WEBVTT - Beware Tech Tycoons with Piranha Tanks - with Katie Prescott 0:00:15.410 --> 0:00:25.770 Pushkin. As always, London's Heathrow Airport is heaving. Security queues 0:00:25.930 --> 0:00:30.770 shuffle forward. Exasperated parents trundle suitcases to the terminal with 0:00:30.970 --> 0:00:35.090 toddlers in tow. Excited friends gather in bars to toast 0:00:35.130 --> 0:00:38.450 the trip they're about to take. To it all walks 0:00:38.450 --> 0:00:42.250 a heavy set, balding man in his late fifties. He's 0:00:42.290 --> 0:00:46.090 wearing a smart suit and open shirt. But this man 0:00:46.170 --> 0:00:50.130 is not on a business trip. The people walking politely 0:00:50.210 --> 0:00:55.370 alongside him are not colleagues. They're police officers. They guide 0:00:55.450 --> 0:00:59.210 him through the corridors of Britain's busiest airport and deliver 0:00:59.330 --> 0:01:06.650 him to their American counterparts. Here civility acts. The US 0:01:06.810 --> 0:01:10.770 guards promptly cuff his wrists, wrap chains around his waist, 0:01:10.930 --> 0:01:14.130 and march the man who's more accustomed to champagne and 0:01:14.170 --> 0:01:18.810 private jet to the very back of the United Airlines player. 0:01:19.970 --> 0:01:23.570 It's the eleventh of May twenty twenty three, and Britain's 0:01:23.690 --> 0:01:28.090 answer to Bill Gates is about to be extradited to 0:01:28.170 --> 0:01:35.290 the US on charges of fraud. Once the darling of 0:01:35.370 --> 0:01:39.050 the UK tech world, a member of the BBC's executive 0:01:39.050 --> 0:01:44.610 board and a government advisor, Mike Lynch settles in for 0:01:44.650 --> 0:01:50.730 the eleven hour flight. As London disappears beneath him, he 0:01:50.810 --> 0:01:53.890 knows it will be a very long time before he 0:01:53.930 --> 0:01:59.930 sees Britain again. I'm Tim Harford and you're listening to 0:02:00.050 --> 0:02:29.490 Cautionary Tales. Mike Lynch was the co founder and CEO 0:02:29.770 --> 0:02:34.570 of the UK's biggest software company, Autonomy, which he sold 0:02:34.570 --> 0:02:39.530 to the American hardware giant Hulet Packard for eleven billion 0:02:39.650 --> 0:02:44.490 dollars in twenty eleven. But Hewlett Packard quickly began to 0:02:44.530 --> 0:02:50.010 suspect something was awry at Autonomy. Soon, legal battles raged 0:02:50.090 --> 0:02:53.890 on both sides at the Atlantic, revealing shocking truths and 0:02:54.050 --> 0:02:58.930 incredible oversights. This story of a tech tycoon is a 0:02:59.090 --> 0:03:04.090 rags to rich's tail, a business thriller, and ultimately a tragedy. 0:03:05.050 --> 0:03:07.730 Here to tell us all about it is Katie Prescott. 0:03:08.130 --> 0:03:11.530 She's Technology business editor at The Times and the author 0:03:11.570 --> 0:03:16.650 of the book The Curious Case of Mike Lynch. Katie. 0:03:16.970 --> 0:03:18.210 Welcome to Cautionary Tales. 0:03:18.450 --> 0:03:19.810 Hello, good to talk to you. 0:03:20.450 --> 0:03:23.850 So we should talk about Mike Lynch. His background. First, 0:03:24.690 --> 0:03:29.610 fabulously wealthy at the time we pictured him walking through 0:03:29.690 --> 0:03:32.930 heatho airport. But he didn't start wealthy. 0:03:33.050 --> 0:03:35.810 No, quite the opposite. And actually, you know, in a 0:03:36.210 --> 0:03:42.570 Britain which is still riven by class, Mike Lynch was 0:03:42.730 --> 0:03:47.410 the outlier. So. Born to an Irish immigrant mother who 0:03:47.450 --> 0:03:50.930 was a nurse and a father who was a fireman 0:03:51.130 --> 0:03:55.010 but also a Irish parentage. They moved to Essex when 0:03:55.330 --> 0:03:58.450 he was quite young, and very early on they spotted 0:03:58.970 --> 0:04:02.090 that he was a super bright little boy, and his 0:04:02.130 --> 0:04:05.570 mother put him forward for a scholarship at a private 0:04:05.570 --> 0:04:07.010 school when he was eleven. 0:04:07.210 --> 0:04:10.730 Yeah, so's he's studying with a bunch of rich, or 0:04:10.730 --> 0:04:15.170 at least upper middle class kids whose parents are paying 0:04:15.730 --> 0:04:18.050 large sums of money to send their kids to this school. 