in all cases of slow-motion-train-crash style incompetence, no-one was to blame.

Truly, a very familiar story to tell. But the next one is also perhaps a little familiar in its broad sweep.