
Chris Anderson on Free 05/12/2008
What this covers
Chris Anderson discusses with Russ Roberts why digital goods increasingly trend toward free rather than merely cheap. Anderson argues that collapsing marginal costs in storage, processing, and bandwidth—falling faster than Moore's Law—push any digital product toward zero price. Roberts challenges this throughout, pressing on whether we're really seeing free or simply very low prices propped up by cross-subsidies. Their exchange spans the economics, psychology, and business mechanics of free, drawing on examples from Google's speech-recognition training through GOOG-411, Wikipedia's volunteer labor, and Craigslist's demolition of newspaper classifieds.
The conversation turns on several interlocking claims. Anderson emphasizes the penny gap—the psychological chasm between one cent and zero that makes micropayments fail despite seeming economically trivial—and argues that only true zero-price, monetized elsewhere (via advertising, freemium tiers, or user data), can overcome this barrier. He distinguishes "old free" (the razor-and-blades model, where you pay for the blade) from "new free," where digital economics actually permits giving away 99% of a product. Roberts counters that much of what passes for free still relies on unproven mechanisms: advertising revenue assumes exposure drives purchases; gambling-subsidized free attracts the wrong users. He also highlights cases where "free" simply means drastically reduced price—used books at fifty cents, not literally zero. The discussion covers the gift economy and non-monetary incentives, the collapse of traditional media's advertising advantage against measurable pay-per-click models, and the structural difference between digital distribution (near-zero to users) and physical atoms. Anderson's forecast—that every industry becoming digital will become free—provokes skepticism from Roberts about whether the economics genuinely support such a wholesale shift.
Anderson argues that the collapsing marginal cost of digital inputs (storage, processing, bandwidth) pushes prices not merely toward 'cheap' but all the way to 'free,' creating a new dominant business model built on cross-subsidies, third-party payers, and non-monetary economies that conventional economics struggles to predict.
- Digital inputs fall in price faster than Moore's Law, so any digital good trends toward zero marginal cost
- The psychological 'penny gap' makes free qualitatively different from cheap, killing micropayments
- Revenue migrates to third-party (advertising), freemium, and non-monetary incentives like reputation and attention
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Google's free directory assistance (GOOG-411) was not monetized through advertising; users were unknowingly training Google's speech-recognition algorithms—correcting it on regional accents and proper nouns—thereby creating value for Google in exchange for the free service (a labor-exchange model).
“what you're doing whether you know it or not is training their speech recognition algorithms”
Economics is good at understanding individual non-monetary motivation (Adam Smith's Theory of Moral Sentiments addresses pride, ego, reputation) and even altruism, but it is poor at making ex ante predictions about which markets will function better as gift markets versus for-profit markets—economists could not have predicted Wikipedia would be a high-quality product.
“we're not so good at is making predictions about say which markets will work better as gift markets rather than for profit markets”
The 'penny gap': economically the difference between one cent and zero is insignificant, but psychologically the chasm is often uncrossable, because any non-zero price invokes the brain's 'is it worth it?' value calculation—mental work that can exceed the monetary stakes—so micropayments carry the psychological baggage of a price without the economic virtue of a price; the solution is to round down to zero and find another revenue source.
“the difference between a one-sentence $0.00 probably isn't significant and yet psychologically the chasm is often uncrossable”
Whether monetary transactions or cultural norms govern an exchange depends on the situation: people readily let an urgent person cut a grocery line for free (a sympathetic appeal) but recoil at being offered cash to do the same, showing that the role of money versus social norms—not class—drives acceptance of market solutions in informal settings.
“if you say I'm in a hurry here's 20 bucks per alia split it up and you cut in line nobody thinks it's attractive”
There is 'no such thing as a free lunch' in the economic sense not because of externalities but because of opportunity cost: even when something has no out-of-pocket cost, no expectation of reciprocity, and no favor owed, it still consumes time and the resources used could have been put to another use—though it can be genuinely free from the consumer's narrow perspective.
“the idea in economics when we say that nothing's free is that everything has a the fancy ugly jargon is an opportunity cost that something else could have been done with those resources even when there's no out-of-pocket”
Spam is what you get when the marginal cost of communication becomes so low it is effectively free—a tragedy of the commons—illustrating that zero marginal cost is not always desirable.
“we call that spam yeah and so that's exactly and that's what you get... a tragedy of a common”
Bloggers compete with professional newspapers by giving away what was previously charged for, and although no individual blogger is as good as a major paper, in aggregate they do a better job because they can focus with infinite granularity on narrow subjects of specific relevance.
“in aggregate we do a better job because we because we have we can focus with infinite granularity on subjects and narrow relevance”
The advertising/media model is expanding to non-media domains (search, email, software services) even as it fails the traditional media companies themselves, because newspapers must now compete with more measurable, effective advertising (Google AdWords/AdSense) and the shift from pay-per-impression to pay-per-click exposes that traditional media cannot prove its advertising works.