0:04:18.130 --> 0:04:21.850 But he's in on merit because he is incredibly sharp 0:04:23.010 --> 0:04:27.770 and academically. Where did his interest really shine? What was 0:04:27.770 --> 0:04:28.650 he doing at school? 0:04:29.010 --> 0:04:32.450 He loved computing and he loved engineering, and I think, 0:04:32.530 --> 0:04:36.770 unusually for a tech entrepreneur, he wasn't a lonely geek 0:04:37.130 --> 0:04:40.490 at all. Actually, far from it. He had lots and 0:04:40.530 --> 0:04:44.250 lots of friends and he led societies. So he started 0:04:44.290 --> 0:04:47.850 an electronics society for example. He also started a jazz band. 0:04:48.090 --> 0:04:51.330 We like that because he loved playing the clarinet, and this. 0:04:51.250 --> 0:04:53.410 Would have been with the late seventies early eighties that 0:04:53.450 --> 0:04:56.890 he was starting these societies. So just as the personal 0:04:57.370 --> 0:05:02.170 computer revolution is beginning and people are getting their z 0:05:02.490 --> 0:05:05.410 X eighty one's or whatever, he's a teenager. 0:05:05.170 --> 0:05:07.650 That's right. He said that there was a place local 0:05:07.730 --> 0:05:09.770 to school that he used to go so he could 0:05:09.770 --> 0:05:14.730 gaze at a commodore sixty four. It was an era 0:05:14.810 --> 0:05:18.730 where the kind of computing revolution was taking off and 0:05:18.770 --> 0:05:22.090 he was really enjoying that as a young boy. 0:05:22.250 --> 0:05:25.130 Yes, there was this sort of sense that you computers 0:05:25.130 --> 0:05:28.130 were things you could take apart and put back together again. 0:05:28.850 --> 0:05:30.690 You could use them, for example, to play. 0:05:30.530 --> 0:05:34.050 Music, and that was his early business actually with a 0:05:34.130 --> 0:05:37.890 mate from school, making synthesizers, so you could kind of 0:05:38.370 --> 0:05:42.290 muck about with music in a way that you couldn't before. 0:05:42.450 --> 0:05:45.770 That's really how he got into the business of tech. 0:05:46.090 --> 0:05:49.730 And then he goes to Cambridge, so this is presumably 0:05:50.450 --> 0:05:51.610 nineteen eighty two, ninety. 0:05:51.370 --> 0:05:53.170 Eighty three, nineteen eighty three. Yeah, that's right. 0:05:53.250 --> 0:05:57.650 So tell us about Cambridge in the early nineteen eighties 0:05:58.410 --> 0:06:01.770 and the connection to the tech industry, because I think 0:06:01.810 --> 0:06:07.450 a lot of people will picture these amazing Gothic chapels 0:06:07.450 --> 0:06:11.210 and these magnificent buildings, and this is where Isaac Newton 0:06:11.810 --> 0:06:14.850 held court. But there was something else going on in 0:06:14.890 --> 0:06:16.530 Cambridge by the time Mike Lynch showed up. 0:06:16.610 --> 0:06:18.770 Yeah, it's an incredibly exciting time I think to be 0:06:18.810 --> 0:06:22.170 in Cambridge if you loved computing and the bit to 0:06:22.170 --> 0:06:25.530 you painted tradition is absolutely right. Mike Lynch went to 0:06:25.650 --> 0:06:31.450 Christ's College, which is Charles Darwin's former establishment, and its 0:06:31.690 --> 0:06:36.090 picture perfect with absolutely pristine lawns and that very famous 0:06:36.130 --> 0:06:42.650 sandstone colored brickwork. But also there was the famous Cambridge 0:06:42.650 --> 0:06:49.090 Computing Lab where people were putting together computers and experimenting 0:06:49.130 --> 0:06:51.050 in this new age a bit like people are today 0:06:51.090 --> 0:06:54.970 with AI. Actually I find the parallels really similar. And 0:06:55.050 --> 0:06:58.890 Mike Lynch was really interested in something called signal processing, 0:06:58.930 --> 0:07:03.490 so this is kind of audio video technology. And he 0:07:03.890 --> 0:07:06.610 had a professor who really took him under his wing 0:07:07.730 --> 0:07:11.890 and convinced him to stay on and do a PhD. 0:07:12.130 --> 0:07:16.250 Yeah, and he was working on noise canceling headphones and yes. 0:07:16.210 --> 0:07:18.290 Exactly all sorts of things like that. And they had 0:07:18.290 --> 0:07:22.090 a project going on to digitize audio as well and 0:07:22.130 --> 0:07:24.450 sort of try and smooth out some of the crackles 0:07:24.970 --> 0:07:26.770 from old gramophone records. 