Data And The Future Of The Web
After I asserted several times that data is the key to the future of the web, Umair Haque gave my head a good spin by asserting that data is in fact a commodity. Umair is half right — we are increasingly overrun by data, and SOME of it is a commodity.
Context & Ripple Effects
The commodity argument has been running through the blogosphere for months: A VC's claim that web startups are a commodity in October 2007 drew the same line Umair Haque now draws one layer down — if the code is a commodity, so is much of what the code produces. Haque's rebuttal to Publishing 2.0's 'data is the key' thesis is consistent with his longer-running frame: back in November 2005 he was already arguing, via Herbert Simon, that attention rather than information is the scarce resource.
What makes the exchange more than semantics is that user data has just become visibly contested property — Jason Calacanis's November 2007 post on Facebook users and their data showed platforms and users fighting over who owns behavioral records. Karp's concession that 'some' data is a commodity implicitly splits the market into abundant raw data and scarce structured or owned data.
First-order effects
- Startups pitching proprietary datasets as their moat face a direct challenge: if Haque is right, their defensible asset shrinks to whatever filtering or relationships they build on top of undifferentiated data.
- Publishers like Publishing 2.0 must decide which side of Karp's own split they sit on — treating their content-generated data as a commodity input or as an owned, differentiating asset.
Second-order effects
- If raw data is commoditized, competition shifts to attention and curation — exactly the scarcity Haque flagged in 2005 — pushing platforms to compete on interfaces and filters rather than collection.
- Ownership disputes of the kind Calacanis described at Facebook spread beyond social networks: any service holding user behavioral data faces pressure to justify its claim over records users arguably created.
Third-order effects
- The debate prefigures a structural question about the web's incentive system: whether value accrues to those who generate data, those who aggregate it, or those who own the rights to it — a question that touches platform economics and eventually regulation.
- If 'some data is a commodity' hardens into consensus, the industry stratifies between commodity data layers and scarce proprietary layers, reshaping which web businesses are defensible at all.
The trend: Web value is migrating from accumulating data toward structuring scarce attention around it, with data ownership becoming the next contested layer.