Thursday, November 03, 2005
Research Note - Google Print, The Economics of Being Evil, and What Eric Schmidt Won't Tell You
Google Print is dividing opinions in almost every corner of the mediaverse. So much so, that Eric Schmidt recently took the time to explain in the WSJ, that the fuss is, really, much ado about nothing.
Is this really the case? Despite all the hoopla, almost no one has taken the basic trouble to understand Google�s latest initiatives from an economic and strategic point of view - especially not the tech community, which has been the most vocal in the belief that Google Print will be less evil than the alternatives. So let�s try and understand whether Google Print is something that content creators should rationally support.
The premise of this argument is, as Tim Wu has nicely recently delineated in Slate, whether authors should prefer exposure or control. This is, indeed, a restatement of Google CEO Eric Schmidt�s central point in the WSJ: that authors stand to realize, essentially, marketing economies they could not realize without Google.
Why, then, should content creators be nervous? If they�re essentially getting something for nothing, are their fears irrational? Shouldn�t they jump on this arbitrage for all it�s worth?
I think that Google Print is classic example of a strategy with unintended consequences. I think that content creators� fears are not just highly rational, but also, in fact, are deeply rooted in the history of their industry.
Let me explain why, by following the dynamics of Google Print to their logical conclusion, and adding a few points Eric Schmidt has (rather conveniently) left out of his exposition of Google Print�s costs and benefits.
��Indeed, some of Google Print's primary beneficiaries will be publishers and authors themselves. Backlist titles comprise the vast majority of books in print and a large portion of many publishers' profits, but just a fraction of their marketing budgets. Google Print will allow those titles to live forever, just one search away from being found and purchased.�
Of course, that�s true. But to really understand potential costs and benefits to content creators, what�s much more important is what Schmidt, in a not-so-huge coincidence, leaves unsaid: the fact that Google Print is a network good. It�s value to Google increases disproportionately with the number of books indexed. Because of this simple network effect, the benefits to Google of indexing such huge amounts of books are, well, bigger than huge.
So the question we really have to ask is this: do authors also stand to gain disproportionately from the size of the network? If not, it is Google who is the primary beneficiary of Google Print � not publishers and authors. If so, Google must tie their share of revenue to the size of the index. One way to do so is to share ad revenues with content creators.
In this case, not only is Google going to share no advertising dollars with content creators � and make no mistake, ad revenues from Google print are going to be significant � it is going to take essentially a marketing fee as well. That�s pretty cheeky.
In fact, Google doesn�t share revenues to fully reflect the network value of your good in any of it�s services; for example, in its AdSense service, you may receive 70-80% of revenues generated by ads on your blog, but the reality is that your blog increases the value of the Google ad network by a greater amount than the revenues you generate (much less 70-80% thereof).
At the same time, Google Print is also a killer example of an edge competence based strategy. Like core competences in the 80s and 90s, edge competences are going to dominate the post-network economy of the 21st century. By making info about books more liquid and plastic, Google atomizes upstream and downstream segments in the value chains. For example, it dilutes Amazon�s market power directly, by massively reducing switching costs � and, in general, the market power of anyone on either side of it�s value chain segment. Value shifts away from the core, and towards the edges.
The problems with Google�s revenue sharing arrangements, as well as the strategic implications of edge competences, bring us some way closer to the real strategic issue with Google Print: network effects create natural monopoly dynamics in this market. Just like there�s only economic room for one eBay, BoingBoing, or, well, Google, there�s only economic room for one global index of books. And this is the real heart of the problem.
Where else do we see similar dynamics? Well, in the currently evil media industry, which is an oligopoly dominated by a handful of players in each market. There, marketing economies of scale and scope mean bigger and bigger players invest greater and greater amounts in more and more homogeneous blockbusters. This is bad for consumers; it�s terrible for content creators.
This is the reason why so many content creators are fed up with the current state of the media industry. Whether it is record labels, film studio, or newspapers � the structure of the industry means that content creators have almost zero power to negotiate fair deals with publishers.
