broadstuff

I was on holiday at the time so I'm a bit late onto it, but this piece in The Economist chimes a lot with my experience:

Firstly, "market research" for new technology startups is nothing like what usually works for more tradtional start up businesses:

Anyone who has tried a hand at starting a high-tech business—seeking to turn a clever research idea into something customers will pay good money for—quickly learns that everything taught in business school is next to useless. The mistake is to think of start-ups as just smaller versions of established businesses. They are nothing of the sort.

Existing enterprises, even so-called SMEs (small-to-medium establishments), usually have no shortage of marketing and financial data. Their success depends on how well they execute their data-driven plans. Start-ups, by contrast, are faith-based organisations that survive on passion and belief—with nothing to go on but a bunch of hunches.

Our experience is its far more about modelling scenarios, what-ifs, probabilities etc and laying down "marker stones" to know where you are, so you can tell which scenarios are probably playing out. For this reason I was very interested that Silicon Valley Entrepreneur Steve Blank has come to similar conclusions, but with a lot more data and experience, and has set up his own accelerator - i-Corps - to do this (I know, I know re Accelerators* - but read on, there is some interesting stuff here):

What distinguishes an I-Corps start-up from a typical university spin-out is the way it forces researchers to stop fixating on the technology they have developed. New ventures, they are taught, are all about finding customers, what distribution channels to adopt, how to price the product, who to partner with, and more. From day one, the mantra is “get out of the lab”. Participating academics have to make countless cold calls to potential customers—something few research scientists and engineers have ever done in their professional lives and most initially find awkward.

The I-Corps programme is based on the premise that all new ventures are little more than a series of untested hypotheses—in other words, optimistic guesses about market size, customer needs, product pricing and sales channels. With so many unknowns, the programme teaches participants to treat their start-up as if it were a typical research project, amenable to the same iterative process of hypothesis testing and experimentation.

Stopping Technology Fixation is IMO one of the absolutely critical issues for Technology startups, the problem is its virtually impossible for techies who are in love with their technology to do - so the second bit, on treating the startup dispassionately as a live project, seems like a very clever way of forcing that dissociation. Anyone who has studied innovation and startups will know that which technology/company/method succeeds is highly unpredictable in the initial stages, and all sorts of factors come into play - exactly because all sorts of factors are untested.

But I was also impressed with the following, because in my experience turning around tech companies, these are usually the most critical issues - making sales and making profits:

Mr Blank’s blueprint draws on two complementary techniques—“business-model design” and “customer-development process”. The first uses a technique pioneered by the authors Alexander Osterwalder and Yves Pigneur in their book “Business Model Generation”. This provides a handy graphical template (called a "canvas") that helps fledgling start-up teams envisage their most likely customers. With the canvas' nine separate blocks covering such things as resources, value propositions, customer relations, cost structure and revenue streams, it captures the complete essence of a business plan. By filling in the individual blocks, the hopeful entrepreneurs are forced into the sort of business brainstorming many are unfamiliar with.

Customer-development process, meanwhile, provides a real-world way of testing the business plan's various hypotheses, as written in the canvas' nine blocks. The testing is done by getting potential customers, suppliers, partners and channel operators to challenge the assumptions. If invalidated, the I-Corps team has to pivot, and revise the entry in that particular block of the canvas. The process is repeated until all reiterations, in all nine blocks, are finally exhausted and the business plan emerges robust enough to be realistic.

There are other methodologies than those suggested here of course, but in my experience the "80/20" impact is the act of focussing on the business model (how you will make money) and the customer model (who will buy). Its damned hard to do, especially for the techies who are struggling with making the new wotsits work, but absolutely critical. You don't have to have the "right" business model initially (just look at Google) but you have to know what a "wrong" one looks like, and what the envelope of success looks like.

Also, they emphasise that there will be significant twists and turns in the business (or "pivots", as the fashion is to call it these days), its a given not a failure

The I-Corps curriculum emphasises that failures which force participants to pivot and change their assumptions are an integral part of the learning process. That can mean rethinking the market, changing the price, even altering the product itself. During the eight-week programme, I-Corps teams repaint their business-model canvases literally dozens of times. The average team confronts 100 or more potential customers while honing its business plan and tweaking its product.

