disintermediation part 3: universal spiff

Retail stores are struggling because it's cheaper and better to buy things on the web. I see it in the store, but I wonder if what I'm seeing is highly rated, and I know I can get it cheaper online. Yet manufacturers still need to get goods into customers' hands to encourage them to buy them. Some companies have tried manufacturer stores, but they are hampered by not undercutting other stores (my local Sony Style store closed a year ago).

I worked in tech support at a company, and if someone in support was able to convince a customer that they needed a product (often a paid upgrade to the new version that addressed the problem) , the employee would transfer the phone call to sales and get a small commission on the sale. This is called a "spiff": A bonus or other remuneration, given for ... promoting the goods of a particular manufacturer.. The problem is if the customer later on bought the product because of that person's assistance, there was no way the employee could get compensated. I've also gone out on sales calls where my technical knowledge sealed the deal, yet the sales person got the commission.

My thesis is companies have a certain promotional budget for the sale of an item, and they should be willing to hand it over on ANY sale. They already do this on click-through ads for web sites, but they should extend this to the world of people.

It would work somewhat like this:
Enter the e-mail address of the person who led you to buy this product: _____

So whether someone demonstrated the laptop at the Sony lifestyle store, or someone showed it to you at a convention, or you saw a book at a bookstore but guiltily purchased it online instead, or your mechanic told you to buy some accessory, or you read a rave review about something on a blog, you can credit the person who led to your purchase.

This would encourage and liberate millions of people to sell things, and it would break down the artificial division between salespeople selling a product and the many people whose efforts can lead to a sale. It would provide a financial model for company "stores" and marketing events that just demonstrate items yet often spur sales.

Companies wouldn't allow any e-mail address, they would have a known list of approved promoters: all their employees, partners, people on certain web sites. The system has vast potential for "abuse", where someone who had nothing to do with the sale would offer to split the spiff with the buyer. But these all lead to increased sales!

disintermediation part 2: Amazon should be/buy UPS

I used to compare items for price using MySimon and Froogle, but now I just go to Amazon and pay whatever. However, there are still companies competing on price. You can buy LCD TVs from obscure companies in Nebraska that sound like they are operating out of someone's garage, some dot com sites are still trying to preserve the buying circles (get 15 friends to buy the same product and it's 5% less), and of course there are various auction methods.

It's all a waste. A third party coming between the manufacture and me can do no more than jack up the price. My thesis is simple: the lowest possible price of an item is the price at which the manufacture is willing to put it on the loading dock, plus shipping to your door. Everything else is overhead.

You may respond "But shipping 50,000 widgets to Walmart is a lot cheaper than shipping widgets to 50,000 addresses." Sure, but Walmart has its overhead. They have to uncreate 50,000 items and put them on the shelves of big stores, then wait for them to sell.

No manufacturer wants the hassle of shipping to 50,000 customers. But manufacturers don't really want the hassle of shipping even to 100 stores either. The whole process of getting goods to you is called logistics:

Logistics involves the integration of information, transportation, inventory, warehousing, material-handling, and packaging, and occasionally security.

Many manufacturers already outsource this. What's stupid is the manufacturer pays for one logistics chain that gets widgets from it to the store, then the store pays for another chain that gets the widgets to customers' houses.

You could argue that manufacturers need stores to aggregate individual purchases into bulk orders; they can't run their assembly line in response to individual orders dribbling in, but they can intelligently respond to a 50,000 item order from a store. But bulk orders just hide inefficiency and mistakes. Some of a big order winds up on sale or dumped in discounters and outlet stores, which jacks up costs for the store. You could claim that stores should know their customers' tastes better than the manufacturer, e.g. the Super Bowl is coming up but economic conditions are poor so cheap 720p TVs will do well while 1080p will languish; but that's part of the problem: manufacturers need this information as much as stores do! A manufacturer whose customers are big stores knows less than a manufacturer whose customers are the actual users.

For years I've believed the obvious answer is for Amazon (market cap $ 52 bn, revenue $22 bn) to buy or become UPS (market cap $58 bn, revenue $ 45 bn). This results in one chain from the loading dock to the customer! Returning to my thesis, the manufacturer tells UPS/Amazon what price it's willing to put it on the loading dock, and Amazon/UPS puts it in the customer's hands. The manufacturer and Amazon/UPS can work out whether to manufacture 50,000 at once or do smaller runs, it all ends up in one supply chain. And Amazon/UPS has all the customer knowledge and programming smarts to offer sales approaches that reduce prices like bulk-buying circles, limited-time promotions, auctions, reverse auctions, overstock, etc. but in conjunction with the manufacturer.

