Re: [projectvrm] Is VRM an Ideologically-inspired Dead-end?

[Joe Andrieu < >]

On Mon, Mar 18, 2013, at 10:28 PM, Don Marti wrote:
> begin Joe Andrieu quotation of Sun, Mar 17, 2013 at 08:54:44PM -0700:
> 
> > But my real point was conditional: If it turned out to be true that
> > people can't succeed with intentcasting, then it is necessarily
> > true that vendors can't succeed with tracking.
> 

That was actually Chris Savage's quote, as he deconstructed what I think
was original Graham Hill's argument.

There's all sorts of reasons that (non-self-) tracking might do better
than intentcasting. Privacy concerns are a good one, although we could,
in theory, ameliorate many of those if we had robust vendor identity,
for example so our intentions could be read only by those certified to
be in that business, and if we had some reliable way to assure the
information would only be used as intended...

But even larger than that is that pesky law of large numbers. Intent
casting is higher fidelity in the micro, but until it hits the macro of
at thousands of users in a given sector, there's very little trend data
to figure out what's missing in any particular thread.  Tracking happens
everywhere to everyone. It's leverage of the law of large numbers is
great, even as it's privacy violations are unacceptable.

What's best about intentcasting versus tracking is that it shifts the
analysis from an open ended question looking at every piece of data
available to a specific inquiry with highly focused data. There will
still be guesswork, but it will be hyperlocalized. Instead of companies
like BlueKai and Audience Science monetizing the thrilling fact they
they think you are pregnant and looking to buy a car, intent casting can
make it explicit that you are in the market for a specific type of
vehicle. The shopping experience isn't done, you may still change your
mind, but as you close in on your particular ideal car, vendors
can--theoretically--stay right along with you, giving more relevant
options and learning more about your underlying needs. 

When lots of users start working with this kind of data, the *system*
can start to predict "reliably good results" surprisingly well. It's
like Amazon's "people who bought this also bought these other products".
It's not supposed to be 100% advice for you to also buy those things,
but it is often highly relevant information. Get intent casting to that
level of penetration and things will start to get interesting.

Perhaps best of all, once you buy your car, they can stop trying to sell
you one.  Talk about annoying waste.

-j

p.s.
Yes, by "the system" I mean one or more vendors. As came up in the other
thread, it's when you can aggregate demand that the power of analytics
shows up in force.

On Mon, Mar 18, 2013, at 10:28 PM, Don Marti wrote:
> begin Joe Andrieu quotation of Sun, Mar 17, 2013 at 08:54:44PM -0700:
> 
> > But my real point was conditional: If it turned out to be true that
> > people can't succeed with intentcasting, then it is necessarily
> > true that vendors can't succeed with tracking.
> 
> Not necessarily.  Tracking could be revealing
> information that the user would not choose to share,
> whether because it's too much trouble or because of
> the user's privacy concerns.  And not all purchases
> are planned or mindful.  Maybe most aren't.
> 
> -- 
> Don Marti                      +1-510-332-1587 (mobile)
> http://zgp.org/~dmarti/        Alameda, California, USA
> 
 


-- 
Joe Andrieu
SwitchBook

 
+1(805)705-8651