Re: pivoting, I increasingly think of a startup as a bit like a new genetic algorithm, so it has to go through a number of self winnowing cycles before it can navigate the ecosystem its launched into. Another analogy may be a new species that has emerged blinking onto the landscape and has to adapt to its niche.

Its too early to say whether this particular program will work, after all execution is 9/10 ths of the lore, but I think the proposed areas of focus are in the right direction. The one thing this piece does not say, of course, is that for any startup taking money from an accellerator, incubator, seed investors etc - Cave Contract! Get a lawyer to look it over.

*Yes, in the Bubbletime we are also sceptical of any New New Accelerator story, but this piece has some interesting points.

WSJ notes that many companies are still using the Freemium model, but its not necessarily successful:

Yet for some, the "freemium" strategy is turning out to be a costly trap, leaving them with higher operating costs and thousands of freeloaders. That's what happened to Chargify LLC, a provider of billing-management software to small businesses, which used the freemium business model when it started out in 2009. The Needham, Mass., company gave away its software to merchants that billed fewer than 50 customers a month. If a merchant wanted to bill more than 50 customers monthly, then the business owner would have to start paying $49 a month.

Most Chargify users never became paying customers. Within a year, the company was on the path to bankruptcy. Chargify eventually put up a paywall for all users. Last month the 12-employee company became profitable, with more than 900 paying customers. The starter plan is $65 a month.

If only they'd read us in 2008, when we explained why FreeConomics overall would fail. (I can tell you that presenting ^that^ paper in 2008 at the O'Reilly Berlin conference was "interesting" as Free! was at the time the latest prayer to a cash-strapped bootstrapping startups' needs. In 2012, maybe not so much controversy as to what does and doesn't work - as the article notes:.

Start-ups are attracted to the freemium model because it is "deceptively simple"—lure users with free services until they're hooked, then charge for extras, said Vineet Kumar, a professor at Harvard Business School, who is currently working on a study of the model.

"The problem is, it's not always obvious what features should be free and which should be paid," he said. What's more, he added, offering too many features in the free version risks "cannibalizing your paid customers," while not offering enough might generate little interest all around

The freemium approach doesn't make sense for any business that can't eventually reach millions of users. Typically only 1% or 2% of users will upgrade to a paid product, said David Cohen, founder and CEO of TechStars, a start-up accelerator since 2007 with five U.S. locations.

About a quarter of the roughly 50 companies that participate in TechStars every year use a freemium strategy, he added.

The strategy also often isn't effective for businesses whose range of products is limited in scope, because paid users generally expect to get better or different versions of what they've already received free of charge. And it rarely makes sense for companies that sell products or services mostly to large businesses. Enterprise clients typically have budgets for buying goods and services, thus, they aren't as drawn to free products.

Its that low takeup rate (in 2008 people were casting about numbers like 10% takeup, we thought 5% would be damned lucky) and lack of business customer interest that caps the Freemium strategy in any given niche, and if its a popular niche there is always the next entrant with a bit more VC funding runway left who will still be free when you are trying to charge. And as we predicted, the most likely way Freemium would actually work was via offset funding (Ads, Virtual Goods, Buy-through goods) - ie other revenue streams than the product itself. Successful Apps (which have the best Freemium profile - low setup cost, low marginal distribution cost, minimal support cost) just about all use offset funding of one stripe or another.

Anyway, here is a summary of the knowledge today:

Steps to Making Freemium Work

1 Limit free offerings to avoid cannibalizing your paid customers. Expect your free-to-paid conversion rate to be less than 5%.
2 Promote both free and paid offerings, rather than focus just on acquiring paid users.
3 Create an incentive for free users to refer their friends to your service or product.
4 Clearly distinguish between free and paid plans, so users can see value in paying for more.
5 Create a natural progression from free-to-paid features, such as the option to buy extra storage space on a cloud service once a user runs out of free space.

In defence of Ad based Freeconomic advocates however, we were wrong (so far, anyway) when we foresaw that the biggest risk was in using Ad funding to fund the Freemium business - we pointed out in 2008, if you ain't paying for the free lunch, then you are the free lunch (or as it was better put later, users are not the customer). Needing to feed the Ad Monster forces the company into ever more risky data mining activities, and playing ever more loose with customer privacy. We thought that more companies would have had this aspect blow up in their faces by now, but we have been proved naive about how little reward people require in exchange for putting large amounts of their personal data online.