“the media model is sort of is sort of you know failing the media but succeeding everywhere else”
The 'gift economy' (Wikipedia, Craigslist, the blogosphere) produces things of genuine value with no one making money, driven by non-monetary incentives like reputation, attention, and expression, which prove not only sufficient to generate things we used to pay for but often better at generating participation and labor donation.
“the recognition of all these non-monetary incentives like like reputation attention expression etc the proof not only to be sufficient to generate you know things we used to have to pay for but often better at generating participation”
What drives price all the way to zero (beyond economies of scale, which only push toward low) is the ability to monetize the user base for other purposes—getting them to use your other products, cross-selling, brand lock-in, or advertising—as in the Yahoo/Google case.
“what's driving zero is something a little different I suspect the Yahoo Google story it's that you're going to get access to these folks for other purposes”
Falling and even very low prices are not psychologically the same as free; the leap from 'cheap' to 'free' is extraordinary and is what makes the phenomenon distinct, since technology alone would only suggest cheaper prices, not zero.
“low it's not the same as free psychologically so the leap to free which is really quite extraordinary is a different thing”
The internet is the first time we have had near-zero marginal cost of both manufacturing and distribution combined; earlier analogues like pharmaceuticals had low production cost but still faced distribution scarcity.
“the Internet is the first time we've had sort of the sort of you know low margin of near zero marginal cost of manufacturing plus near zero marginal cost of distribution”
Whether people devalue something because it's free depends on its history: things that used to be costly and become free (e.g. the Village Voice going free) tend to be valued less, whereas things never paid for (e.g. Google search) are not devalued and users even get angry when they're not free—so the marketing challenge for a free book is to reset expectations so readers value it as if they paid.
“do people value something less because it's free and typically the answer is if it used to be if you used to be costly and now it's free they tend to value it less”
Craigslist demonetized roughly half the newspaper industry by making classifieds free; this is the marketplace in action, benefiting millions of users while the losers are a few institutions that didn't move with the times, and there is no law that any industry must remain its current size—a $4 billion industry can be turned into a $4 million one while delivering far better information service to consumers.
“when Craigslist demonetized half the newspaper industry by making classifieds free that's yeah that's the marketplace in action”
There is 'old free' (cross-subsidies like the 19th century saloon free lunch, razor-and-blades, free checking) where one thing is free but you pay for another and don't spend less overall, versus 'new free' enabled by the internet where things are genuinely free to the consumer because digital economics drives marginal cost to near zero.
“there's basically old free and new free... it is the razor and blades it is the free checking etcetera it's all across subsidy of sorts where where one thing is free but you pay for another”
In the physical world of atoms you can give away maybe 1% of your product (free samples) and sell 99%, but in the digital world the logic inverts: you give 99% away for free because it costs nothing, so that the 1% premium buyers can subsidize everyone else.
“in the digital world you want to do just the opposite you want to give 99% of your product away for free because it cost nothing to do so”
Digital economics (high initial development cost but near-zero marginal cost of production AND near-zero marginal cost of distribution) almost requires maximizing reach, because scale costs nothing and enables business models from advertising to freemium; this combination is essentially unprecedented and is what kick-starts free as the dominant business model.
“web businesses are all about scale because scale because you know because scale costs nothing”
Everything online is driven by ego—people contribute to promote themselves, blurring the line between content and advertising, such that a Microsoft engineer's blog about their project is simultaneously self-expression and commercial promotion, making the terms 'media' and 'advertisement' increasingly meaningless.
“everything online is driven by the most powerful force in the universe which is ego”
It is easier for a newcomer to introduce free than for an incumbent to respond, because migrating an existing large user base to free is costly (you must buy massive new capacity to satisfy demand with no new revenue) whereas a newcomer with few customers can promise the same thing at much lower economic consequence.
“the newcomers can can introduce free and disrupt the market much more easily than the incumbents can respond”
The media model is a three-party business—producer, consumer, and a third party (advertiser) who pays for access to the consumer—which has existed for a century (radio, free-to-air TV, controlled-circulation print) and is now becoming a core business model for nearly anything on the internet, including software and services that were never previously advertising-supported.
“the medium model is a three is a three party business where you have producer you have a consumer and then you have the third party who pays for access to the consumer”
Anderson's proposition is that every industry that can become digital will become digital, and every industry that becomes digital will become free, so sooner or later virtually every industry will have to compete with free or become free.
“every industry that that can become digital will become digital and every industry that is digital will become free”
Roberts plays skeptic: the internet has mostly made expensive things cheap rather than free—e.g. a used hardcover history book on half.com for 50-75 cents, or a personal webpage for $100 instead of $20,000—and sellers price below cost mainly to build scale and reputation, cross-subsidizing via advertising or network effects, but these are still not literally zero.
