Notional Slurry Logo

Archive for November, 2008

Redisintermediation exemplar: John Cope’s Toasted Dried Sweet Corn

When we lived in Hanover, PA a few years ago, we started buying boxes of a Lancaster delicacy: John Cope’s Toasted Dried Sweet Corn.

It’s good, and difficult to replicate. No other cornmeal or bready preparations are substitutes. There are no doubt a number of delicious recipes possible, but the one that is printed at the top of the box (or bag, these days) is still best, just as using Jiffy Mix for corn muffins is better than the supernumerary suggestions of waffles or even johnnycakes.

Also, in re John Cope’s effort: it’s cheap.

Buy it from the manufacturer, or their distributor, in 12-bag cases, and including shipping it’s less than $3.50 per box.

Zingerman’s Deli, here in lovely Ann Arbor, charges more than $11 for a single 7.5 oz tin. Because it’s artisanal, no doubt. Or maybe the tin is worth the effort, since it’s made by hand by Russian Amish people specifically for Zingerman’s, and flown here sustainably or something. Because that would be a $7 metal tin, I guess.

Plus shipping, if you don’t live in lovely Ann Arbor.

And if you search for it at Amazon, you can pay a mere 100% markup. Plus shipping and handling. Or for some kind of odd bulk repackaging I’ve never seen before, slightly less.

Let’s just sit ourselves down a minute, in these days of local communitarian sentiments and economic crisis and belt-tightening and thoughtful economy and direct compensation of artists and craftspeople for their intelligent work and sustainable transportation and stuff… and think about those alternatives.

Less than $3.50 per unit, net, for twelve you could share among friends. Said money sent direct to the manufacturer, I assume. At least closer to them than any alternative in the supply chain.

Or $8 or more for retail pricing of the same volume. In a metal bin, if you’re really fancy.

I note, by way of a fucking point: It is not illegal, to date, for purchasers to enter into informal agreements with one another to collaboratively seek bargains by sharing informational or practical costs.

I’m going to spend December thinking about that, OK? The whole damned Black Friday of a month.

And my mind may wander from dried sweet corn to beer, or other foods, or books, or magazine subscriptions, or toys, or DVD rentals, or copy-editing one another’s writing, or consulting referrals, or newspaper article writing, or photography, or design, or gardening, or building houses and communities. It might make a bit of sense for me to look at cartoonish John Cope, with his stereotypical bushy beard, and think a minute.

Just one minute. Especially if I’m tempted to play at Black-bumper sustainability, and conspicuously consume artisanal foods without thinking about the supply chain that got them to me.

Don’t present a single solution to a complex problem. Just don’t.

Google Maps must have purchased a new suite of road information recently. Or maybe they algorithmically tried to “improve” the dataset they had. Used to be it knew local geography pretty well; now, not so much.

When asking for directions from our farm (on Walsh Road, Webster Township) to the Dairy Queen in Hamburg Township, the driving (not walking!) algorithm suggests we stay on northbound Scully. If you saw it from a distance, the satellite image would lead you to believe that, yes.

Google Maps algorithm FAIL

Except that many years’ fierce argument at the county border has left a nasty but potent gate blocking the road, which will persist into the foreseeable future.

Google Maps algorithm FAIL (closeup)

If you were to drive up the rough, mainly untended Scully Road on a snowy day, trying to get (say) to a hospital in Pinckney or something, the least that would happen is you’d waste a half-hour trying to back out of the last few hundred yards without ending up in a ditch… once you arrived at the impassable gate at the border, and well after you had trespassed on a private road at the end.

The De Lorme Michigan Atlas & Gazetteer, a nice old printed book I keep in my car, and which is so obsolete that it shows little red lines for roads of all sizes and characters, manages to catch the gap.

Now every dataset contains errors or missing information. But every time that dataset is used to make a single, summary statement, based on a single model? Badness can happen in unexpected ways. In fact, I am obliged to be curmudgeonly about it because of my professional experience in these matters: it is always wrong to present a single answer for any multi-objective or highly constrained decision-making problem. Big, fat period.

I can’t complain, in all honesty, about advice given by a black box operations-research algorithm that on inspection I knew was incorrect. You get what you pay for. But I can complain about a cultivating a misleading user experience in a ubiquitous data-driven decision support system that presents only one solution at a time to the decision-maker. Hell, every iPhone in the world has one of these on it; they’re all wrong, too.

No, I don’t think I am feeling lucky, Google. And you didn’t even ask.

I want to see a sheaf of routes. The little “adjust the route and recalculate a new one using my milestones” handles Google introduced a few years back are a beautiful thing, a cunning artifact and a useful tool! And of course, the standard “avoiding highway” or “fastest” toggles let me reach in and fiddle with the search method. But only indirectly.

I want the objectives right there, not combined. I want not just to surface the meter (to use a phrase Dan Cooney’s taught me), but surface all of them. I want choices coupled to clearly differentiable supporting arguments.