I saw this advertised today on the Twitter, and went along to the page linked (here), abuzz with hope - so what did I learn?

In Summary:

1. Don't compete with Amazon.
2. Don't sell stuff Amazon sells
3. Don't sell stuff Amazon resellers sell
4. Don't sell stuff in countries where Amazon sells it
5. Don't sell stuff where people know that they want already, because they go to Amazon first
6. Sell branded stuff, but only that Amazon doesn't sell
7. Sell to iPad owners because they are richer....
8. .....and Amazon is too busy pumping Kindle
9. Sell stuff with high margins (and low delivery costs)
10. Mobile is The Future (unitil Amazon does it, of course....)

Tongue in cheek I know, and to be fair, there are some snippets of useful stuff there - but I don't think the above summary is far wide of the mark!

Today the CEO of Yammer argues thus (after having sold to Microsoft for $1.2bn, oddly enough). In essence he argues that its hard to:

(1) Find an idea that the Big Guys haven't thought of
(2) Prove it in the c $5m runway the SV VCs will give you these days
(3) If it takes off, keep it going long enough to grow before some major player swats it with a competitive play, patent suit or other or spoiling manoeuvre.

As you'd expect, some prominent SV VCs take issues with this, Marc Andreesen arguing that there is an infinite amount of creative talent out there and the big companies are ponderous and often not well run (this being the same Andreessen you recall who watched his own Netscape get crushed by a fairly inept Microsoft's Explorerer in the mid-nineties, of course....but then he made good on Skype. Maybe one forgets....)

We have wondered aloud about the End Of Silicon Valley in in the past, but have always noted its ability to resurrect itself. But today, as well as the above, three other different articles, on 3 different aspects of this issue, seem to be painting an emerging picture:

1. Firstly, the Old School VC's aren't making any money anymore - Business Insider

Last year, in an attempt to cash in on the late-stage "private IPO" investing frenzy, Kleiner startled the Valley by launching a $1 billion late-stage "Digital Growth Fund." Many saw this fund as Kleiner's attempt to play catch-up and associate the firm's name with hot companies it had missed. Some also saw the fund as a sign that Kleiner had lost not just its edge but its valuation discipline and predicted that the firm might get clobbered on some of its late-stage investments.

These concerns were well-founded.

Armed with a boatload of new cash, Kleiner rushed out and made late-stage investments in, among other companies, Facebook, Groupon, Zynga, and Twitter. In most of these investments, moreover, Kleiner did not buy stock from the company--thus providing additional growth capital for the company--but, instead, bought stock from existing investors, helping them cash out with huge gains.

Kleiner's investment in Twitter may yet pay off, but it has been far from a home run.

Twitter is unlikely to make up the numbers. Many of the older VC firms took big hits in the DotCom bust (I recall seeing a McKinsey report that said no investment after 1997 made money). KP has been widely felt to be off the pace for a few years, and this was seen by many observers as a last ditch attempt to get back in the game.

2. Secondly, even the New Boys are backing out of their new investments- WSJ

Some of the early backers of Groupon Inc, including Silicon Valley veteran Marc Andreessen, are heading for the exits, joining investors who have lost faith in companies that had been expected to drive a new Internet boom.
....
Shares of Facebook continued to slip last week, falling to about half their May IPO price, as sales restrictions expired for some pre-IPO investors. Since Zynga's December IPO, its shares have dropped 70% as investors worry about its ability to keep making money from people who play its digital games.
....
The Groupon investment flurry that ended in January 2011 was part of a recent Silicon Valley trend of prominent investors jumping into young companies shortly before their IPOs. Critics say the phenomenon builds unsustainable valuations for those young companies.

"Groupon would never have gotten this big without that late-stage money," said Bill Gurley, a general partner at venture-capital firm Benchmark Capital.

At least they got in earlier, but its not a viable model long term when the IPO money knows its going into an almost pump 'n dump scheme, so no surprises then that.....

