A thread on this interesting piece by Hagiu & Wright on the data advantage. hbr.org/2020/01/when-d… @theplatformguy #bigdata #antitrust 1/10
The piece makes 2 key points: (1) there is a data advantage only under some conditions & (2) the data-enabled network effect is less powerful than the standard network effect 2/10
My views: the first point is uncontroversial. The second point is not necessarily true in all circumstances, and anyway it doesn't necessarily matter when discussing the data-enabled network effect 3/10
Where it is true that the data-enabled network effect is less powerful than the standard network effect, there is an important implication: interoperability is more important than data access. See for example this piece on interop: knightcolumbia.org/content/protoc… 4/10
The Hagiu & Wright piece is a nice change from usual anti-enforcement arguments (ie the conditions are often or generally not met, data is not like oil, it's more like sunshine, data experiences decreasing returns, and the data network effect is not really a network effect) 5/10
Hagiu & Wright contradict anti-enforcement arguments: (1) data accumulation even with decreasing returns can still improve the product dramatically (I made this point many times, see mlex.com/GlobalAntitrus…) (2) the data network effect is a type of network effect 6/10
A missing point in Hagiu & Wright concerns quickly outdated data. They note this means new entrants don't need large volumes of historical data. But it also means that where there is mandated data access, it should be access to a real-time data flow, to ensure data freshness 7/10
The Hagiu & Wright piece should be read in combination with current commentary on the data advantage. Many pieces have an agenda or a headline message (sometimes oversimplified) but are actually nuanced and provide arguments in both directions (sometimes unintentionally) 8/10
E.g. this A16Z piece suggesting that data moats are overrated, but advising to "secure proprietary data sources" as a "moat against competitors" a16z.com/2019/05/09/dat… 9/10
Finally, my own standard slide on this topic [end] 