“but the books aren't zero there's 75 cents it's just really cheap the software's not 400 or 800 or $1000”
Free has a 'magic quality' that both grabs attention and turns on greed instincts while simultaneously triggering suspicion (because we've been suckered by fake-free offers before); economists tend to underweight this psychological impact, as illustrated by crowds' irrational zeal for free t-shirts fired into baseball stands.
“free has a sort of magic quality that does it does is both it both sort of grabs your attention turns on all of your greed instincts and also we're so rightly suspicious of it”
Free-but-monetized-by-gambling or advertising models carry adverse-selection and effectiveness risks: a free flight subsidized by gambling may attract non-gamblers who never generate the needed revenue, and advertising-funded free relies on the unprovable assumption that exposed consumers actually buy the products—'the mystery of advertising for the last hundred years.'
“if you make it free but you can gamble while you're there it tends it might tend to attract people who don't like to gamble but just want the free flight”
The reason everything on the internet trends toward free is that for the first time in history all the key inputs—storage, processing, and bandwidth—are halving in price roughly every 18 months, so whatever you do today will be cheaper tomorrow and eventually so close to zero you might as well call it zero.
“all the inputs in this case storage processing and bandwidth are basically falling in price are having in price every 18 months”
In the 2002-2004 email market, Google disrupted incumbents by giving away a gigabyte (then two) of free storage based on the insight that storage cost was falling even faster than Moore's Law (halving every ~14 months), forcing Yahoo to respond first by matching then by offering unlimited free capacity, because the curves meant whoever priced at zero first would eventually find it profitable.
“we're going to give away a gigabyte and then they went to two gigabytes and for free and so now yahoo is like okay now how do we beat that”
The newspaper industry's decline reflects two dimensions: commodity content (e.g. sports scores available from ESPN, Yahoo Sports, 50 other free sources) gets driven to marginal-cost-zero by competition, while highly relevant narrow-interest content (which readers are price-insensitive about) cannot be served economically by professional media organizations that cannot scale down to that economic level.
“there is commodity versus non commodity and relevance if it's a commodity ie I can get at other places then... in a competitive market price balls of marginal cost bingo zero”
Being first to market is sometimes overrated because latecomers can learn from early movers' mistakes, but where being first matters, giving the good away (or even paying people to use it) to build a base of users is a valuable strategy.
“a lot of people emphasize the virtues of being first into a market I think that sometimes overrated because late comers can often learn from early comers mistakes”
Google's invite-only rollout of Gmail was a clever strategy that gave the product cachet, made people want to find a way in, and incidentally limited the rate of expansion (and thus the cost ramp) during the early period.
“remember in the early days you couldn't just sign up you had to be invited and that was a very clever thing it gave it a little cachet”
Many service industries previously requiring expensive professional experts (divorce lawyers, tax accountants, travel agents, brokerages) have turned into software and consequently become free or nearly free, leaving the consumer to pay only residual fees like court filing costs.
“all these service is used to be expensive propositions offered by professional experts now their software and now they're free”
Most bloggers depend on the existence of professional old media (New York Times, Washington Post) to critique, link to, and analyze, so the new media needs the old media—though the speaker concedes he cannot quantify 'most.'
“the new media needs the old media most of us most of us who blog are relying on the existence of online newspapers to critique and make fun of and analyze”
Ryanair has redefined the business it is in—not 'a seat on a plane' but broader travel and ancillary services—making money from rental cars, hotel reservations, cargo, baggage fees, food and drink, duty-free, on-board and website advertising, and subsidies from off-the-beaten-path destinations, with the CEO promising eventually free flights subsidized by on-board gambling; the underlying jet-fuel and aircraft costs are flat or rising, but technology has lowered the cost of everything around the flying experience.
“they redefined the business they're in they're not in the in the share on a plane business there in a kind of a mush hole broader come travel and stuff around its business”
The capital cost of large server farms remains very high, but because each generation of hardware serves far more people, the marginal cost on a per-user basis falls lower and lower—so 'free' refers to marginal not capital cost.
“we're obviously talking about the marginal cost obviously the capital cost of of you know big server farms is still very high”
Anderson plans to give away the digital forms of his book (audio, ebook, web book) for free since their marginal cost is zero, and to produce physical versions monetized differently—an advertising-supported edition with sponsors, a normal paid edition, and potentially distributed free at conferences or via transit handouts—because as an author he is in the 'me business' of propagating ideas widely rather than the book business.
“I'm in this sort of me business which is to say propagate my ideas far and wide let me figure out how to monetize them”
There is one thing better than free—being paid to take a product—and people are already paid to do things they value, such as receiving a free computer or TV for joining survey activities.
“there is one thing better than free and that's being paid to open an email account”