Anyone know Jeff Bezos' e-mail address?

disintermediation part 1: ordering direct

I have a backpack from the Gap. It is utter garbage that's falling apart. So what exactly is the Gap adding when it has a Chinese factory sew its label on their part 53AF47Y?

Just let me order the best product direct from its manufacturer.

This applies whether I want a custom one-off or just their generic part. As I wrote responding to a Gizmodo post about custom manufacturing:

I want the best worker in the factory cranking out quality $10 clothes to make me custom stuff after hours for five times the price (which will still be less than what some US store charges me for their worthless label). That just requires some bilingual entrepreneur to set up a web storefront to take my order and hook me up direct with the manufacturer.

This is a huge potential market. It cuts out the non-existent added value of some dumbass US brand supposedly getting me better quality goods while in reality they're just twisting their suppliers' arms to reduce costs.

Sadly for the USA, this innovation is more likely to come from a Chinese bilingual near the factory rather than someone living in America.

Meanwhile every American brand that only spends on marketing while ordering up generic crap (clothes, luggage, etc.) from overseas factories DESERVES TO DIE. The actual factory can be the brand, and can offer custom work through these entrepreneurs. The moment Happy Dong Manufacturing gets a great rating from Consumer Reports and you can order its widget on eBay, a lot of USA specialty stores will really suffer.

non-support: Suddenlink can't help itself

My cable bill appears on my bank's billpay site, but when I click "View Bill" nothing displays. My bank can't fix it. So first I tried Suddenlink's "Chat now":

S Page: my problem with viewing my Suddenlink statement through my bank's online bill payment system. Can you assist with that?
Dana: Unfortunately, no. We provide the statements to your Suddenlink.net account. ... Viewing it via your bank's online bill pay, could be an issue with their system.
S Page: No, wsc.suddenlink.net is definitely doing something wrong. My bank's billpay site shows all my other electronic bills fine. It presents the amount and date of my Suddenlink bill fine, but when I try to view the Suddenlink bill, a) it only works if I reduce my browser security settings and b) even when I do that I get "JSPG0036E: Failed to find resource /WEB-INF/jsp/lang/en/atl_cferror.jsp".
S Page: This is a highly technical flaw with https://wsc.suddenlink.net/EUR_ViewBill/Controller/ProcessCheckFreeAuthorisation?data=...
Dana: That could still be an issue with their site. Have you been able to view your bills at that site before?
S Page: As I said, *every other electronic bill* I get works fine. Problem a) has been around for over a year, but the error message b) is new.
S Page: Do you have a bug reporting system? I want you to enter in it "Customer reports two problems with wsc.suddenlink.net's presentation of online billing to another billpay system. ..." I would be happy to provide more details but there's no point if you don't have a method to report problems with this system.
Dana: We do not have a method to report this that I am aware of. I can report it in your account.
Dana: Everything that I am finding on that error appears to be software issues so far.
S Page: Well that's completely lame. I'm trying to help Suddenlink fix a problem with a service it provides! Suddenlink must have a director of web software engineering who needs to know that your electronic bill presentation system isn't working. Yes it's a software problem.
Dana: (no response)

I tried again with Melanie on the phone. She also had no means to report that their bill-presenting software is broken, the best she could do is e-mail her supervisor.

It's the same sad pattern as Earthlink support and Symantec support. They can all help a customer with certain classes of problems, but the company is structured to be incapable of letting a customer help them.

The future of the web is supposed to be autonomous bits of software talking to each other on behalf of customers, but dinosaur companies aren't set up to support the interaction.

skiing: mogul love 18 ways

O mogul field, how do I love thee? Let me count the ways:

classic short radius turn

Approach mogul, plant pole on top, as you crest the mogul the tips and tails of your skis clear the snow, so you can turn both feet simultaneously to swing the skis around in a short-radius turn and slide down the backside of the mogul towards the trough before the next bump. This is how most skiers attempt moguls, but there's more to life.

pivot slip

Turn your feet in-place on the crest so that your skis don't follow the arc of a short turn, instead they pivot within their length and you transition into a pure sideslip. Disadvantage: hard to control.

pop off pillow below

Don't turn so much, so your skis head downhill more than slide sideways. Guide your skis towards the soft snow in front of the next bump while keeping your body low. They'll hit that soft snow, your upper body will pop up and the unweighting will turn your skis the other way. Disadvantage: can be herky-jerky

leaper edge change

Ski diagonally to the bump, let it launch you, in mid-air tilt your shins and ankles to change from your uphill edges to your downhill edges, and land in the trough before the next bump. For bonus points, land past the next bump. Works better in shallower bumps. Disadvantages:needs room, takes guts.

