Bargaining over links

In a recent blog post internet investor and entrepreneur Chris Dixon brings some much-needed microeconomics to the heated aggregators-vs-publishers debate. But although he asks the right questions, I think he gets to the wrong answer. In a nutshell, he claims that if Google were less dominant publishers would be able to demand payments from it. But even if there were only lots of small aggregators and no dominant Google, I think aggregators would still have the upper hand in negotiations with publishers. The reason is that aggregators, far from being merely parasitic tapeworms, create value for consumers.

Chris starts by attacking the naïveté of the claim that simply because they receive (monetizable) traffic from Google and the like, newspapers should just accept the favour and be thankful. He notes that

Newspapers, like all websites, are suppliers of content to Google. In most markets, with genuinely competitive buyers and suppliers, the revenues are shared between buyers and suppliers in proportion to their relative bargaining power.

When the NY Times agrees to accept traffic from Google (implicitly or explicitly), both parties stand to make some profits that they wouldn't have made otherwise. What Chris is noting is that in general, in transactions like this, which of the two parties makes most money is a matter of bargaining power. If the Times had more power it could get a larger share of the pie simply by demanding that Google make a payment (or else stop linking to the Times). Bargaining power comes from being able to walk away from the deal with minimal cost, and here the Times would lose a lot more than Google if all the Google-Times traffic were to stop. The reason for this is that, as Chris notes, "If the NYTimes decided to opt out of Google tomorrow, Google users would barely notice," while the Times would see a serious drop in its traffic.

So far so good. But Chris then argues that Google's bargaining power comes from its market dominance, and here I think he is wrong. Imagine – as Chris does – a world with a competitive search-engine industry in which no single player had more than a 5% share. Consider the negotiation between one of these and the Times. If the Times were to talk away it would lose a full 5% of its search-engine traffic; on the other hand, the small search engine would barely notice the drop in traffic – because, again, its users would barely notice the difference. Its doesn't matter that the search engine is small; what matters is that the loss is all the Times'. Neither party can seriously hurt the other, but only one can inflict pain.

(To be fair, I don't think Chris fails to see this, but he avoids this conclusion by positing a world in which search engines "differ primarily according to which content sites they index". To me that seems implausible as a starting point precisely for the reasons above. Note that there were no exclusive deals between newspapers and search engines even before Google's rise.)

This points to a fundamental asymmetry between aggregators and publishers, one that stems from the fact that aggregators give publishers traffic that otherwise they would not have. This means that publishers can't just "walk away" from the negotiations without taking a bigger hit than aggregators.

In turn, I think, this stems from the fact that aggregators satisfy a fundamental consumer need that newspapers don't: the need to have, arranged in front you, a list of links to content from across the web that is relevant to a search query (Google), a political persuasion (Huffington Post), a professional community (Digg), or a social group (Facebook, Twitter). It is a need not just for good content, but also for good curation that is editorially and commercially divorced from the content it aggregates. Newspapers can't offer this, and this means that aggregators are here to stay as a separate link in the value chain of news. They create consumer value and it is only natural that they should capture some of it.

***

Even if you accept this, at this point you might argue that as an industry aggregators don't necessarily increase the pie. It may be true that aggregators are good for consumers but if they didn't exist at all people would be forced to go directly to newspapers' sites. Who is to say that the resulting traffic would be any less than what newspapers get today with the help of aggregators? Are aggregators good for newspapers?

At one level the answer doesn't matter. Business is not fair. Value-chains get disrupted all the time by new entrants who can offer something new. As long as this isn't because of collusion or an abuse of market power it is all fair play.

But the question has some bearing on the emotional debate about "fairness", and also on corporate strategy. To take an extreme example, imagine that all the big newspapers merged or got together in a cartel and negotiated with aggregators. If aggregators as a whole were not bringing any new traffic, newspapers could walk away from the negotiation, block all links, and aggregators would suffer a serious hit (while newspapers would be unaffected). Aggregators would have to concede and pay newspapers. If the money is good enough, this would be a reason for newspapers to merge or at least collaborate (whether the latter would be legal is a separate question).

More realistically, if a significant portion of the traffic that aggregators give newspapers is genuinely incremental then things are less clear. If the extra money that newspapers make due to aggregators is roughly equal to the money that aggregators make thanks to newspapers, then a sensible compromise would be to "split the pie" so that each party keeps the money it makes from advertising and neither pays the other – as is the case today.

Note that this is not as far-fetched as it seems. You would expect newspapers to monetise the visitors they get from aggregators much better than aggregators themselves. Newspapers have better ad units, show more ads per page, and entice some of these visitors to stay and read other content. They also get to turn the occasional visitor into a regular reader or even a subscriber. So, say that newspapers collectively monetise aggregators' users 3 times better than aggregators themselves. Then if only 1/3 of aggregator-generated traffic to newspapers is genuinely incremental (over what newspapers would get anyway) then both sides would be making similar profits to each other and today's "no payment" scheme could well be the outcome of a negotiated solution.

Clearly, assessing whether this scenario describes the status quo is a difficult if not impossible task. But to the extent that it is plausible, this exercise suggests that newspapers may be wasting their time when they ask governments for permission to negotiate jointly with aggregators. Even if they are allowed, the negotiations may well lead nowhere.

Comments

[chris dixon]

The referenced Chris Dixon here. Interesting argument. I'm not sure we disagree. I was really just trying to debunk the argument I saw frequently that: Google must be helping - otherwise why not noindex your site.

I think your argument includes the premise that aggregators increase the pie (total traffic to news sites). I'm not sure that's the case.

[Nico Flores]

Chris,

Thanks for your comment. I don't think my argument assumes that aggregators (singly or as a group)increase the pie. The only assumption is that aggregators are (singly) in a position to take traffic away form newspapers; but that doesn't mean that they created any of this traffic (singly or as a group).

To clarify this, I think there are two different things here:

1. If a given aggregator were to stop linking to a given publisher, would the publisher stand to lose a lot more than the aggregator?

2. Does a given aggregator increase the pie available to it and a given publisher?

I think (1) is generally true, even if (2) is not -- that is, aggregators are individually strong vis-a-vis publishers even if they don't increase the pie.

To see this, imagine a world with 20 aggregators, each with a 5% market share. If one of them went out of business, most of its users would go to one of the remaining 19 aggregators, all of which would continue sending people to newspapers as usual. So newspapers wouldn't notice any change, which means that the pie wasn't increased by the marginal aggregator. However, if the same aggregator stopped linking to a given newspaper (instead of going out of business), the newspaper would experience a 5% drop in its aggregator traffic (by the argument in my post).

In other words: _given_ that there are aggregators out there, and assuming that there is a segment of the audience that relies on them, then any aggregator is in a position to hurt a newspaper at almost no cost to itself.

Yes, there's a second assumption here - namely that once people are in the habit of relying on aggregators they'd rather switch between aggregators than go back to newspapers' front pages. This could be because aggregators either (a) just stole some of the pie from newspapers, (b) increased the pie, or (c) a mixture of the two. Which of the three it is doesn't matter for the strategic picture, but it may matter in the blame game.

I may get around to writing a new post with all these points - I can see that the one above isn't clear on them.

Best,
Nico