3. Thirdly, the huge Californian Public Service pension fund (CalPERS), a stalwart VC funder over the decades, is considering pulling out of funding VCs - PE Hub.

The nation’s largest public pension fund, CalPERS, appears likely to slash its investments in venture capital in what could be a blow to this still recovering asset class. Venture capital has wallowed through nearly a decade of difficult returns and sluggish fundraising. The decision by the California Public Employees’ Retirement System to largely abandon new commitments to venture funds will likely prolong the agony.

......

Over the past decade, venture capital has been the worst performing asset class in CalPERS’ private equity portfolio. It has delivered a net return of 0.0%, according to June 30, 2012 pension fund data. Part of the poor performance stems from the continuing fallout from the dot-com bust, which saddled many venture funds with negative returns [See 1 above} . Performance was hurt again by the recent financial crisis, which had an impact on almost all asset classes.

In the past year or so, venture returns have improved as IPO activity has risen. Still, venture capital has fared poorly relative to other asset classes at CalPERS. Buyout funds, which represent 57% of CalPERS’s private-equity portfolio, have produced average annual returns of 15.4% over the past 10 years, while credit related investments, another big CalPERS investment focus, generated 14.1% annual returns.

Granted they only put about 6% of their c $240bn fund into VC, but that is a very big number for the VC community

Of course, if CalPers had just shorted the stocks of the companies their VC investments were funding, like the banks did, they would have made out like bandits - WSJ again:

Profits made by banks underwriting Facebook Inc.'s initial public offering as the stock fell were distributed this past week from a pool of about $100 million, say people with knowledge of the deal.

The banks together made the money through a process known as stabilization, which is a standard procedure in IPOs, though Facebook's deal size made the profits larger than normal, these people said. The payout comes in a week that saw Facebook stock hit new lows, falling Thursday below $20 after an IPO price of $38 a share.

Thats on top of the c $175m they were paid to do the IPO. Easy when you know how.....but not good news for the whole VC ecosystem. Private profits, public (offering) debts....

Both the FT and the Grauniad have opined, dewy eyed, on the stupendous lessons for future successful management from the Olympics. The Grauniad's effort is entitles "Olympics: the key to our success can rebuild Britain's economy" and has its core thesis is that success is a mix of infrastructure investment and private elan, that of course the Grauniad's oppoenents (the Conservatives) don't understand this:

...we need to sustain these principles so they become structurally and culturally embedded for continuing Olympic success, but they should also be applied elsewhere. The problem is that they are born of an ideological hybrid that wrong-foots our political class. They are mostly rooted in liberal social democratic values that understand the importance of public investment, public organisation and institution-building. But they also involve an unashamed recognition that in the end individual application, resolve and will to win are indispensable.

David Cameron and London mayor Boris Johnson are happy to celebrate the element that is rooted in competition, elitism and individual effort. But they flounder the instant the conversation moves to the role of public investment and the necessity of understanding and sustaining our unique sport ecosystem, just as nearly every Labour politician flounders the other way round.

Now the Grauniad is fairly left of UK centre, but the FT is ver "small c" (and probably large) conservative to a T (or "C") but it too thinks amazing lessons can be learned (I'm not allowed to cut and paste what they say, as the T&C's are "Please share this article with others using the link below, do not cut & paste the article. See our Ts&Cs and Copyright Policy for more detail. Email ftsales.support@ft.com to buy additional rights").

Anyway, they too waxed lyrical on the Legacy, and the benefits of running with large numbers of (lowly paid) volunteers, and that Vancouver got improved transportation, more affordable housing, and top notch athletic facilities. But, we got Zil lanes, people thrown out of their flats by landlords, and the unedifying spectacle of 2 football clubs scrapping over what bargain basement price the "top notch athletic facilities" would be flogged off for afterwards.

Now don't get me wrong - the London Olympics were a superb example of a very complex project, delivered to a very high quality, brought in on time. And the Team GB athletes performed superbly, and the Volunteer Army were superb, and won everyone's hearts. But it was certainly not on budget (even with all those near-free Volunteers) - that went up from c £4bn to c £12bn - so you'd kind of expect the delivery quality would be Rolls Royce class! And, what we also got was gold medal quality finessing of commercial greed, and there are some very searching questions that need to be asked about the ticketing, the pricing, the giveaways to commercial sponsors, the pursuit of perfectly innocent people on trademark etc infringement, and changes made to British laws to make this happen. However, they probably won't ever happen because the Olympics comes around once every 2 generations or so, so all lessons will be forgotten - and Britain had a very good Games, so it will likely all be brushed under the carpet in the carefully cultivated ensuing euphoria.