0:07:27.330 --> 0:07:31.290 He doesn't become an academic, he's more of a business person. 0:07:31.570 --> 0:07:32.290 Tell us about that. 0:07:32.410 --> 0:07:35.890 He then started a company called Cambridge Neurodynamics. 0:07:36.250 --> 0:07:38.210 Must be clever because I've got no idea what that means. 0:07:38.290 --> 0:07:40.490 Well, that's exactly why they picked the name. It sounded 0:07:40.530 --> 0:07:42.770 clever and it sort of picked up on keywords at 0:07:42.770 --> 0:07:45.290 the time that were hot in the tech world. 0:07:45.570 --> 0:07:50.410 So this is the early nineties before Google. But one 0:07:50.450 --> 0:07:54.610 of the products he works on for a while is 0:07:54.650 --> 0:07:56.970 the sort of thing that Google turned into a trillion 0:07:57.010 --> 0:07:58.290 dollar business. 0:07:58.530 --> 0:08:01.610 People were trying to work out the best way of 0:08:01.690 --> 0:08:06.570 sifting through vast amounts of digital information. Mike Lynch emerged 0:08:07.050 --> 0:08:12.810 with a company called Auton, and the business sifted through 0:08:12.810 --> 0:08:17.130 information rather than in a clunky keyword search kind of 0:08:17.170 --> 0:08:21.210 way by as Mike Lynch described it, in marketing material, 0:08:21.370 --> 0:08:25.730 looking for concepts. When you're talking about corporations, which is 0:08:25.770 --> 0:08:28.730 where this software ended up being sold to, who've just 0:08:28.770 --> 0:08:32.330 got huge amounts of information. It seemed like a real 0:08:32.690 --> 0:08:33.530 magic bullet. 0:08:33.690 --> 0:08:35.810 It was a huge thing. I mean, those of us 0:08:35.810 --> 0:08:38.170 who are old enough to remember the world before Google, 0:08:38.330 --> 0:08:41.450 but after the world wide web. So these few years 0:08:41.570 --> 0:08:45.010 of there's a lot of stuff on any computer with 0:08:45.050 --> 0:08:48.290 an Internet connection, but there's no way to find it, 0:08:48.850 --> 0:08:51.690 and so we used to swap notes about well, how 0:08:51.770 --> 0:08:54.130 do you search for a thing? And so oh well, 0:08:54.290 --> 0:08:57.130 so one idea is, if there's something there's a particular 0:08:57.450 --> 0:09:01.610 expert in the field, you might search for that expert's name, 0:09:01.890 --> 0:09:05.570 and then web pages which have that expert's name, which 0:09:05.650 --> 0:09:08.610 is going to be unusual, those web pages will probably 0:09:08.610 --> 0:09:12.490 be relevant what you're looking for. That sort of kind 0:09:12.490 --> 0:09:15.930 of hack, rather than just as we would do today, 0:09:16.250 --> 0:09:18.530 to go to Google or to go to an AI 0:09:18.730 --> 0:09:20.330 and just say, oh, I'm just looking for this sort 0:09:20.330 --> 0:09:21.770 of stuff and it will just find you. We take 0:09:21.770 --> 0:09:25.450 it for granted that this stuff will be found. But 0:09:25.650 --> 0:09:29.690 it is an incredibly difficult problem, and a problem that 0:09:30.010 --> 0:09:33.970 Google turned into a trillion dollar business. But Lynch was 0:09:34.010 --> 0:09:35.450 on it before Google even existed. 0:09:35.730 --> 0:09:40.530 It's very hard in today's age to understand quite how 0:09:40.610 --> 0:09:45.410 novel this was. Lynch always said he regretted not selling 0:09:45.410 --> 0:09:48.970 it to the consumer market and not turning it into Google, but. 0:09:49.010 --> 0:09:51.730 Instead so autonomy. So it's not Google, it's not doing 0:09:52.090 --> 0:09:55.170 consumer directed web search. So what sort of search and 0:09:55.210 --> 0:09:57.570 what sort of information retrieval is it doing and who 0:09:57.650 --> 0:09:58.210 is paying for it? 0:09:58.410 --> 0:10:01.970 All sorts of enormous corporations in the end, as well 0:10:02.010 --> 0:10:06.690 as organizations within governments around the world. So rather than 0:10:06.730 --> 0:10:11.930 finding information in neat databases, look through your emails, voicemails, 0:10:12.370 --> 0:10:16.610 things like word documents, this kind of enormous trove of 0:10:16.650 --> 0:10:22.090 information that corporations create. And that's who he was selling 0:10:22.130 --> 0:10:24.570 it to, which made sense as a market at the time, 0:10:24.570 --> 0:10:27.130 because it seemed incredible They've got the money right, very 0:10:27.210 --> 0:10:30.490 lucrative and a more obvious place to sell something to 0:10:31.170 --> 0:10:32.570 than the consumer market. 