Monthly Archives: February 2008

EBay changes

Leave a reply

Fascinating. EBay is eliminating seller’s ability to give buyers negative feedback. Nicer for buyers, worse for sellers. That’s really the tip of the changes. Amazingly they have noted that sellers leave negative feedback eight times more often than buyers. And it seems like they are going to substitute some more private in house means of giving sellers a way to negatively effect buyer reputations.

I love the PR speak: “Donahoe revealed that eBay will update its feedback system to reinforce healthy, vibrant trading and keep bringing buyers back to eBay.” It appears they have noticed that when buyers get negative feedback they don’t come back.

Markets work if there is a balance of buyers and sellers. So, presumably as market maker eBay had decided it to make it more attractive to buy. This is analogous to how the credit card companies presume the buyer is right and the seller is wrong when disputed payments arise.

Buyer reputation isn’t as actionable as you might think. Sellers must accept bids, even if the buyer has a lousy reputation. Buyer reputation is more important if the buyer is also seller; but that’s pretty rare.

Watching the soon to be CEO of eBay speak to their top 200 sellers at their eCommerce forum is fascinating. This is an elite club, no doubt they account for a huge slice of eBay’s income. He, in effect, tells some of these 200 that they won’t be kicked out of the club. The real meat of what they are thinking is hidden in this talk; though you have to suffer through a lot of organizational change. All that change is a signal of how significant the changes they are into.

For their elite sellers they are setting up very significant carrots and sticks (they call that motivate and reward). They appear to be very concerned about how some bad sellers are causing buyers to flee. Discounts on fees for well behaved elite sellers. Search engine ranking that disadvantages lower reputation sellers, elite and other.

It is interesting that they are setting standards to raise the tone of the market place. This is one of the kinds of standards settings that I find most interesting. For example how the airline industry once agreed to set a floor on exactly how lousy a “sandwich” can be. I’m more interested in the collaborative cases where floors, for example professional ethics, are introduced.

Apartment Social Sites

1 Reply

This is mostly a publicist’s piece in about a web site that sells to apartment building owners a web site where their tenants can network.  I’m impressed that the vendor can sell these to landlords.  Most landlords know better than to let their tenants help organize against him.  I recall as story from years ago about a landlord suing a site because his property had such awful ratings.  There is a nice success story near the end of the article about a child birth counselor getting a lot of buisness by advertising via the site.

My town of 40 thousand people has a surprisingly successful email list.  It survives because there a tremendous pool of people here who know how to manage an email list of this kind.  Replicating that success, using volunteers, would be hard.  In time such skill will be more common place, in the meantime I find it somewhat more plausible that a firm could do it.

control is failure

Leave a reply

Great quote.

You’ve just got to keep reminding yourself that control is failure. — Daniel Taylor

That probably sums up my theory of both parenting and technical management.
The quote appears in Deep Economy.  Taylor’s president of Future Generations a 3-4 Million dollar a year nonprofit working in Afganistan, China, Peru, and India on programs.    The idea is to create seeds by aiding communities to activate problem solving on a problem, and then to replicate that across regions to achieve scale.

Here’s another quote: “Change happens not because of how we invest our money.  Change happens because of how we invest our human energy.”  I don’t think that’s quite right.  It’s right to focus on the human energy; coordinating that is the hardest part.  A franchise business model solves this by creating a rigid structure and channeling the labor through that.  But what does he mean by “our human energy;”  if we means by that the community members then yes.  If he means by that the organizers then not so much.

The first quote is a delight because it’s clear he knows that when you create these systems you can’t prescribe how the human energy will be coordinated; you can only attempt to create a context in which that coordination emerges.  This is particularly necessary when you maybe uncertain what problem your solving, how your going to solve it, and what issues will arise as soon as you get into it.  At that point you want frameworks that enables the human energy to remain engaged.  At that point it’s unlikely control is going to be much help.

