Getting to "yes" in a world of "no"…

Archive for December, 2011

Is it time for a “Small Manufacturing Tsar”?

Few people have heard of the Tech City Investment Organization, and even fewer realize that one of the main reasons it was set up was to encourage overseas companies to park themselves in Shoreditch, to “create more employment more quickly”. Hence the eyecatching entrance of Google and Cisco into the whole Tech City parade: of course, what we’ll do once all the web programmers in town suddenly disappear into their corporate maws is another matter… so yes, that was a great idea, Dave. ‘Great‘, as in… ‘utterly awful‘.

Of course, it’s true that TCIO would also like to encourage angel investors to invest in Tech City startups, but… have you seen any doing this yet? [No, neither have I]. And much as I enjoy chatting with Silicon Valley Bank’s smart & engaging Oscar Jazdowski, trying to apply SVB’s heavyweight expansion capital funding model to Tech City’s frothily lightweight, early stage social media startups is not unlike trying to crack a nut with a pneumatic Hammer-O-Tronic 9000. Well, you could… but, frankly, you’d probably lose your nuts in the process.

As normal, you don’t have to scratch very far beneath the surface hype to see the real problem. For even if you don’t remotely believe TCIO’s apples-for-mangos statistics for the allegedly dramatic expansion of tech startups in Tech City (and I for one don’t), it’s surely terrifically unfair that those [quite literally] poor darling hotdeskers have such an extraordinary burden of government rhetoric dumped on their collective shoulders. Much as I love them as individuals, how exactly are these self-funded software engineers supposed to change the fortune of UK plc? Or is it rather more likely that 95% of their startups will pivot, pivot, stutter and fall, leaving their principals to drop their principles & grab jobs with the webbed-up corporates now conveniently situated next door?

It’s a little bit galling that almost all the attention goes to Tech City, when there has (according to other recent [and probably more reliable] statistics) been a huge surge in startup activity right across the London area. I’m not blaming TCIO’s Tech City tsar Eric van der Kleij for this, but I can’t help but notice that as soon as Tech City gets mentioned in the media, it’s as if everyone’s suddenly trying to behave like an idiot venture capitalist circa 1995, but crossed with all the idiot parts of the Lean Startup circa 2011: people trying to fund VC wetly disintermediated dreams with pin money. Does that notion make any sense to you? Nope, I didn’t think so. 😦

If we can somehow manage to resist the temptation to inhale too deeply from the Tech City bong, the big question is this: what would actually make a difference? What positive thing could the Government actually do? Well…

What about a ‘Small Manufacturing Tsar’?

Being realistic, to make a macroeconomic difference you’d need at least a hundred startups to go totally platinum – to become billion dollar export-heavy industries all in themselves. And that’s a big ask, seeing as barely any UK companies have managed this trick of late.

Yet I think one sector that stands a finite chance of achieving this is modern lean manufacturing, and – not for the first time in the last few years, it has to be said – the world viewed from my seat appears to be radically different from what Whitehall seems to be seeing.

For me, modern manufacturing is totally rock & roll, and offers a real chance of making a macroeconomic difference to exports, tax revenues, and to UK plc in general. Electronics development (and even mechatronics development) has gone garage, while hardware startups are just starting to go properly guerilla too.

Yet many people – even otherwise well-informed and erudite commentators – still struggle to see the differences between, say, British Leyland circa 1975 and a miniaturized mechatronic startup circa 2011, when (apart from the “both making things” bit) you’d have a lot of trouble finding any similarities. I hate to say it, but a lot of this is just academic snobbery, self-appointed faux-elitist nonsense from people who excel at engaging with theoretical knowledge, but disdain getting involved with anything remotely practical. Score one (own goal) for academe, sadly.

The central communication problem here is that small manufacturing has no obvious champion, no clear focus, no accepted role models – the closest Vince Cable gets is when he talks up Brompton Bikes (what about Grahame Herbert and his Airframe folding bike, then? 😉 ).

