I don’t get it. Consider this wonderful quote via Ars Digital:
Microsoft has been tracking this information for years through its various sites, including MSN, Hotmail and others, keeping a vast database on tens of millions of individuals, each assigned a user ID Microsofties refer to as a GUID, or global user ID.
How refreshingly straight forward that is.
The drivers for this kind of stuff are very strong. The demand for models, even lousy models, of users is high. The cost to accumulate them is very low. All the major web properties must have this kind of data and the skills to do the model building aren’t rocket science. Amazon is quite up front about it. Double click and others have has been doing it for years.
So I don’t get why this doesn’t appear to be more wide spread and obvious a practice. I shop for tires, say, I don’t currently see tire ads appear through out the rest of my day’s browsing. Why not? Don’t get me wrong, I’d love to find some scheme to temper this erosion of privacy; but that doesn’t mean I’m willing to pretend I don’t see how strong the drivers are.
The question faced by all these operations isn’t persay about privacy – they are not in the privacy biz. Posssibly it’s about a combination of social contract and market maturation. But that doesn’t really explain why the feedback loop closed yet? I reveal a clear demand signal, for tires, but the advertisers are not receiving and actting on that signal. The data is there, the demand is there, so where is the breakdown?
Amazon demonstrates that you can close this loop. If I reveal an interest to Amazon they rapidly customize things to address my presumed desire. It’s actually kind of fun to toy with them. They have solved both the technical and the social problem. As far as I can tell eBay hasn’t.
Why haven’t the big site independent advertiser brokers (google, double click, etc) closed this loop. I guess there are three hypothesis. A. They have and I’m just blind to it. B. The market just hasn’t gotten there yet. C. There is some social contract they haven’t figured out how frame up right. The difference between B and C is slight; one is technical engineering and the other is social engineering.
I don’t get it.
Years ago when I worked in that industry, there were a few companies doing it. Matchlogic (owned by excite and went down with them) had a pretty cool system to keep track of I-forget-how-many millions of users. The trouble was that it, at least at the time, actually wasn’t cheap to run. It took them a relative large amount of effort for relatively little gain (in clicks/acquisitions/signups/etc).
Around the same time I recall that DoubleClick were winding *down* their system doing it. The story I heard was that it did work (increased the results by something measurable) but their technology cost to do it made it not worth it in the end.
Of course it’s some time ago and with Moore’s law and substantially cheaper hardware and storage costs, who knows…
Hmn, maybe I should go back and work in that industry again. 🙂
Interesting, so that means there is almost certainly a patent thicket in the way of any new entrants.