0:10:32.730 --> 0:10:37.650 So Autonomy, Mike Lynch's company in the late nineteen nineties 0:10:37.810 --> 0:10:41.170 is jumping into this space where there's this opportunity because 0:10:42.010 --> 0:10:45.810 corporations have started generating enormous amounts of digital information that 0:10:45.890 --> 0:10:48.530 is unstructured, there's no way to make any sense of 0:10:48.530 --> 0:10:52.690 the stuff, and Autonomy does that and it's a very 0:10:52.690 --> 0:10:56.210 successful company. So what was it like to work at Autonomy? 0:10:59.170 --> 0:11:01.810 I think what sums up the atmosphere was the Piranha 0:11:01.930 --> 0:11:06.690 tank in reception, something not something you often see in 0:11:06.770 --> 0:11:10.690 the roundel of the Autonomy logo, and Lynch used to 0:11:10.770 --> 0:11:13.890 joke that that was the tank that he would throw 0:11:13.970 --> 0:11:15.370 failing salespeople into. 0:11:15.730 --> 0:11:16.090 Charming. 0:11:16.370 --> 0:11:18.970 It was a very very difficult place to work. Mike 0:11:19.090 --> 0:11:23.450 Lynch was a very hard taskmaster. He could be brutal, 0:11:23.770 --> 0:11:27.090 he could be very aggressive, and I'm afraid I heard 0:11:27.490 --> 0:11:30.210 very early on from researching the book some real horror 0:11:30.210 --> 0:11:33.690 stories about how people were treated within Autonomy, and that, 0:11:34.250 --> 0:11:38.050 I think, because the tech world in Britain is very small, 0:11:38.250 --> 0:11:41.050 much much smaller at that time, leaked out very quickly 0:11:41.130 --> 0:11:42.930 and it got Mike Lynch's certain reputation. 0:11:43.050 --> 0:11:44.970 When you say a certain reputation, I mean that all 0:11:45.050 --> 0:11:47.170 kinds of things go on in companies that we could 0:11:47.170 --> 0:11:51.090 describe as horrible. But what was he actually doing. 0:11:51.530 --> 0:11:55.650 There was one woman who worked in the marketing team 0:11:55.890 --> 0:12:00.490 who he thought wasn't working hard enough, and he would 0:12:00.570 --> 0:12:04.170 yell at her repeatedly, and after an incident when he 0:12:04.210 --> 0:12:08.050 saw her and some colleagues mocking him for his James 0:12:08.130 --> 0:12:10.410 Bond obsession, because I loved James Bond, and he named 0:12:10.410 --> 0:12:13.410 all of the meeting rooms after James Bond villains. He 0:12:13.490 --> 0:12:16.410 put her desk first inside his office and then just 0:12:16.490 --> 0:12:18.650 outside so he could look over her shoulder. The whole 0:12:18.650 --> 0:12:20.930 time and see what she was doing. There were tons 0:12:21.010 --> 0:12:23.450 and tons of examples of people being fired at will. 0:12:24.130 --> 0:12:27.250 People were very scared of him, so. 0:12:27.370 --> 0:12:31.250 Very demanding, cool, mean, angry. 0:12:30.930 --> 0:12:34.050 Bullying, and I think I'm afraid it was deliberate cruelty. 0:12:34.570 --> 0:12:38.250 I think he enjoyed the power. He had a tiny 0:12:38.290 --> 0:12:42.530 team around him, and I got the impression they were 0:12:42.610 --> 0:12:46.290 like his henchmen. And that's how the company was run. 0:12:47.210 --> 0:12:50.010 So how does it do financially? Because it grew very 0:12:50.050 --> 0:12:52.930 fast and it was floated on the London stock Market 0:12:53.010 --> 0:12:55.530 and became a foot Sea one hundred companies, So this 0:12:55.570 --> 0:12:59.810 is one of the one hundred biggest companies listed on 0:12:59.810 --> 0:13:02.090 the London Stock Exchange. So that sounds good. 0:13:02.490 --> 0:13:03.810 It does, and I think we have to be very 0:13:03.810 --> 0:13:06.770 careful here. It's certainly benefited from the dot com boom. 0:13:07.170 --> 0:13:11.170 It actually had a very ino it software product. However, 0:13:12.050 --> 0:13:14.810 at the time that it went into the Footsy one hundred, 0:13:15.090 --> 0:13:19.250 that's because its valuation, so what shareholders vestors thought it 0:13:19.370 --> 0:13:22.290 was worth was way above the money that it was 0:13:22.330 --> 0:13:23.090 actually making. 