Hence what I think the UK needs right now is a ‘Small Manufacturing Tsar‘ – someone who can champion, focus and promote small UK manufacturing, to help the country’s small manufacturers get the mindshare and attention they need. They’re not British Steel and British Aerospace, they’re small (yet hugely scalable) companies who don’t happen to fit the kind of startup pigeonholes and templates that funding bodies like the Technology Strategy Board would like them to conform to. They are, as per the title of a recent survey, “Born Global” – because proper manufacturing may now start small and ‘Lean’ (in the correct sense of the word), but it starts with a foot in every country. And that’s… quite a lot of feet.

And so here’s my open question to the Business Secretary: Mr Cable, do you have any current plans to champion, focus and promote the small manufacturing sector? Might I suggest that a Small Manufacturing Tsar might help to do this?

Risk-adjusted money, cars, supply chains, startups, and why the Internet is killing us all…

Here’s a big insight I want to share with you, something which could change how you think about startups and finance: but all the same, as with my talk on Dangerous Reasons, you probably won’t like it

For all the obvious reasons, I don’t tend to watch a lot of TV: but Storyville’s “Inside Job” (by Charles Ferguson) a few days ago was irresistable viewing and set me really thinking. The way Ferguson tied a succession of leading “economists” and their universities in deservedly squirmworthy knots was exquisitely masterful: but for me, the star turn of the whole show was the former chief economist (2003-2007) of the International Monetary Fund Raghuram G. Rajan, a man so utterly modest he doesn’t even appear in the Inside Job IMDb credits.

What Rajan presented to an elite conference in September 2005 [timecode 38:30 – 40:40] was a paper called “Has Financial Development Made The World Riskier?” (hint: the answer is ‘yes, it bl**dy has’). His idea was that even though investment bankers were being incentivized to do big deals that increased their companies’ risk exposure, they were not being compensated in risk-adjusted money. Let’s say your company is worth £10m and currently has a 10% chance of failing: you’re now offered a deal that’s worth a further £10m and gives you personally a huge bonus today but increases your company’s risk of failure to 50%… would you take it? It would seem that lots of people on Wall Street basically answered ‘yes’ to this question… all of which led to the overleveraged, hyper-inflative, non-‘AAA’-securitized fake mortgage context for the 2008 Credit Crunch etc. That is, they weren’t making more money while taking less risk, but making more money while taking more risk. If individual agents had been compensated in what Rajan calls “risk-adjusted money” for such insane deals, the bonuses would have flowed in the reverse direction (and 20x greater magnitude, too).

While I was talking with my friend Alan Ferdman yesterday, I was struck by a strong parallel between Rajan’s insight and the car industry. Everyone seems to have a car mechanic story about how fancy-looking car X suddenly developed a computer fault Y and had to be scrapped rather than fixed. That’s an example of fragility: how the risk of electronics failure reduces the effective value of a thing. If roughly 25% of a certain car model will develop a computer fault within 2 years (but you’ll never know in advance until it just happens to happen to you), then that’s an example of an object having a different risk-adjusted value. Alan mentioned how Keith Armstrong had discussed whether a short-lived EMI (electromagnetic interference) fault could have led to Toyota’s “Sudden Unintended Acceleration” problem: fascinating, awful, and terrifically fragile, whatever the actual cause.

Supply chain fragility affects us all too, whether via oil affected by civil war in Libya, hard drive component factories in Thailand destroyed by a tsunami, or a PCB materials factory in North-East Japan hit by a nuclear evacuation. In the MBA-optimized Just-In-Time world we live in, we are so (virtually) close to the producer that the cost of the things we need becomes inherently more volatile: how can we give peas a chance when we don’t know if we’ll still be able to afford them tomorrow? Or eggs? Or beef? Or bread?

Weaving all these strands into a single picture, and I think you’ll find an economics mega-gauntlet being thrown down in front of you. Might it be true – the challenge goes – that for almost all companies, economic development in real terms has now ceased? If a company needs to take on substantially more risk in order to grow, is that company really growing? In risk-adjusted money terms, the answer is that it probably isn’t – it’s merely appearing to grow. It would rather live in more risk than stay as it is: but that way lies chaos, mayhem and disaster.

