Adword Arbitrage

    March 8, 2005

Seems appropriate on the eve of the boom’s anniversary to revisit arbitrage and bubbles.

Jeff Nolan points to a definition of arbitrage (risk free profit), but in the context of a poor analysis. Jupiter analyst Niki Scevak takes issue with an article by Bambi Francisco on how keyword arbitrage is leading to a bubble in Google Adsense valuations. Unfortunately, the article is behind a costwall, but Niki’s post summarizes:

Bambi cites Nextag as an example of a ‘search arbitrageur’ because they bid up the term ‘dvd players’. Yet Nextag does not buy the keyword and simultaneously sell it for a riskless profit. I can’t say enough that this assertation is categorically, absolutely and unconditionally wrong (again to put it mildly).

Nextag takes a generic keyword, filters that user through the decision of what make and model, and often what price range the consumer is willing to pay for their dvd player, and then sells that more qualified lead to merchants. It furthers the consumer through the purchase funnel. Extra value is added. Ask any merchant bidding upon the term ‘dvd player’ and they will tell you it performs poorly, in terms of direct conversion.

But the term ‘dvd player’ is more valuable to vertical search engines like Nextag than it is to merchants and so Nextag can afford to pay a higher price than can merchants.

I won’t pick on Niki’s analysis or assertion that Bambi’s article is horribly misleading information, but instead explain how the Nextag example is arbitrage at work.

Say the keyword ‘dvd player’ costs $1 on Google’s auction market. On Nextag’s market it costs $2. Nextag does add value, by decreasing search costs for vertical merchants, say by $0.50 (lets not drift into a discussion of how they are extending the Long Tail, but this is about Fat Tails). Nextag also does have its own transaction costs, say $0.50. Overall, Nextag pockets a profit of $0.50 for their ‘value add’ — but if you assume the transaction was straight through processed and there were no credit or operational risks, this is risk free profit.

In this simple example, they are buying from one market where there is a low price and selling in another where the price is higher. How did they do this? By having superior information than other market makers or buyers.

Now assume that there was a delay from when they purchased from Google and when they sold to another party. That would be risk, the price could go up. Common in storable commodities such as natural gas. But on the other hand, it would create the ability to retain inventory as a hedge against price volatility.

So is there an AdWord bubble? Not by the Nextag example. As arbitrageurs, they are doing the opposite, and others will follow, gradually bringing the prices in both markets towards equilibrium. Who knows if the market is in contango (trending up) or backwardation (trending down), there are many other forces at play. As we are figuring out the economics of the long tail, I fall back on Says Law, where the market flocks to both scarcity or abundance to bring it back into equilibrium

Foreward: Five years ago I created RateXlabs with the good folks at Oncept tand published some whitepapers on bandwidth arbitrage conditions applicable to today’s computing commodity market. Finally I have a use for all that information, to share it and hopefully not mislead.

Ross Mayfield is CEO and co-founder of Socialtext, an emerging provider of Enterprise Social Software that dramatically increases group productivity and develops a group memory.

He also writes Ross Mayfield’s Weblog which focuses on markets, technology and musings.