0:13:23.210 --> 0:13:24.690 People were very very excited. 0:13:24.730 --> 0:13:27.050 I think it was making something like forty four million 0:13:27.170 --> 0:13:30.690 pounds a year and it was valued at four billion 0:13:30.810 --> 0:13:34.210 pounds because people were very excited about its future prospects. 0:13:34.890 --> 0:13:39.290 And we're seeing that similar valuation difference in AI businesses today. 0:13:39.810 --> 0:13:43.850 So yes, it was doing well for a startup, but 0:13:43.930 --> 0:13:46.530 its revenue certainly didn't match up to its valuation, and 0:13:46.570 --> 0:13:48.850 it very quickly actually fell out of the Footsy one 0:13:48.930 --> 0:13:50.770 hundred when the dot com bust happened. 0:13:51.090 --> 0:13:54.170 I would love to talk about the bust, but let 0:13:54.250 --> 0:14:05.250 us do that after the break. We are back and 0:14:05.370 --> 0:14:08.730 I am talking to Katie Prescott, the author of the 0:14:08.850 --> 0:14:13.330 que Serious Case of Mike Lynch. Katie, we promised to 0:14:13.370 --> 0:14:17.770 talk about the dot com boom and bust. Before we do, 0:14:17.890 --> 0:14:21.410 we should probably introduce a new character, and that is 0:14:21.930 --> 0:14:26.930 the Reverend Thomas Bays. Lynch was obsessed with him. So 0:14:27.010 --> 0:14:30.570 what did he do and why was it relevant to 0:14:30.650 --> 0:14:31.330 Mike Lynch. 0:14:31.490 --> 0:14:35.730 He was an early academic in the field of probability. 0:14:36.250 --> 0:14:39.850 He came up with this concept that sounds very simple, 0:14:40.050 --> 0:14:44.250 that what has gone before informs the probability of something happening. 0:14:45.010 --> 0:14:49.210 So if you walked into a room where people had COVID, 0:14:49.490 --> 0:14:51.450 it is more likely that you would come out of 0:14:51.490 --> 0:14:57.730 it with COVID, and Lynch like embrace this both as 0:14:57.770 --> 0:15:00.690 part of his business, but he also, i would say, 0:15:00.730 --> 0:15:04.370 mainly used it to inform his philosophy of life. Right, 0:15:04.490 --> 0:15:08.130 So he was constantly weighing up the odds of things happening, 0:15:08.770 --> 0:15:12.330 and his teams were very used to him saying I 0:15:12.410 --> 0:15:16.290 put twenty percent on that, eighty percent on that. If 0:15:16.290 --> 0:15:19.050 he was hiring someone, he'd say, Okay, who are they 0:15:19.490 --> 0:15:22.650 like that? We know, yes, just trying to work out 0:15:22.730 --> 0:15:25.610 the odds of things happening. And one of the things 0:15:25.650 --> 0:15:28.610 that struck me in writing this book is that everything 0:15:28.650 --> 0:15:32.650 about Lynch's life was completely improbable. And the working title 0:15:32.650 --> 0:15:35.970 of the book was actually improbable Odds because it's so 0:15:36.090 --> 0:15:40.690 unlikely that someone from his background would go to private school, 0:15:40.770 --> 0:15:44.290 would go to Cambridge. He was born with no fingerprints. 0:15:44.650 --> 0:15:47.370 Most startups fail, I don't know there are There are 0:15:47.450 --> 0:15:49.370 so many improbable things. 0:15:49.370 --> 0:15:51.290 He was born with no fingerprints. He slipped that one in. 0:15:51.850 --> 0:15:53.610 He was born with no fingerprints. 0:15:53.690 --> 0:15:55.570 Wow, he really is a Bond villain. 0:15:55.690 --> 0:15:58.690 Okay, yeah, And that was part of the joke. He 0:15:58.770 --> 0:16:00.690 used to say to analysts in the city. Look, I 0:16:00.690 --> 0:16:02.770 have got no fingerprints. I'd make a brilliant criminal. 