As financial managers, we are supposed to deal with this by factoring in contingent liabilities, not only real (contractual) ones in our actual accounts but also predicted exposure scenarios in our management accounts. However, for all the supposed rigour of risk management, I think it’s actually almost impossible now to conceive and generate a sufficient number of scenarios to cover all manner of supply chain, finance, channel, and customer fragilities. If the world has apparently become richer while simultaneously becoming more fragile (far beyond merely ecological viewpoints), has it actually become richer, or is it the same, or perhaps even worse?

The whole Just-In-Time thing is simply a way of concentrating, multiplying and sharing fragilities, rather than reducing fragility. Yet that is the channel distribution model the world is increasingly moving towards, in terms of mutualizing the optimized returns on its deployment (in the form of lower customer prices) yet exposing everyone in the chain to more fragility. We save money, but the price of our electricity and petrol is suddenly hugely volatile.

In a lot of ways, it would be useful if we could price up our startups in terms of risk-adjusted money: in many ways, that’s what VCs and lead angels do when they negotiate valuations with entrepreneurs. However, the very ground upon which we build our startup shacks has itself become fragile: one of the first things that happened to my own startup was that the fabless Korean company making the image sensor I planned to use went under – the vast chip foundry making its sensors ran out of money (so the story went), and without any hope of production it folded too. The problem is that if we can’t conceive the ten thousand ways in which the invisible infrastructures we implicitly yet absolutely rely on – water, air, broadband, staples, paper, biros, hard drives, RAM chips, sensors, cables, lead-free solder, whatever – can fail, we can’t possibly price up such pervasive risk. Perhaps this is the real reason startup funding dealflow choked: the inability to price up infrastructural fragility.

Arguably, an even bigger problem is that almost all the social media and digital media startups I see being developed, talked about, and talked up seem to based on an even bigger folly: that there is an infinite amount of ‘juice’ to be squeezed out of the systems around us, by using social hacks, GPS hacks, network hacks, recommendation hacks, gift hacks, etc. But if you stop to think this through this for even a moment, you’ll see that the implicit claim here is that startups can somehow disintermediate the economy out of its dark place: that we can macroeconomically grow by cutting out middlemen (middlewomen too, presumably) or channel owners.

And so I think Internet discourse is killing us: it promotes the idea that we should be applying our ingenuity to shortening and tightening supply chains (i.e. increasing fragility, increasing systemic risk) rather than trying to reduce risk and fragility. We should instead be making everyone rich by reducing fragility, yet the only obviously applicable skills from the MBA smorgasbord seem to increase it. As a design, engineering and finance community, we’ve lost the knack of building worthwhile things – of even trying to judge what is worthwhile and what is merely fragile.

Even though the dismally low-aim Lean Startup community tries to claim that the important question is “Pivot or Persevere?“, I think the only question of real worth is “Worthwhile or Fragile?” Make money, sure, but stop obsessing on optimizing social hacks and instead start building things that make a positive difference – things that make the world less fragile, not more. Do you know how?

(Again, I told you you wouldn’t like it.)

Applying “Five Whys” to The Lean Startup itself…

The biggest shadow hanging over nearly every business book is the ghost of unwarrantable universalism. Which is a fancy way of saying that, whenever any business writer claims “Methodology X worked for me, so it will work for you [despite the fact that your situation is almost certainly completely different]“, you know that pretty much every bullshit warning bell in your head should go off.

To be precise, this is because every single business success story is a particular historical narrative (i.e. of what actually happened), not a general scientific template (i.e. of what might happen in a hypothetical business): and in fact, given the terrifically small number of startups that achieve success with a high-growth development curve (often with numerous strokes of luck generously helping them on their way), such rare businesses are almost all outliers in the business landscape. Yet the temptation to post-rationalize all that ‘outlieritudinosity’ and to promote the particular approach you took as being a universal entrepreneurial path to success is apparently overwhelming: for if bookshops were taxed on the number of glowingly-confident head-shots of cashed-out entrepreneurs they had on their shelves, they’d surely all fold within a day, mmm-hmm?