So, for the truly ambitious manager of Machiavellian mind on the make, we leave this post to posterity to ensure that the real Olympic lessons of success are not lost - and without further ado, here they are:


THE 10 REAL MANAGEMENT LESSONS FROM THE OLYMPIC GAMES

1. Massively underestimate the costs of your project

2. Massively overstate the benefits of your project

3. Get VIP/Celebrity support in early - the Chatterati will then write reams of breathless copy, and the few sceptics are drowned out

4. Find a sponsor with Very Deep Pockets (a national Government, say) and then ensure that their reputation is on the line, so if anything goes wrong it is their problem. From now on you can't lose - and can overspend till the cows come home.

5. Use free or cheap labour wherever possible, and fire them up to believe they are involved in a great enterprise. Bonus points if you can make them pay to work.

6. Ensure that your Commercial Sponsors can privatise their profits while your large backer can tax their citizens to make all the huge losses public

7. Give those Commercial Sponsors the keys to the kingdom in terms of enforcement of commercial privileges. Close parish magazines, sue mom and pop bakeries, trademark 1000 year old words in the English language - all good stuff

8. When the huuuuuge overspend becomes public, argue that it would have been even bigger, but for the heroic efforts of The Brave Management

9. Scalp the paying punter - create artificial shortages that increase desperation and inflate access prices in all ways you can.

10. Fireworks. LOTS of Fireworks. And lots of celebrities, and feelgood about ourself stuff - so everyone goes Yay, Paaaarty and completely forgets about points 1 to 9

Here endeth the Lesson. See you in Rio

(Update: Matt Schofield points out Lesson 11: Win way more golds than expected. Success has a thousand fathers....)

Yes, we told you it would crash, in June 2011.

I twt'd this on Friday:

"the Groupon biz model is to use Other People's Money to grab eyeballs, then IPO the f*cker before it comes crashing down - Dotcom II"

And then worse came out the woodwork with their filing. Seems they are technically insolvent and the bigger they get, the more they lose, along with the fact that later investors were paying for earlier investors' exits, and the IPO will no doubt pay for the current investors' exits. Greater Fool Theory in full technicolour! (In fact some have been more direct, calling it a Ponzi scheme)

We can only apologise to the readers of our Bubblewatch series, in that we never predicted an attempt to IPO a Greater Fool Theory dotcom business before the big one had gone out. Move the notch up one (see chart above). DotCom II is nearly here now....But a whole bunch of dumb pension funds will invest - with your money. And This Time It Is Different, as everyone seemed to arguing at D9.

Of course it is...........

There is nothing wrong intrinsically with an online coupon business, there was something seriously wrong with one at that starting valuation.

But the reason we are not writing this blog from our own Caribbean island is that its a lot harder to predict exactly When.

When Marissa Mayer announced she was joining Yahoo on July 16th, Twitter was largely positive in its praise of her, but Yahoo's stock went down slightly - see graph above (on an announcement of flat results, on a day when the NASDAQ also went down slightly, so the Marissa Effect was at best neutral, maybe slightly negative)

Today, when Mark Thompson announced he was leaving the BBC to joing the New York Times Twitter was more sceptical ("UK Publicly Funded TV/Radio Broadcaster to be responsible for US Commercial Newspaper revenue" was typical) but the stock market price went up a bit.

In recent months Twitter has been hailed as a predictor of a company's future success, even it's stock price:

Want to predict what will happen in the stock market tomorrow? Just use Twitter, say Californian researchers. A team at the University of California has created a computer model which allows them to 'predict the future' of the stock market by scanning the social network. It has up to 11 per cent more accuracy than other computer models. Researchers found that scanning for volume of Tweets and how those Tweets linked to other subjects could 'predict' stock prices a day later - and outperformed other computer investment tactics. The software looks at volumes of Tweets containing references to companies or products, and how they link to Tweets on other subjects. It can predict both the volume of trading and the value of a stock the next day.