But remember, in it�s current incarnation, the media industry is, for all it�s manifold sins, at least an oligopoly. Content creators do have some power � but the structure of the industry means only if you�re you�re fortunate (or unfortunate) enough to be Tom Cruise.
From my discussions with content creators, many of them intuitively understand these historical dynamics, and apply them to Google Print. As they point out, Google Print essentially makes Google a monopsonist � the monopoly buyer and marketer, or publisher, of books over the www � and this is, in fact, (something most of them find unbelievable) one of the few things that threatens to make them even worse off than before.
When they ask whether Google Print really offers a fairer deal, they have to balance losses to a monopsonist to gains from marketing economies they couldn�t realize alone. Now, certainly, the latter gains are potentially nontrivial � revenues may double, triple, or, as Schmidt points out, more.
But the former losses to the monopsonist are potentially catastrophic � the monopsonist can essentially buy your good at as low a price as he or she wants, because he or she has all the market power behind the transaction.
So the real problem isn�t whether Google�s evil or not. It�s that it is almost too efficient at what it does; it�s ruthlessly brilliant solution to rationalize the books market also positions it as, paradoxically, an even bigger threat to content creators than their current worst enemies; the marketing droids who publish their books and demand the lion�s share of revenues in exchange.
At least content creators have some market power in today�s industry; in a Google Print world, the economic implications are that they will have none. From that perspective, to put simply, sometimes, you�re better off with the devil you know.
Nice blog entry. But I don't follow how having a big index (of books, web pages, whatever commodity content) creates a natural monopoly.
What are the barriers to entry for actors like MSN, Yahoo and Amazon.com to create their Google Print replicas? What are the barriers for a user to switch from Google Print to another book search service? Other than habit?
I can see the operational difficulties in creating a good book search engine, but not the strategic. Especially as publishers are likely to be interested in supporting competing services to limit Google's market power.
Would be great if you could clarify your reasoning for an interested reader.
This is just simple network effects. The utility I derive from an index of 1000 books is greater than 10x the utility I derive from an index of 100 books.
You can think about this intuitively as analogous to a phone network - the more people that join (books that are indexed) the more likell I am to find something that maximally satisfies me.
There are natural monopoly dynamics because this means the first-mover/biggest network wins - there are usually almost never for consumers incentives to switch (think of the phone network example).
Or you can check my edge competence presentation to understand it quantitatively - look at the Google network economies graph and think how Print will steepen the slope.
This is the first good write up of Google Print I have read.
I urge others to share this essay with their friends, because Google is dead wrong in terms of their rationalizations of Google Print.
While I do think getting more books online is a great idea, I simply don't buy the idea that Google's plans are the way to go.
I get that a larger index can give a user greater utility. But index size alone is not a good proxy of utility. The ranking algorithms used also influence utility - a lot. I.e. dump Google's web index of today into AltaVista of 1997 and Google would be far more useful due to better algorithms.
A comparison of Google Print to the phone network example don't convince me.
If I am on the phone network (together with *a lot* of other people) and switch to another, non-compatible, smaller phone network I will realize less utility. There are huge incentives to stay on the old network, as mass migration from the old to the new phone network is unlikely. The new network cannot clone/force the people using the old network, and are the basis for the network externalities, to switch. Therefore switching is unattractive. I cannot see these dynamics with Google Print or search engines.
If the basis for network effects is a commodity (customers are not a commodity, things often are), does the first-mover really gain a natural monopoly? If there are network effects, aren't they too weak to create a natural monopoly? Why should I not think about this in terms of economies of scale, where the largest player gets advantages but not a nautral monopoly?
Actually, I am with hencrkc on this one too : network FX on phone networks are realized because by joining the network I can talk to someone else, thus increasing the utility for both of us. I don't see the analogous effect for Google Print : if I can find and/or consume a book, how does it increase someone else's utility? My consumption on Google Print is passive, and the network only facilitates searching for the final good. Contrast this with phone network where the network is actively deployed for voice transmission during the act of consumption between me and my counterparty. From another point of view, think about the switching costs: if someone were to replicate Google's index with a reasonable searching algorithm, my cost of switching over to that service will be zero. This would not be the case with a phone network because I would need my friends and family to switch too because our utilities are connected.