As far as Lean Startups go, Eric Ries is a smart, sharp guy, one who is amply self-aware and reflective enough to know all the above… but he still persists in making universal claims. Using one of his favourite Lean techniques (the “Five Whys”), we might therefore probingly ask a series of five ‘Why’ questions, aiming to get to the root of this issue rather than getting mired in the surface details.

Firstly, why is he so certain that his consulting recommendations for startups (which in turn were built largely on his personal experience at a single company, IMVU) are well founded enough to drive a global ‘movement’?

I suspect the answer is that Ries is mainly talking about Internet businesses with a framework you can use to carry out rapid customer experiments. Even though he used to talk the Lean Startup up as an extension of Steve Blank’s Customer Development, Ries’s book actually seems to have pivoted round to the quite different idea that the customer doesn’t know best, but performing lots of A/B experiments on them helps you learn from their aggregate behaviour what path to follow. So, Ries’s certainty seems to stem from his experience at IMVU using the process of engineering statistical experiments on microtransactions (even if the early-pipeline currency is ‘attention’ rather than ‘dollars’) to learn genuine data about its customers, and his belief that it was this process that led to its success.

But why (the second why of five) does he think his personal experience at IMVU just happened to give him a unique view of startup business truth? From his account, IMVU made plenty of mistakes early on and was manifestly not customer-led or even really customer-engaged: Ries’s eventual epiphany seems to have been the realization that his company was in a position to use tech to improve itself, not just its product. And so the answer to [why #2] would seem to be that Ries’ view of the Lean Startup was as something that uses tech to institutionalize a kind of ‘customer awareness therapy’, i.e. that numbers from experiments can help your company ‘Get Real’. Essentially, it’s a tech answer to a social awareness problem that enabled IMVU to drop its stealth-mode conceptual baggage and get on with the real business of engaging with people and satisfying their needs with avatars that don’t walk. 🙂

But why (3/5) does Ries think his particular tech solution is a universal solution to the social problem of customer development? I suspect the honest answer would be that this opinion came from his background as a CTO: the answer he developed relied on appropriating tech tools to develop knowledge of IMVU’s customers. He was apparently amazed to find that the tech tools he just happened to have in his hands already were able to help his company get what it wanted, and this amazement has somehow driven his quest to help other people use their own tech tools to get what their companies want. It’s the joy of techs: a kind of engineer empowerment exercise.

But why (4/5) does Ries think engineers should be empowered in this way? Again, given that he started as an engineer and that most of his effort over the last few years seems to have been involved outreaching to countless other engineers, it seems that his Lean Startup movement is actually just a collective business land-grab by software engineers. By trying to reduce social problems to technical problems, the idea is to turn everything into an engineering exercise, a conceptual business structure where you don’t need to know about product or marketing or PR or personal relationships, your customer experiments will (eventually) tell you all that, by way of pings on a learning metric chart. As a result, it’s resolutely anti-theoretical, and indeed anti-pretty-much-everything-that-isn’t-engineering: even his “anti-waste” angle is supported only by a few carefully selected narrative tales of the unexpected. Does he have any real data to back up his assertion that non-Lean-Startup businesses are wasteful? What he manifestly fails to mention is that Lean Startup development often takes a lot, lot longer: sure, you may waste less physical resources, but you can end up wasting a vast amount of time to do so. Time never really gets factored into the economics equation: it’s somehow assumed you have plenty of it. (Errr… no, you don’t].

But why (5/5) waste all this effort on driving what is essentially a business land-grab by engineers? The biggest universal claim implicit in Lean Startup is that startups only need engineers: and that reeks to me of a parochial engineering-centric mindset being extended far beyond its genuine validity. I suspect what Lean all comes down to is that Eric likes engineers, and wants them to prosper: but surely promoting his particular experience as a universal way of running startups is a bit of an extreme way of doing that?

Really, should you lean on him?

Tech Note: how to fix MT9P031 low-light banding streaks…

A quick note to Aptina and to any developers experiencing odd-looking “streaks” in very low light with the Aptina MT9P031.