GS turn across two/three/four bumps

Ski over the bump, use the unweighting to effect the edge change, extend your legs to the side but don't twist too much, and make a huge turn across the mogul field as you simply absorb subsequent bumps. Have faith, the power of your turn overwhelms the forces from the bumps. Disadvantage: how strong are you?

two tight turns down the spine

Make a tight turn or a pivot turn at the crest of the bump but rather than slide down its back, make a second turn again on the convex part. If you imagine your skis resting on a sphere, your tips and tails are always clear and ready for another turn. Works best on longer "beached whale" bumps. Disadvantage: it's harder to turn on a downward slope.

hop turn

Instead of a scrabble slide down the back of the bump, just do the entire turn in mid-air. This avoids any worry about your skis clearing the bumps nearby. Pick a suitable landing spot: on the backside, in the trough, in the pillow of soft snow before the next bump, or even on the next bump. Disadvantages: Could be a big drop, your back might hate you.

carve the water line

Don't slide down the backside of the bump, don't slide into the next bump, don't even pivot on the crest. "Simply" follow the 'S' shaped path between the bumps that water would follow down the hill. Retract your legs as the path flattens out, extend them to the side after the path changes direction and curves. Disadvantages: In two turns you'll be going ridiculously fast, sometimes there's no room for a carved turn.

aggressive turn finish uphill

Most people trying to link turns in moguls stop turning when their skis face across the hill, which doesn't reduce speed. As they improve they try to stop turning when their skis point to the start of their next turn, which usually picks up speed. To control your speed, keep turning your skis (whether skidded or carved) until they point uphill; you'll need to practice twisting your knees at the end of turns. (If turning is good, more turning is better!) You can either steer to a bump slightly uphill or skid backwards into the next turn. Keep your upper body facing downhill and the extreme counter-rotation of your upper and lower body builds up big forces; the moment you relax your skis will swing into the next turn. Disadvantages: none.

avalement avec le deep knee bend

"avaler" is French for "to swallow". As you ski to a huge bump, instead of popping off it or launching, crouch low to absorb it. Crouch ridiculously low to look French; when you plant your downhill pole on top of the bump your hand should be above your shoulder! Disadvantage: it's harder to pivot when you're that low.

le jet turn

Instead of pivoting as you crest the bump, shoot your skis forward and turn the tips in front of you. After a stylish drop into the trough, bring your upper body back over your skis Disadvantage: unless you're French, you'll look dopey.

backpedal the feet up the face

Retracting your legs as you ski up the face of a bump picks up speed. It's possible to resist the upward pressure of the rising face to slow down slightly. One way I've heard it described is you move both feet as if backpedaling from bottom to top on a bicycle. As you approach the face of the bump, push your feet forward and let the bump push them upward. Disadvantage: you have very little time to get this right.

push the dolphins back under the water

You can think of this as the reverse of back-pedaling up the face, you're moving both feet to pedal forward and down. But riding dolphins SeaWorld-style is more poetic. You don't want bumps to throw you (unless you want to do leapers or hop turns). After you reach the crest of a bump, you need to aggressively push the tips of your skis down and into the new turn to avoid an unintentional launch. Especially in bumps in new snow, you want to feel as if you're driving your skis back under the water — as if you're astride two dolphins that crest out of the water and then you drive them back under. Disadvantages: none.

punch the gearlever into third

The pole plant on the crest of a bump helps timing and gives you something to twist around. But as you move downhill you don't want that pole plant to hold you back and you definitely don't want it to pull your shoulder back so that your upper body faces across the hill. So reach downhill and plant the pole, but immediately push your hand further downhill in an aggressive move. It should feel as if you're rapidly shifting a car's gear lever from second to third. Watch freestyle skiers, their hands always return to downhill in front of them. Disadvantages: none.

one-and-two independent leg ("Hey, it's slightly better than a stem christie")

You want slightly more weight on the outside leg in any ski turn, and in moguls the outside leg has more room. If you're in a slight snowplough the outside leg is pointing in the direction of the new turn but your inside leg is pointing the wrong way and has less room. and winds up perched on the bump. So you start the turn with the outside leg and simply lift up the inside leg and put it down next to the outside leg. It gets you down the gnarliest bumps. To help avoid this, before you start turning bring your skis closer together and crouch lower. Disadvantages: less fluid, less control at the start of the turn.

skidded parallel turn to out-of-control traverse (oops)

Skidding a turn, leaving your pole behind, and not turning your skis all lead to you facing across the hill and bouncing into the steep sides of nearby bumps. Disadvantages: not stylish

sideslip with optional caught edge and shoulder plant (ouch)

Nearly all mogul turns involve some skidding, so skip the turn part and just do the skid. Act like a novice snowboarder and sideslip over and down the bumps fast and smoothly. Imagine that you're a 70s skier on 7 foot planks and it'll be another 3 seasons before you can make parallel turns. Disadvantages: by definition in bumps the snow isn't flat, so there may be NO correct edge angle for effortless sideslipping.