Henrik and M.,
Thanks for the comments. I know you are both very smart and have studied some network stuff.
So here I think you are missing the forest for the trees, because my phone example is direct FX.
Search is very clearly dominated by network fx - they're just indirect; there's a third party involved.
You should think about network utility from an economic perspective - in this case, the network buyer is an advertiser - not other consumers.
Think about it that way. Clearly, my value to the advertiser increases when you join the network. So advertisers will always defect to the bigger network, because of increasing marginal utility.
This, in turn, creates pressure for scale: natural monopoly dynamics just like in broadcast or comms networks, because average cost decreases (or average revenue increases, if you like) in network size.
Scale in this case is Google "indexing all the world's information" - going across as many domains as possible.
Here is the barrier to entry. Amazon, MSN, and Yahoo are always playing catch-up, because indexing is getting more and more capital intensive as it moves to higher value domains (viz, now it's up to the point of satellite time for images and scanning books by hand).
Think about the limit case: the endgame is that all the consumers, and so all the advertisers defect to whoever has indexed all the world's information first, and then to whoever does it with a minimum of transaction costs.
Pretty clearly, this is likely to be Google.
Put another way: you won't switch away from Google if someone offers a better solution in *one* domain (web search, blog search, video search, whatever), when everything is delivered through a single command line, unless that domain is worth more than all the rest of them combined.
And it's pretty clear that this is Google's endgame, as more and more domains show up in what used to be web search.
Now, your logic works in the social search space - guys like Jeteye and Rollyo - who realize direct network effects, because users are helping each other search.
But even there, if you want to think strategically, you have to factor in the other side of the two-sided market: network buyers, or advertisers.
As usual it is one of your brilliant posts. I do get a few bits of what you are saying though I not as well read in network literature :)
What you said in the post was all correct but then as you pointed out in your comments itself it is the use of "phone network" as the analogy that caused the confusion.
All the way you in the post you were mentioning about indirect network externalities and the phone network is the classic example of direct FX.
Let me dare to say that I guess an analogy of a broadcast(say TV) network would have been more appropriate :) instead of phone example.
Phone is a two way network and thus exhibits direct FX and "broadcast network" & google here as pointed out by you is a one way network and they exhibit indirect externalities always.
Now how do I say this, well I also understood a post that you wrote ages ago where you referred to a paper by Economides on Network Economics :)
You are absolutely right - I should have used a broadcast, not a comms network, example, because the externalities are indirect.
I'm glad to see you are learning (and fast, if you ask me) - nice catch!!
Well, first of all, Google Print just changed its name to Google Book Search.
I've been a Google Print Partner for several months now. This was an extension of my experiment in electronic publishing. My thought was that some sampling on the Debbie Fields model would increase sales of my titles, which are sold through Amazon, Powells, Fictionwise, Diesel E-books, Elibron and hundreds of web sites in affilate programs. My distributor is Lightning Source and they've certainly done a wonderful job of making my titles available. But the material is, for the most part, very narrow interest; the kind of thing where, if you are not interested, you won't download it for free. Sampling is a way to overcome that since it lets customers see whether or not the information is what they are looking for. So, in return for a share of ad revenues, I allowed Google to sample 20 percent of 66 different titles. So far they've put up eleven of them. Two of them have actually had a slight increase in sales, but this may be a coincidence. It is too soon to tell since sales reports take a long time to work through the system. But I have already earned a small amount of ad revenue. I have opposed Google's wholsesale scanning of library books, in my "The Fight For Copyright" blog and elsewhere, but the original program seems to be quite sound. It, like all electronic publishing, is a work in progress. Let me point out that Google Book Search is a feature of Google and doesn't appear on other search engines and that Google is not the whole world here. My titles are available in many other venues, espcially Amazon in all of its variations.
I'm selling some copies every month. I expect that Google will eventually cause the number sold monthly to rise since customers will be able to browse PART of the text. The trick here is not to give away the whole thing.