0:16:03.090 --> 0:16:07.570 So he's this brilliant guy. He's a bully. He has 0:16:07.650 --> 0:16:11.290 set up this business, he's launched it on the stock market. 0:16:11.610 --> 0:16:16.850 So tell us then, about what happened when Autonomy ran 0:16:16.930 --> 0:16:18.490 into the dot com bust. 0:16:18.730 --> 0:16:23.410 Lynch was only thirty at this point, running a startup 0:16:23.530 --> 0:16:27.890 out of a coworking space in Cambridge, and it went 0:16:28.170 --> 0:16:31.210 from footsy one hundred company too, and you can see 0:16:31.210 --> 0:16:35.770 the share price plummeting, and Lynch really wanted to work 0:16:35.890 --> 0:16:39.650 to get that share price up again. Now, on one level, 0:16:40.050 --> 0:16:43.170 it was still doing well and still making the sales 0:16:43.250 --> 0:16:46.530 that it was and its technology was still valued. Just 0:16:46.570 --> 0:16:49.170 because it had dropped out of the list of shares 0:16:49.250 --> 0:16:51.090 hadn't changed what it was doing. 0:16:51.170 --> 0:16:53.130 I mean, it's worth being clear. The reason it was 0:16:53.130 --> 0:16:55.770 a footsy one hundred company was because it was valued 0:16:56.010 --> 0:16:58.610 by investors very highly. And the reason it was valued 0:16:58.610 --> 0:17:01.290 by investors very highly was not because it was making 0:17:01.450 --> 0:17:04.650 enormous profits. It was because they were very optimistic about 0:17:04.650 --> 0:17:08.930 its future prospects. So when they become less optimistic about 0:17:08.930 --> 0:17:11.130 all these tech companies, it drops out of the foots 0:17:11.250 --> 0:17:11.690 one hundred. 0:17:11.690 --> 0:17:14.490 But it's not because it's doing anything wrong. 0:17:14.610 --> 0:17:15.850 Yeah, it's not done anything wrong. 0:17:15.930 --> 0:17:19.290 All the big tech companies did drop at that point 0:17:19.290 --> 0:17:22.570 to me, even the likes of Microsoft. This is the 0:17:22.610 --> 0:17:27.650 first time that Lynch faced the reality of running a 0:17:27.690 --> 0:17:30.010 public business. You've got the glory of being in the 0:17:30.010 --> 0:17:32.370 foot seat one hundred, of being, you know, the young 0:17:32.450 --> 0:17:36.370 cool guy in Britain's tech scene during the dot com boom, 0:17:36.570 --> 0:17:39.130 but then during the bust. Actually, you've got the pain 0:17:39.650 --> 0:17:43.090 of trying to convince the press and trying to convince 0:17:43.410 --> 0:17:46.890 people in the city who are scrutinizing your company and 0:17:46.930 --> 0:17:49.290 telling investors whether they should put their money into it. 0:17:49.770 --> 0:17:53.170 You've got the pain of really trying to get them 0:17:53.210 --> 0:17:57.010 on board. And he really hated criticism, especially when it 0:17:57.090 --> 0:18:00.170 came to Autonomy. He had very very thin skin. 0:18:00.770 --> 0:18:02.970 He is subject to a level of scrutiny that he's 0:18:03.010 --> 0:18:05.810 not used to. He doesn't like it, and he wants 0:18:05.850 --> 0:18:10.610 to prove that Autonomy is for real. There's a certain 0:18:10.730 --> 0:18:14.930 pressure to make the numbers do what Mike Lynch wants 0:18:14.970 --> 0:18:16.810 them to do at this point, so. 0:18:17.170 --> 0:18:22.770 Every quarter Autonomy would publish their results publicly. 0:18:22.890 --> 0:18:25.130 This is what we sold, this is what it costs. 0:18:25.250 --> 0:18:27.330 Exactly right, so how much money we made And kind 0:18:27.330 --> 0:18:30.970 of unusual for a British company, because reporting every six 0:18:31.010 --> 0:18:34.490 months here is more normal. But Autonomy had been listed 0:18:34.490 --> 0:18:38.130 on the Nasdaq and Lynch always wanted to make parallels 0:18:38.210 --> 0:18:42.010 with his company and the very successful tech businesses in 0:18:42.050 --> 0:18:46.050 the US, so he was forced every three months, or 0:18:46.130 --> 0:18:48.770 he had chosen every three months to publish. 0:18:48.850 --> 0:18:51.290 Their results as a condition of Nasdaq listing. 