And if all else fails, straightline short sections, pumping your legs up and down like Jonny Mosely to absorb the bumps.

Disadvantages: can your knees take the abuse?

Armed with these techniques you can approach a mogul field as artistic goal-driven surgery. Here are some operations:

• Descend as slowly as possible yet smoothly, by turning uphill and steering uphill to bumps.
• Le French totale, all deliberate lower leg actions and huge knee bends.
• Stickwork, punching the gearlevers and planting poles to set your downhill progress.
• Maximum absorption, using retraction and extension over and down bumps to increase or reduce speed (I suck at this, in a skateboard halfpipe after pumping the deck five times I wind up motionless at the bottom)
• Minimal effort, just directing and pivoting the skis in advance so that as you reach each bump it turns your skis for you.
  
• Stay off the snow, doing leaper and hop turns off every bump instead of absorption
  
• Two monster GS turns to cover the whole thing.
  
• Keep all your turns in a narrow corridor regardless of conditions.

art: Rothko coolly torches SFMOMA

SFMOMA's Rothko (No. 14, 1960) quietly overpowers everything else in the museum. It's not the Titanic, it's the iceberg that sank the Titanic. It's suffused with pure emotion, but it's a receding internal burn. The red is alive but not fiery; the mouth (or dot of the exclamation point?) underneath is a deep inky blue but it's not heavy, you can see through it to the background of quiet earth. The brushwork is sensationally good. Ten feet tall, it's bigger than you but it's taking emotional states available to everyone and exalting them to a mystical, noble level. Duchamp invented so much, Picasso and Matisse were big stars, but in 2200 I'm confident people will rank Rothko higher than any other artist of the 20th century.


Rothko painted a similar design in 1961's Number 207, it's brighter on darker and blacker. The canvas is burning up:

computers: the heck of Linux audio, the hell of documenting it

For no apparent reason sound stopped playing in certain applications on my computer.

I'm running the Kubuntu flavor (incorporating the KDE Plasma Desktop) of the Ubuntu distribution of Linux, which makes certain decisions about how sound should be configured. But I've also installed latest Firefox and Thunderbird and the Flash plug-in and the VLC media player; these make their own decisions about how to play sounds. Kubuntu appears to provide an audio "software stack" consisting of the following layers: hardware audio drivers, the ALSA low-level Linux interface to audio, the PulseAudio server to mix audio streams and talk to different hardware, and KDE's Phonon audio abstraction. That's complicated enough, but great minds can and do disagree if this is the best approach to audio; some people think audio should be simpler, some think it should support playing across the network to your Bluetooth headset when you walk into the garden and automatically switch to 5.1 surround sound if you login to your friend's home PC. Those disagreements result in multiple software packages and approaches to audio. So somehow some piece of software additionally installed the GStreamer multimedia framework and the esound daemon on my computer. And each of these layers can be configured to talk to, bypass, or emulate the other layers, so maybe I have an emulation of OSS or JACK or some other software approach. Each program tries to detect what's available and use the right approach, but each program and layer can be configured to use different layers. Who knows where the problem is? Who knows what the right solution is?

Each one of these programs and packages has sound-related documentation. But each doesn't know how my system is configured, so their instructions are inadequate, incomplete, or wrong.

So I search for "ubuntu sound problems". One of the results is "Comprehensive Sound Problem Solutions Guide", even though it was written in 2006! And it has 1,680 follow-up comments! Another search result confidently instructs you to remove the PulseAudio sound engine altogether, even though it's what Kubuntu uses. Another tells you how to compile source programs to add sound support, even though Kubuntu came with dozens of sound packages and other programs have installed many more. Most instructions tell you to make unexplained undocumented changes to configuration files, even though there are configuration dialog boxes in Kubuntu and it attempts to configure itself correctly without these files.