Oh sure, there are of course are plenty of systems where control is a means to success.  Your lucky if you’ve got one of those, since such systems scale more easily.  Once you get them blocked out you can replicate them: some standards, some mass produced capital equipment, some training, some quality metrics, some architectures of control.  You see lots of that in large scale franchising.  And it appears that to a degree the folks at Future Generations do just that.  If their seeding attempt succeeds they attempt to replicate the pattern for solving a problem developed in the first community into others.  Stamping out of duplicates.  But the problems they are grappling with requires much more adaptable problem solving schemes; so I suspect that the duplicates feature substantial variation.

Interesting.  This is a kind of grass roots franchising model.  If one community puzzles out how to run a farmer’s market, a community tool shed, a library, a wifi network, knitting circle, etc. etc. then others can mimic that.

Notice how there is a two step process here.  In the first phase you need to attempt to get a solution to emerge that works for community members, at all.  In the second stage you want to strive to get it to replicate.  These probably call on different kinds of entrepreneurial skills.  But the build out is where you can create really significant change.

Snap

Leave a reply

Boy, I’m late to this party! Freebsd’s usual file system, UFS (unix file system), has snapshots. It’s had them for years. I wonder why MacOS doesn’t appear to have them? In anycase, these are great!
There is a nice tool kit for managing them be found in ports freebsd-snapshot. Written by an old friend too. That gets you 95% of the value of a netapp. So, now I have hourly and daily snapshots of all my main file systems and can trivially recover lost files and do diffs of things like log files. e.g diff /snap/var:hourly.1/log/messages /var/log/messages.  Suddenly it is easy and cheap to see what happen recently.

This is also a big help in doing proper backups; i.e. you can backup snapshots. Orchustrating a good snapshot for backup purposes is a nice programming puzzle. You want to tell every running process to flush it’s state to disk and to lock out changes during while you make the snap. That’s a lot easier if the file system supports snapshots. Of course this flush to disk and lock intervals should create a suitable restart/recovery point for the application. You can see example here for doing just that with mysql as the application.

Disintermediation: the landlord

Leave a reply

Here’s an amusing thought.  Quoted over at calculated risk: “We’ve attracted a lot of borrowers who are really renters  It is disheartening as a servicer to see the willingness [to walk away]  [borrowers] simply don’t care.”

Ha!  The bankers have disintermediated the landlord; and like usual when you cut out a middleman you discover, a while later that he had a function which surprise surprise you now have to fill.  And since you haven’t been filling that role well to date there are more surprises to come.  No doubt there is a business opportunity there; a variant of the condo management companies, but this time beholding to the bank rather than the condo board.

Holding the Bag

Leave a reply

The Irish decided to impose a 33 cents tax on those plastic bags that retailers use. I presume this was in part a sin tax, in part an attempt to tax the externalities created by the bags, and in part a bit of stunt. The end of the story is that it’s wiped out the use of plastic bags, shifted social norms about bag usage. It’s a fascinating story about deciding to change the way a society behaves; making it acceptable to shun those who exercise their freedom to use plastic bags.

For some reason that story is currently slotted in my mind with the health care debate’s latest point of discussion.

The question at hand is should we put everybody into the system, or should we allow people to opt out. Allowing people to opt out would presumably create incentives for people to game the system. Or putting it another way it would allow people to gamble that they aren’t going to need health care during the next time interval.

Boffins think that this single choice would have a sharp effect on costs. Plans were we all join look likely to cost us each about $2,700 a year, while those were we allow gaming would cost $4,400 per year. Intuitively that sounds about right – you just take a guess on how many people would decide to take a swag at getting away with no healthcare, say 25%, and then you figure that group is likely to be somewhat less likely to need care, say 60% less; an you get similar numbers.