Like the basic Google Search results themselves: ten routes at a time, ranked somehow. Or not even ranked, but handed to me as a Pareto-equivalent set of alternatives, some faster, some bumpier, some with bigger roads, some with more gas stations, some more scenic. Heck, maybe I just want to know there are at least ten ways to go back and forth, so I can stage a race, or not get bored on my commute, or defend against unwanted SUV invasion by a foreign county or something.

At least sometimes. Stop assuming I’m feeling lucky.

Next time, we can bitch about the misleading user experience and illusory authority created by the fuckin’ weather forecast format. Everybody complains about the weather forecast, but nobody does anything about it.

Lather, rinse, repeat.

links for 2008-11-28

links for 2008-11-27

links for 2008-11-26

links for 2008-11-25

links for 2008-11-25

links for 2008-11-22

My personal acceptance test for a community development effort

There’s a lot of local chatter lately (local in my social network, here and afar) about community development as opposed to tribal consolidation. That is, developing meetings and infrastructure that bridge between disparate groups who otherwise never meet and interact, vs. team-building and strengthening the internal cohesion of well-formed groups themselves.

Here’s an acceptance test I’m thinking of using for a long-term project of the bridge-building type (the one I call “real” community development). It’s hard to know whether your notion of how to run things actually fosters and enhances diversity rather than just consolidating pre-existing barriers, so I’m musing about a general-purpose challenge that discriminates them. Maybe:

  1. Pick up a local Yellow Pages.
  2. Open to a random page. You may want to slice the book up into chunks to ensure uniform sampling.
  3. Randomly select an entry on the page, maybe with a blind stab, and note the category it’s in. Plumbers? Lawyers? Dentists? Libraries? Landscaping? Escort Services? Restaurants? Used and Rare Books? Knitting? Jewelry? Wedding Planning? Septic Services?
  4. Repeat the previous two steps to select a second category at random.
  5. If you can create, announce, and populate an open-format unconference-style meeting that will attract at least five people who actually work in each of those two categories professionally, your community-building effort may have a chance.

You don’t need to ever repeat with the same two categories. But it might be interesting to walk ahead by adding a new category and dropping the oldest category in each successive meeting.

This may be a bit of a stretch. Wedding Planners plus Shoes I can see a path to; House Painting and Office Supplies, less so.

But what’s a test without a challenge?

Is this a good time to reveal credit card terms?

Citi just sent us one of those ominous plain envelopes from South Dakota, and (as I expected) inside is a Notice of Change in Terms and Right to Opt Out. They’d like to move this particular card from Prime + 4.99% to Prime + 8.99%.

The interesting thing to me is that we pay down our cards regularly and substantively, but (entrepreneurship being what it is these days) do maintain a balance. So I’m going to assume, based on the little I know about financial actuarial practice and business analytics, they’re reaching into their customer base and targeting those who they gauge willing to pay the increase and unwilling to bail out of the program entirely. The suckers, in other words.

I suspect, though I have no evidence on hand, that they may even be ameliorating rates for those in serious financial difficulties, since defaulting on credit cards (no matter what Congress does to change the bankruptcy laws) doesn’t give CitiFail any more money when you get right down to pocketbook accounting, or operating capital. And I bet they’re intentionally pissing off, or dropping outright, people (like my Mom) who are long-standing customers who pay down their balances immediately.

Intellectually, I don’t envy them. The times, and the crisis, are setting them against us. And if they make any wrong moves in the course of their collapse, they will not merely fail financially but end up being rendered in cartoons as big fat men with tophats and pocket watches. The only thing they can do (from their standpoint) is try to get what they can, trim the fat, and harvest the rest.

I find myself wondering if maybe, just possibly, there is an opportunity here to… well, the phrase coming to mind is “reverse the predator–prey relationship”. That’s pretty optimistic. “Level the playing field a bit” might be the more measured phrase. “Enable a collective defense” might be even better.

A transparency play.

I know that companies like Wesabe already aggregate (but keep private) credit account information. Not just how much money people owe, but I assume also information on Terms of Service and interest rates and suchlike.

What I’m imagining this afternoon is an anonymized but public aggregation of “everybody’s” terms of service, with a deep-ranging analytic system wrapped around it. How many people in your county have got better Terms? How many card contract structures are there, actually? What can we infer about Card Company X’s actuarial ruleset from the data on 100000 people’s contracts? What can those people do, looking at one another’s contracts, to identify opportunities to improve the “product” they are offered?

For example, when I called Inisha at Citi just now to bitch about these changes, she was (a) unable to tell me my Card Membership Date, (b) able to tell me that my Terms of Service cannot be seen online at all, anywhere, (c) able to put me on hold and propose a compromise interest rate somewhere closer to my original, without obviously working up a sweat. The implication to me is that this amounts to intentional obfuscation and hiding of crucial information, coupled with a big set of simple contingent reactions to protest. I have to assume that Citi keeps it hard to find anything out about the complex state of one’s terms specifically so that their representatives can offer a pig-in-a-poke incomparable alternative when challenged.

So I’d copy over my contract information somewhere, and even be willing to reveal limited balance data, if I was confident of its anonymity and knew that it was useful to reveal patterns in Citi’s (and others’) business practices and sales offer strategy.

I bet some other folks might be willing to do that too.

links for 2008-11-20

Older entries »