View Single Post
  #2  
Old 02-14-2002, 10:33 PM
Guest
 
Posts: n/a
Default firsthand information



You gave an example about how a big player showed up out of the blue, and moved the volatility. This left the analysts scrambling.


There are several ways to get this information. You can

1) wait for it to hit the tape,

2) pay strategists and traders to deduce it,

3) pay salespeople to collect it, or

3) pay clients to deliver it.


Inadvertently, using expensive manpower, you do pay clients to deliver it. Talking to traders and salespeople, and submitting to friction, is the only way they can acquire proximity to whatever unique value magnet you control.


Put differently, you have to deliver ever more fantastic services, to subsidize the manpower. In other words, the information topology is not, like, self-inflated. Like, you you aren't paying the client directly for information collection.


Why must I attack this topic at night, when my mind is totally burnt out!? Let me clarify what I mean by "topology" while I collect my thoughts:


http://www.twoplustwo.com/cgi-bin/ne....pl?read=35263


You are manufacturing the service out of information, gathered and processed by people. I say, you need people to process it, but not to gather it. And you can also make it cheaper for the client to deliver it, through some type of better automation. When it is cheaper for the client to deliver it, and he is paid directly based on quantity and quality - rather than indirectly, through some lumped-together analog carpool of services - the supply will be better regulated - price-regulated like everything else in the economy - towards optimality.


But let me digress into some of my thoughts on "quality" people. I once worked in the customer service department of the major manufacturer of a near-monopoly product with daily delivery and a short shelf-life. It was my job to lean on a list of several hundred smaller clients, on the phone, so as to squeeze their demand information out of them in time to coordinate production and delivery logistics with the plant, the distribution nodes, and the fleet drivers.


Notice, we did not give clients any price breaks for getting their orders in on time, all I did was lean on them - and even sweet-talk them - to try and get enough out of them to punch into the coordination database. There was a cutoff point, past which they might not end up with the exact inventories they wanted but, short of that cliff, they bore little cost for postponement and silence. But that is not the point of my story.


The point is that our department was staffed by three decent people working fairly hard. I thought it could have been staffed by two excellent people working very hard, or one absolute freak, using three headsets and two Nextel DirectConnect pagers at once. But equilibrium for the department was four people. Why?


Because four retards are more flexible, more adaptable, more resilient, than one freak. You can afford to pay each of them half as much as the freak - for a total of twice what you would pay the freak - and still come out ahead. If one of them is sick, they can still function. If one gets a better-paying job and ditches, they can still function. If their job description and necessary talents change slightly, they are more likely to have someone who can adapt, or you can have a smooth transition by firing and replacing just one of them. You can plug and play.


In other words, I think multi-person structures, acting in a sort of ecosystem like dumb microbes, are actually smarter and more adaptable than high-quality super-people, like at Goldman Sachs. I think, with a combination of the right groupware, the right information-payment topology and pricing metric - and a minimum number of retards to staff it - you can actually create a superior biological machine which services clients better, and costs them less in time and hassle at their end.


But I think the key is automating the incentive structure, to encourage and discourage behavior patterns on part of the client. Today, you basically massage them and steer them into doing what they need to do for the whole system to work. But, like you said, it takes mega manpower on both sides to man the interface, and dance the dance without stumbling or falling out of step.


More advanced automation also is a way to connect many small brains into one big brain, where individuals are pruned out or reinforced in like synaptic connections, devoted regions, and hemispheres. Automation fine-tunes and delivers the correct incentives and constraints to each point in the chain, even where the coordinated points are inside a single institution. People don't always have optimally-synchronized aims even with the same organization


I think, the utility of the inputs is not conveyed to clients, in the form of a structure of behavior incentives, in as optimal and streamlimed a fashion as it might be.


Maybe I'll try again tomorrow!


Thanks, Javelin, for keeping the topic alive, and giving me a chance to get it right - someday.


I need a good 40-hour session of 30-60 - followed by a 12-hour drive with loud music and radio fuzz - to put my brain right. THEN, I'll tell you what I'm talking about, as clearly as you have told me. Maybe being tracked by an endless/video-game stream of expendable Haitian assassins would help...


Know any?


eLROY
Reply With Quote