I invite you to take a look at my titles to see how this actually works.
Francis Hamit, Owner
Francis Hamit Electronic Publishing
I think I should provide a follow-up on Google Print/ Google Book Search and how it all worked out.
First of all I have withdrawn from the Google Book Search Program. Here is why:
Google Book Search never put up more than 19 of the 66 titles provided. When I inquired, once more, why they were not doing what they promised, I got a reply from someone named Dan that their technical experts could not work with files shorter than six pages. This was the first time this objection had been raised. We had taken a lot of trouble getting those files into their system. Over the year that we were in the program, sales of the titles that were "live" there was no significant increase in sales from the sampling of those titles. That might be because the link to Amazon.com never worked properly. As for the promised revenues from advertising, not many ads were sold on those pages and the revenues from those that were amounted to a few pennies a day at most.
Beyond that, Amazon.com made a decision that will also have a dramatic impact on future sales. Since 2001 they have used Lightning Source, Inc. as their exclusive distributor. With very little notice they terminated that agreement in favor of MobiPocket, which they bought last year. Amazon.com, with its global reach, was the source of most of our sales. Converting those files to MobiPocket would require a lot of time and effort which we simply don't have. Nor, based upon our sales to date, can I make a business case for doing so. Moreover MobiPocket is still a French company and their contract states the deal is under French Civil Law.
In other words, we have no practical protection against a business partner which has already acted in a very capricous manner.
That convergence of events does not affect the availability of our titles, which are still offered for sale at Powells.com, Elibron, Fictionwise, Diesel E-books, and many other online retailers in PDF and MS Reader. We will even continue to introduce new titles in those formats, which everyone has or can get. We view Amazon.com's attempt to force people to use MobiPocket as a marketing mistake of "New Coke" proportions. But that will not affect my other business with them. I am a contributor to the Amazon Shorts short literature program, I have a sideline as a dealer in rare and unusual books with Amazon Marketplace and I still own shares in Amazon.com. None of that will change.
These events seem to indicate that the market for old repurposed magazine articles may not be very large. Given the competition from the electronic databases at local libraries, there may not be sufficent demand to justify this kind of distribution. Given the limited time available to me and my assistant, we'd rather be creating new material than repurposing the old. I've always said this was an experiment, but, to be honest, it was also a legal manuever in the copyright infringement cases I recently concluded. You can't lodge a complaint of "Unfair Competition" unless you make some attempt to sell the material yourself in the same markets. The experiment continues. I've just lowered the price on some titles to see if that will increase sales volume. Price point has always been an issue and ours were the result of an informal minimum imposed by Amazon.com. Just as a consumer, I like 49 cents, their price for Amazon Shorts titles, than $1.95.
Going back to the Google Book Search program , I have a sense that it is operated by a team too small for the task and that, like most of Google's initatives, it was neither well planned nor thought out for the long term.
Electronic publishing will continue to evolve, but I doubt it will ever replace print. Electronic publishing titles that can be printed and read like regular books do better than those which can only be read online.
Thanks for this excellent update - I will put it on the front page, and hopefully we can kick off a larger discussion.
It's interesting to apply this reasoning (content providers will benefit from Google Print, but Google will benefit more, ergo Google Print is unfair) to indexing of closed-content on the Web (like old New York Times articles). In the latter case, many content providers who charge for their content want exactly what Google Print is providing - to be indexed and to have snippets of their articles come up in the appropriate search results. If one is to argue that this is because Google is already the dominant search engine, and that this is somehow a sub-optimal situation, then it should be possible to propose a better alternative. In my mind, the solution that the content providers would create would be a bunch of separate proprietary systems (or else one index but with only the major content providers). This _might_ be better for the big content providers (keeping out the small ones and keeping more of a status quo situation), but not for anyone interested in increasing the availability of information.
Also, one economic aspect that is not considered in the above analysis is the cost to Google of scanning older (but possibly still in copyright) books. Another consideration: the future dominance of Google Print is in no way assured. Google is investing in the hope of an eventual pay-off.