In the case of YHOO (Twitter +ve, Stock market slight fall) and NYT (Twitter sceptical, stock market slight rise) Twitter and the market are at odds. Be interesting to see where things are in a year or so's time. As the old adage goes, you can't buck the market......

BBC on Webinos (Webinos is open source Web Operating System technology designed to run on PCs and Android-operated mobile handsets - see Webinos.org)

More than 5,400 developers have downloaded a new open-source operating system designed to enable digital devices to communicate with each other. They are now looking at ways in which Webinos could be used to connect a range of devices such as mobile phones, car stereos, heart monitors and TVs. Webinos is a 15m euro ($18.4m; £11.8m) project supported by more than 30 organisations, including the EU.

So far people in 155 countries have accessed the Webinos website. "People want to control the technology because if you control the technology you control the money," he said. "But it should be free and open to everybody. "Where we are operating is trying to build the open-source community. That's how web browsers started."

Just as the Internet needed open standards to develop to its fullest potential, the Internet of Things (IOT) does too. Right now all the major players are attempting to build walled garden solutions that won't interoperate. Its the same old same old. The other same old lesson is that it always limits innovatio (and quite often an openly available system breaks the walls0

I don't know if this will be the Mosaic of the IOT, but let us hope something will.....

A few days ago I wrote about "Greshams Law of Information" where, given equal transaction values (ie Google links) then bad information drives out good. This is demonstrable on Google these days as it gets (i) gamed and (ii) all sorts of paid for or belief based sites garner far more links than the boring truth with their superior resources (and now, maybe (iii) allowing interested parties to pay for link manipulation).

I defined 2 types of "crap info" on the web:

Case 1 - Info that is largely aimed at flogging something (ie marketing/advertising), where nearly everyone knows its probably crap and ^will try and seek the truth
Case 2 - Info that is largely aimed at promoting a belief, where many people will believe so will "self filter" out any inconvenient facts - and in fact try and repress them

Now, in economic theory, "Thiers law" enters the cycle later on into any new ecosystem, correcting this imbalance for crap currency as one currency becomes the "go to" trusted currency:

For money, the argument is that, in the absence of legal tender laws, Gresham's Law works in reverse. If given the choice of what money to accept, people will transact with money they believe to be of highest long-term value. If required to accept all money, good and bad, they will tend to keep the money of greater perceived value in their possession, and pass on the bad money to someone else. In short, in the absence of legal tender laws, the seller will not accept anything but money of certain value (good money)

In that previous discussion I could see how Thiers law for Information will work for avoiding Ad-crap (my Case 1 above), mainly via social nmediation (recommendation sites for example) but I didn't work out how the Truth would Out against Belief Based Dis-information (My Case 2). Except for a small bunch of people who would pay a premium for the facts, I could not see a viable dynamic that woulld reverse Gresham's Law as the wall of self-interested money and time promoting crap is far, far greater than that promoting truth.

I did, however, forget about looking more carefuly at Wikipedia as a possible solution, but meeting Jimmy Wales last week got me thinking about it. In essence, the only resource as freely available and as dedicated to Truth that can counter the vested interests of crap-spreaders is a very large voluntary effort of crowd-sourcing, that creates a "trusted currency" of truth - of which at the moment, Wikipedia is the best example (Though I did like the advice to Marrissa Meyer given in joyoftech - see comic strip at top).

Which makes it all the more imporatnt to ensure it stays that way, as the vested interests are way ahead of me and have have already sussed that it is the Trusted Currency of the web, and know their best bet is to subvert it

An interesting argument that scientific innovation is probably linear, and certainly not exponential - Cogntive Social Web:

The resources that we devote to science have been exploding throughout history. The world population of scientists has been doubling roughly every 50 years from 1500 to 1900 [4], and has kept an exponential rate throughout the 20th century also, by what seems to be a factor of 10 every 50 years [5] (it is however unclear whether that growth is still being sustained now [5], and even if still exponential the rate has evidently gotten slower). Likewise the number of papers and patents has followed exponential growth during the past century, unsurprisingly since average paper count per scientist is unlikely to change significantly.