It turns out that these are caused (I’m pretty sure) by the analog offset sampling underflowing: if any of the four offset registers (R0x060, R0x061, R0x063, and R0x064) end up getting set to the minimum value (0x101, i.e. -255), then you get these odd-looking streaks. If you then disable the analog calibration (by setting bit 1 of R0x062), the streaks disappear: but unfortunately you also get a huge colour flash whenevr you do (bizarrely, this can look like a pink tartan pattern overlaying the sensor image), so this isn’t apparently a register you can practically change in real time.

After a couple of days of determined register poking, the least-worst fix to ameliorate (if not exactly ‘fix’) these low light streaks therefore seems to me to be to permanently disable both fast sample mode (by setting bit 15 of R0x062) and binary search mode (by setting bit 11 of R0x062). OK, it doesn’t make the problem go away completely, but it does seem to help a lot. Something Linux driver writers might want to know about! 😉

Incidentally, I presume this holds true for the MT9P001 as well because all its black level conditioning registers seem to be identical. So, Nick’s top tip for understanding the MT9P031 is to read the MT9P001 datasheet as well, basically for the bits Aptina left out. 🙂

PS: my next stop is trying out the same thing for the MT9M131 (which appears to have an earlier, slightly less sophisticated version of the same IP block), either by disabling the “rapid sweep” mode [by setting bit 15 of R0x060] or forcing the rapid step size to 1 [by setting bit 4 of R0x060]. Fingers crossed that will help…

Startups and ‘anti-knowledge’ minefields…

For want of a better phrase, you can think of ‘anti-knowledge’ as something passed off as the truth despite not actually being true. Of course, an outright lie would clearly qualify (duh), but probably the most dangerous type of anti-knowledge is a persuasive story built around a lie, particularly ones which serve to define an entire way of thinking.

I discussed a few such startup stories in the “Mythologies” section of my recent UCL entrepreneurship guest lecture, e.g. the cult of the “winner entrepreneur”, or the cult of the “powerful idea”. In fact, the more I looked at the whole entrepreneurship rationale, the more I concluded that it was nothing less than a sprawling minefield of (plausible-yet-false) anti-knowledge. The curious Zen gatekeeper challenge facing entrepreneurs therefore seems to be to find ways of unlearning these stories so that you can tackle the (arguably even greater) real challenge – of making money in a partially freefalling global economy.

Yet even this doesn’t quite sum up the extent of the problem. My wife pointed out this evening how a number of startup ventures I’ve been involved with over the years have been planned around (and built upon) the truth of various plausible-sounding stories about their business context, only to find (once I’d committed the venture to action) that the people telling me that story had deliberately not disclosed the whole story… a certain product wouldn’t ever get CE marking; a certain route to market would only open up once you had three products, not just one; and so forth.

But how can you ever test the truth of persuasive-sounding stories, short of actually relying on them? Startups have few enough positive assets that it’s extraordinarily tempting to treat these accumulated stories as a kind of knowledge capital, when they can just as well – if they are built around a lie or deliberate omission – turn out to be a kind of knowledge liability.

Perhaps there’s a core weakness of human nature at play here, a desire to trust despite the untrustworthiness of the folk business knowledge that surrounds us: and perhaps the entrepreneur’s biggest Achilles heel is when this desire becomes indistinguishable from a need. For we somehow need those manipulated anti-stories to be true in order to believe that we stand a chance of beating the odds… yet we will likely be brutally disappointed when the mine embedded in it blows up under our feet.

This is broadly the same snake-oil business schools sell their students too. That is, that the stories learnt in the entrepreneurship module lectures will help them beat the startup odds – really, that the deluge of positivistic “history as told by lottery winners” case studies and abstract models washing over them will yield some kind of ‘trickle-down’ microeconomic benefit. And business school lecturers would, for their own self-esteem, dearly like the stories they expansively relate to students to have universal resonance and truth, for they would hate to think of themselves as accidental purveyors of anti-knowledge.

But startups are just not like that: last year’s edgy market knowledge quickly becomes this year’s concrete overcoat – as competitors appropriate, adapt and extend ideas, so those same ideas become retrogressive. If you’re not on the shoulders of giants, you’re probably under their feet. What price knowledge then?