0:18:51.610 --> 0:18:56.690 So he had to not only keep up with the 0:18:56.810 --> 0:19:01.010 forecast or the predictions that Autonomy made about their expectations 0:19:01.050 --> 0:19:04.570 for their sales, but also the predictions that would be 0:19:04.650 --> 0:19:08.610 made by city analysts in London who were making their 0:19:08.610 --> 0:19:13.890 own predictions, and really that set a target every three 0:19:13.930 --> 0:19:17.570 months to hit, and when Autonomy didn't hit that target, 0:19:17.730 --> 0:19:19.370 its share price would fall. 0:19:20.250 --> 0:19:21.730 He didn't want that to happen, and he. 0:19:21.730 --> 0:19:24.970 Didn't want that to happen. If you fast forward towards 0:19:25.050 --> 0:19:29.010 the financial crisis of two thousand and eight, companies stopped 0:19:29.010 --> 0:19:32.210 buying software in the way they had done because budgets 0:19:32.250 --> 0:19:38.370 were squeezed, understandably, But to continue hitting those targets, Autonomy, 0:19:38.410 --> 0:19:41.890 which had always prided itself and boasted about being a 0:19:41.930 --> 0:19:47.090 pure software company, started selling hardware and reselling other people's 0:19:47.090 --> 0:19:51.730 hardware from companies like Dell, So things like keyboards and laptops, 0:19:51.850 --> 0:19:53.530 very basic things mice. 0:19:53.730 --> 0:19:55.410 That doesn't sound like a high margin business. 0:19:55.570 --> 0:19:57.810 Well no, and that's the problem because analysts in the 0:19:57.810 --> 0:20:03.410 city and investors liked Autonomy's pure software model because it 0:20:03.490 --> 0:20:08.170 had an enormous margin because you're not paying to manufacture 0:20:08.250 --> 0:20:11.770 something physical. When you're selling software, it's the dream. You 0:20:11.850 --> 0:20:14.650 create some IP and then you can just ping it 0:20:14.690 --> 0:20:18.370 around and so your margins are massive. But clearly when 0:20:18.410 --> 0:20:21.770 you're reselling hardware that you've bought from somebody else, that's 0:20:21.890 --> 0:20:26.610 very different. And Autonomy bundled all these sales into one 0:20:26.650 --> 0:20:31.970 set of numbers and still told investors that they were 0:20:32.010 --> 0:20:35.570 a pure software business, but actually a chunk of those 0:20:35.650 --> 0:20:40.210 sales in one quarter, the biggest time that this happened, 0:20:40.250 --> 0:20:43.730 twenty percent of those sales came from reselling hardware. 0:20:44.330 --> 0:20:46.410 Just explain to me in a nutshell, what were Mike 0:20:46.490 --> 0:20:49.970 Lynch and his team doing to make Autonomy's numbers look better. 0:20:50.450 --> 0:20:53.730 So they were essentially wrapping the sales of more expensive 0:20:53.770 --> 0:20:57.890 hardware into their software sales and putting expenses that came 0:20:57.930 --> 0:21:02.050 from that in the marketing budget, moving numbers around around 0:21:02.050 --> 0:21:07.090 the accounts. They were clocking sales on their books before 0:21:07.130 --> 0:21:11.490 they'd actually happened by using resellers, which are very common 0:21:11.690 --> 0:21:15.010 in the software industry, but they should really have had 0:21:15.050 --> 0:21:18.970 a deal signed before they passed it on to the reseller, 0:21:19.050 --> 0:21:22.570 and so that created holes in the accounts every quarter 0:21:23.130 --> 0:21:26.930 that they struggled to fill. And then also they were 0:21:27.290 --> 0:21:32.530 writing down sales as sales to original equipment manufacturers or OEMs. 0:21:33.050 --> 0:21:36.170 So they'd say, we're putting Autonomy software in Dell, but 0:21:36.290 --> 0:21:40.050 they were calling all sorts of companies OEMs, including Tottenham 0:21:40.050 --> 0:21:42.610 Hotspur Football club. Linch. She used to describe it as 0:21:42.650 --> 0:21:45.210 the Pinacolada model. He'd sell to an OEM and then 0:21:45.210 --> 0:21:47.730 you could sit back on the beach with your cocktail 0:21:48.050 --> 0:21:49.330 while someone else sold it for you. 0:21:49.690 --> 0:21:56.130 And I'm curious. Autonomy had auditors, professional accountants, independent coming 0:21:56.210 --> 0:21:58.850 and looking at their accounts. What were the auditors doing. 