So that’s 63% more expensive or additional $147/month. Our faux progressive funding system pushes most of the cost of these social programs onto the middle and upper middle class so you can multiply that as you think is appropriate. But that class struggle is less interesting to me today.
What this highlights is how it creates two classes: those who decide they want to take the gamble, in one class, are costing everybody who decides to join, the second class. It is an interesting case of creating a clear cost to the majority group by allowing a minority a freedom. The boffins think that minority is pretty large. Will those who join will shun those who take the gamble? Will the system create shifting social norms, like the Irish experience with the bags, where it becomes unacceptable behavior to take that gamble?
There is another interesting take though. An intertemporal one. I presume that many of those who might decide to take that bet they won’t get sick will be young people. There are plenty of reasons to think that. Young people tend to be less able to afford the insurance. They tend to be less risk adverse. They tend to be healthier. They tend to have less common cause with large institutions. Over time a person is two people; a young person, and then latter a older person. The first of these is raising the costs of the second one.

Actually there are really three people or more people in each individual, since for example, what makes for a healthy young person is a healthy childhood. So the healthy young person that opts out has his head in a bag, selectively blind to both to the past and the future.

Dutch Baby

2 Replies

Having gotten back into making popovers has lead to remembering there is a variant that even easier; where you make them in a cast iron frying pan or dutch oven. It’s also pretty amazing as it comes out of the oven. Same recipe as popovers. We always called these David Eyre’s Pancakes when I was a kid. They it must be a very old dish with many names, for example Yorkshire Puddling, but my favorite is Dutch Baby.
pan_popover.jpg

Two Groups: rich and poor

1 Reply

This is an interesting post on income inequality across the nation’s states. What it adds to his prior posting is a brief synopsis of Jamie Galbraith’s model of what’s going on for the folks at both ends of the spectrum. I find it notable that both groups live on a raft.

The rich raft is the stock market – as the market rises and falls so do they. This says two things to me. One is that the rich are not easily distinguishable from firms for the purpose of modeling, i.e. that distribution of firm size and the distribution of individual wealth are probably the same. But more importantly, as a group, the rich have common cause in creating economic conditions that are, in short form, whatever is good for the stock market.

The poor live on a raft of social programs: the social security, food stamps, the earned income credit, and in the poor states the federal minimum wage. That should create common cause across that vast group to support state policies to improve and maintain those programs.

What I find notable is that in both cases the common cause of the two groups is about the nature of how the state engages with the market as a complementary actor. That again is the model exposed by the voteview work on legislative voting patterns. E.g. that the economic left/right axis practically one to one with the small/large economic actors.  Those on the left looking to provide complements for the small actors and those on the right striving to aid the large ones.

Big Horns

Leave a reply

Before we all got excited about building systems that would leverage Reed’s law we were into systems that leveraged Metcafe’s law, and before that we all wanted to buidl systems that tapped into Sarnoff’s law.  Radio with it’s one broadcaster to many listeners is the classic example of a Sarnoff law system.  Of course before radio you had public speakers.  If you want to build one of these older systems, say by starting your own mega-church, you’ll be wanting some big speakers.  Maybe if you watch this video you can make your own, save some cash for refreshments.

nss_mdns

Leave a reply

To add to a Freebsd machine the ability to resolve domain names managed by zeroconf you can install the port for nss_mdns, and then modify the file /etc/nsswitch.conf. The modification should insert the token mdns on the line in nsswitch.conf that configure the search order for hosts:. I inserted it just after file.

I wanted mdns so my backup server could refer to the laptops by the names they inject into the mdns information cloud.

The port doesn’t provide the latest version (today, i think) which makes avahi complain that it’s not installed. Avahi is a mdns server, and you only need avahi if you want to publish additions information into the multicast dns cloud; or as I want to do reflect mdns accross subnets; think vpn. The libraries are so.1, not so.2; which the avahi codes desires; leading it to complain: “WARNING: No NSS support for mDNS detected, consider installing nss-dns!”

While I was a bit puzzled that the way to add mdns support to your environment was to inject something into the C library, which is what NSS does, I’m getting over it. At this point it almost seems like the right approach.

Update: Sigh, it stopped working … i have no idea.