The documentation is spread everywhere; Google rewards the oldest most out-of-date information; reasonable people don't agree on how it should work. I'm probably screwed. If and when I figure out how to fix my problem I would love to improve the "Linux sound documentation", but in what form (bug report, forum post, wiki page, web page, built-in help) for which component or layer or distribution? Agggggghh.

art: the lonnnngg view, the evolved sense

Judtih Thurman's New Yorker piece on cave art is inspiring. What were our ancestors thinking 32,000 years ago when they drew animals on the walls by torchlight? Their art affects us so profoundly; does that mean they are close kin to us that we can inhabit in our imaginations, or does art transcend species and culture so we're safer admitting those strangers in the dark are incomprehensible? What does it say about art that five times more than recorded history later, no artist draws animals better?

Meanwhile, Denis Dutton wrote a New Yorker op-ed piece dubious about conceptual art, invented by grand old man Duchamp a mere 90 years ago. He mentions the 32,000 year history of representational drawing, and 100,000 years of decorative shell necklaces, but they're passing fads: he points out our ancestors showed off for 1,200,000 years by crafting attractive Acheulian hand axes. "Ooh, nice symmetrical chipping surfaces, let's mate."

It's interesting, but fails to make a case. So appreciation of skill has been around for 1.4 million years and we've only refined our aesthetic sensibility recently? That is no evidence that we'll regress to a more conservative primitive "if it ain't skilled, it's crap!" form of art appreciation. He says

Jokes about Cadillac tailfins and early fax machines were once amusing, and the same can be said of conceptual works like Piero Manzoni’s 1962 declaration that Earth was his art work, Joseph Kosuth’s 1965 “One and Three Chairs” (a chair, a photo of the chair and a definition of “chair”) or Mr. Hirst’s medicine cabinets. Future generations, no longer engaged by our art “concepts” and unable to divine any special skill or emotional expression in the work, may lose interest in it as a medium for financial speculation and relegate it to the realm of historical curiosity.

But I've never heard of those 1960s pieces —I'm already the future generation—and they sound beguiling and intriguing. One and Three Chairs is excellent! Meanwhile mere skill in the Renaissance religious painting (or in sexy hand axe tip chipping) leaves me lukewarm.

Back to cave paintings, they've inspired other fine art: Here's William Gibson speaking on the cave paintings as the precursor to movies, and Steely Dan's The Caves of Altamira lyrics.

skiing: Shane McConkey

Tonight's CBS 60 Minutes had a great segment on wingsuit jumps off mountains (watch it). At 8:15 JT Holmes reminisces about his friend, skiing icon Shane McConkey. Shane died last year when he couldn't release his skis in time — he was combining skiing and wingsuit flying just as he earlier combined skiing and BASE jumping. Fly like a bird, about 1 in 400 risk of dying doing what you love.

The only ski movie I own is
there's something about McConkey:

From ridiculously exposed mountain goat descents to Alaskan peak straightlines, from full-twisting switch backflips in the terrain park to double backflips off cliffs, from building base jumps to swinging on helicopter skids, from rock skiing to stream skiing to the inebriated shenanigans of Saucerboy, ...
The film focuses on the abilities, attitudes and antics that have made McConkey a hero for a generation

Indeed when he died, messages boards were split between "He died doing what he loves, he's my hero" and "He had a wife and kid, unlike a fireman he risked his life for nothing, he's no hero", which shows “hero” has both the Greek ideal for-the-greater-good and an ironic worshipping-false-gods meaning.

After that movie he tried skiing on waterskis which led to his invention the first reverse camber ski, the Volant Spatula, followed by the K2 Pontoon and now a whole category of reverse camber and rocker skis.

cars: death of American brands

I'm surprised how my friends didn't care about General Motors going under. Making things is important to a country's soul, and the alternative of financial shenanigans didn't work out well.

GM has survived, but at the cost of so many storied nameplates. It's hard to keep track of the carnage, here's where things stand, with a little help from Wikipedia.

Hummer: Its sale to Sichuan Tengzhong Heavy Industrial Machinery Company blocked by Chinese government on "environmental grounds" in June, but in October it finally went through for $150 million (a baseball stadium costs three times as much). GM has no plans to continue the nameplate after the 2010 model year.

Opel/Vauxhall: Sold! To consortium lead by Magna Group-backed-by-Sberbank of Russia (55%). GM will continue to own 35% of Opel; while Opel employees will own 10%. Called "New GM Europe" by some.

Pontiac: dropped, all of its remaining models will be phased out by the end of 2010.

Saab: A deal to sell it to tiny Swedish supercar maker Koenigseggeggegegg (sp?) is supposedly still on, though it depends on three billion Swedish Kronor appearing from the tooth fairy troll.

Saturn: Dead now.

GM holds on to Chevrolet, Buick (mostly for China), Cadillac, and GMC trucks and SUVs.

Damn that's depressing. It mirrors the fall of British Leyland.