0:21:59.650 --> 0:22:03.210 The auditors were signing off the accounts behind the scenes, 0:22:03.250 --> 0:22:06.530 there were often questions about how things were being registered 0:22:06.570 --> 0:22:11.890 as sales. The auditors, Deloitte were later fined over the 0:22:11.930 --> 0:22:16.530 Autonomy accounts. There was an extraordinary investigation by the accounting 0:22:16.610 --> 0:22:21.370 watchdog into how they handled them, which essentially found that 0:22:21.970 --> 0:22:25.530 Deloitte Cambridge, which was a branch of the main Deloitte 0:22:25.570 --> 0:22:27.690 office in London. Bear in mind they're one of the 0:22:27.690 --> 0:22:32.090 top four accountants, wanted to keep its relationship with Autonomy. 0:22:32.090 --> 0:22:34.250 It was its only footzy one hundred. 0:22:33.930 --> 0:22:35.850 Customer keep the customer satisfied. 0:22:35.930 --> 0:22:38.730 They wanted to keep the customer satisfied. And the relationship 0:22:38.930 --> 0:22:40.930 was just too cozy with the company. 0:22:41.330 --> 0:22:43.970 And so people looking at these numbers are just going 0:22:44.010 --> 0:22:48.530 to fundamentally misunderstand what is actually going on inside the company. 0:22:48.770 --> 0:22:51.850 Yes, and there was one particular analyst in the city, 0:22:51.970 --> 0:22:55.490 a chap called Dowd Khan, who wrote a rather negative 0:22:55.570 --> 0:23:01.250 note about Lynch, and this sparked a very bizarre reaction. 0:23:02.170 --> 0:23:07.370 He ended up banning Dowd Khan from Autonomy's meetings. From 0:23:07.410 --> 0:23:09.770 their results meetings, so normally they didn't oounce their results. 0:23:09.810 --> 0:23:12.170 They invite everybody in an analyst would be able to 0:23:12.210 --> 0:23:15.490 ask questions, and he said, you're you're not allowed to come. 0:23:15.650 --> 0:23:17.570 This is the equivalent of the White House saying we 0:23:17.570 --> 0:23:20.170 don't want a reporter showing up to the presidential briefings 0:23:20.170 --> 0:23:21.970 because we don't like the way you write your news. 0:23:22.090 --> 0:23:24.530 That a similar sort of energy yes. 0:23:24.330 --> 0:23:28.850 It was. It was very strange, and a little group 0:23:29.010 --> 0:23:32.890 formed of it was quite maverick analysts in the city 0:23:32.970 --> 0:23:37.330 who started to question Autonomy's numbers and be quite vociferous 0:23:37.330 --> 0:23:37.850 about it. 0:23:38.050 --> 0:23:41.930 So they're cutting various corners in terms of their accounts. 0:23:42.290 --> 0:23:44.770 Did they need to do this? Was it existential that 0:23:44.850 --> 0:23:46.570 basically they were they were just going to run out 0:23:46.570 --> 0:23:49.250 of money or run out of funding if they hadn't 0:23:49.850 --> 0:23:53.690 stepped across the line occasionally, or was this entirely Mike 0:23:53.770 --> 0:23:54.570 Lynch's vanity. 0:23:55.130 --> 0:23:57.610 I think there were definitely problems with the technology. They 0:23:57.650 --> 0:24:00.930 weren't investing enough in research and development, and there were 0:24:01.410 --> 0:24:05.810 very serious competitors. Google had created a version of the 0:24:05.850 --> 0:24:09.410 technology that was much much cheaper, but there was still 0:24:09.450 --> 0:24:13.130 selling it. And actually there are some bar charts in 0:24:13.170 --> 0:24:19.170 the book which show how much they inflated the revenue 0:24:19.410 --> 0:24:25.250 each quarter, and you're not talking about enormous sums. They 0:24:25.330 --> 0:24:29.370 probably could have kept going and actually probably pivoted to 0:24:29.450 --> 0:24:32.370 other sorts of technology. So Lynch was very very early 0:24:32.450 --> 0:24:37.210 to the cloud, but he was so very proud. 0:24:37.610 --> 0:24:40.370 Well, he had a lot to be proud about. So 0:24:40.890 --> 0:24:45.730 he'd created this very successful company. Is a pillar of Society. 0:24:45.810 --> 0:24:48.930