Besides available researcher brains, the computing resources available to science have followed exponential growth as well (Moore’s law) in the past 50 years.

Furthermore, each new scientific discovery accelerates the pace of science to some extent, because it gives future researcher a better understanding of the field considered, as well as new tools to work with. That’s “accelerating returns”. You get to stand on the shoulders of giants.

The author also found (as we did when we looked at it) that it waxes - and wanes

And yet despite all this, if you look at the actual understanding of the world that we have achieved, if you look at real impact rather mere volume, the general impression you get is that scientific progress has really been linear. We didn’t “get” physics 50 times better in 1990 compared to 1940. The 1940-1990 progress of physics might even have been slightly less significant than the 1890-1940 one, actually. is

Even in a new, ultra-fast growing field such as computer science, most of the state of the art algorithms currently around date back from the 60s and 70s. Our computers may be 10^6 faster now than in 1960, but our knowledge is only a few times broader, our algorithm libraries only a few times better. Likewise neuroscience and AI have followed slow, noticeably linear growth since the 50s.

The question is why is this happening. .... view is its all about the peanut butter, the increasing friction in any system as it grows:

A widely-noticed empirical fact about science is that within a given field, making an impact is getting exponentially more difficult over time. Just look: the big discoveries in quantum physics had all been reaped by 1930. Most of what could be said in Newtonian physics had been said by Newton himself. Information Theory researchers will never surpass in impact the single 1948 paper that founded their field, no matter how long and fruitful their career. The list could go on.

In essence, the argument is that the first people into a new field grab all the low hanging fruit, ie the good old Pareto Principle is at work. So, as exponentially more people come in it gets exponentially harder to find, leading to a near-linear advance going forward.

So what about the rate of increase of New Fields? The author talks about these as "Paradigm Shifts" that open up the new field:

....in real life the paradigms on which scientific fields get founded tend to evolve over time, which means that the set of available discoveries that researchers can “chose from” will tend to expand over time. Paradigm shifts increase the scope of research that can be done. For instance, inventing transistors opened up the entirely new field of what to do with them. Or coming up with the basics of quantum physics. And so on.

I’d like to argue that these paradigm shifts can be modeled using the exact same model that I used for discovery. Yay for recursive models! We start with a finite pool of possible paradigm shifts, we devote exponentially increasing resources to finding them (“resources” meaning “researchers”, since each researcher tries a little to find out-of-the box ideas), and what happens is that 1) shift volume explodes over time, like papers published did earlier, because they’re a linear function of resources, and 2), the real impact at large of each shift decreases exponentially. Resulting in a linear growth of shift impact… which means in turn that the pool of available potential scientific discoveries increases merely linearly.

What interested me is that he built mathematical models of both the rate of new innovations in a field (as you'd expect, it yielded a power law of decreasing rate of returns), and of paradignm shifts. The Paradigm Shift was what really interested me (see graph at top) as it described exacly what my research showed, ie we advance by pushes forward, then pauses. I usually describe it as layers of S curves (S curves are the running totals of power law curves), so the linear model is interesting.

Anyway, for all the ambitious New Scientists and Innovators among you....

(i) Science advances linearly given exploding resources, and thus its pace will slow down in the future as the resources that we devote to it dwindle. Keep that in mind when choosing a field.

(ii) A piece of advice: as a researcher, you’ve got to work at the 2nd level of discovery, the paradigm level, because it’s so much more impact-efficient. Don’t devote your energy to discovering tiny new things within the old paradigm you inherited from your thesis advisor, rather discover broad news ideas entirely. Be imaginative.

(iii) Do not rely solely on your intelligence and hard work to make an impact on this world, or even luck, it’s not going to work. After all the total quantity of intelligence and hard work available around is millionfold what you can provide —you’re just a drop of water in the ocean. Rather use your imagination, the one thing that makes you a beautiful unique snowflake. Intelligence and hard work should be merely at the service of our imagination. Think outside of the box.

Caveat - as the author writes, "this is not a paper and I don’t have time for this. Take my model for what it’s worth: as something you’ve read on a random blog. If you’ve got a comment, or want to prove me wrong, be sure to post here or send me an email at francois.chollet@wysp.ws!"

But often a model like this is worth a million words.