March 7th, 2013

Don't predict the weather: predict the people

My husband R. is amazing at reading the weather maps at weather.gov. He reads the forums where they discuss the multiple forecasts that are produced before being culled down to something that is then released to the general public, which in turn goes into the giant hopper that is weather.com, which Nate Silver says has a "wet" bias -- they predict more precip than happens.

When it comes to predicting what people and schools will do in response to those forecasts, however, I'm still trying to figure out whether to just ignore what R. says, or to invert it. I think the former is the correct decision. And he's not the only one -- the local radio hosts (and I mean the centrist ones) are consistently in agreement with him on things that seem fantastically insane to me (most recent gem: by Saturday afternoon following the storm named "Nemo" by TWC, it'll all be plowed back and business as normal. In practice, the travel ban still wasn't lifted by then).

I don't pay any attention to the forecast; I look at what comes home in my kids' backpacks and I pay attention to the chatter at the school. If the people working at the school all believe school will open late, close early or not open at all, they're right. Always. And I'm starting to find out why. Here are my theories for school schedule changes regarded as unnecessary/overkill by my husband and/or local radio hosts:

(1) Employees of the school often live in surrounding areas that will have more direct impact than our area. They may (be forced to) cancel (arrive late, leave early) due to children in those areas schools.

(2) Parents of children in our school district who have signed their kids up for xday often have long commutes at inflexible jobs. They may be unable to pick up their kids due to traffic. The primary effect here is to convert what might have been an early release day to a cancel.

(3) Traffic grinds to a complete halt on 2 regularly anyway. And that can force what might have been a normal day to become an early release to guarantee that all the kids get home before dark.

I'm from Seattle. I know you don't go anywhere near a road (not as a ped, much less in a vehicle) when the slush hits, because some Idiot will Hit You, so you avoid it if at all possible. Also, waiting in the slush for buses sucks. I've always assumed that people in SnowLand, which includes around here, cope really well with the white stuff. Well, the cold hard truth is that a lot of people in SnowLand are, like me, Not From Around Here, so the native coping ability is not what it could be. And the natives who can cope with the fact that the building in that location is two buildings after the one that other people use to give directions by have long since figured out that just because _they_ think _they_ could get home in a timely fashion in the middle of a blizzard doesn't mean anyone else will actually show up to work on that day.

Also, there are a few Clever people out here that know it's not worth it to even try, and that a lot of people will like you if you give them a solid excuse not to either.

Baby steps.

AccelaStudy Dutch Essentials (free version on iPad) review

I have not upgraded and it should become quickly apparent why.

The free version seems to have 132 words in it; the premium has on the order of a couple thousand, if I understood the descriptive material correctly.

In the flashcard portion of the app, the English speaker is a male native speaker with a relatively neutral American accent (there may well be a way to customize it, I don't know). However, the Dutch speaker sounds synthetic, with a frequency range that suggests adult male -- but it sounds synthetic, like a program not a recording of a person. For some of the words, the results are compelling (mistig, veertig). For some of the words, the results are less good (oranje) and for at least one, I think it's actually wrong (scharlaken). The result on groen actually sounds like the person whose speech was sliced and diced for the rest of it actually _said_ that word for real.

Scharlaken is particularly weird, because the sch sounds right in verscheidene, but it is pronounced "sh" here. I repeated it a ton of times and I doubled checked the sound clip on nl.wiktionary.org -- it is supposed to be the usual sch and it just isn't.

So you cannot trust the pronunciation here: it sounds synthetic AND it is sometimes wrong in an important way. Can you use it just to learn the words visually/spelling? I did not detect any spelling errors, however, whoever put this together screwed up badly in the word mapping. You could argue about whether you should translate lightning as bliksem (which seems pretty straightforward) vs. bliksemflits (their choice). Ditto the way they translate donder. Whatever. But they translated pacifier (this is in a list of 132 words with _nothing_ about keeping the peace and a few other baby related words like diaper) as vredestichter, rather than fopspeen.

Synthetic Dutch speaker, pronunciation error(s), wildly incorrect translations = I'm going to delete the free app. Just fixing the individual errors in pronunciation and translation quality is not going to be enough to convince me to retry this product, because I'm _learning_ this language. I should not be able to detect these kinds of problems in the basic version of the product.

Babbel Dutch, on a laptop through the web and through iPad app (free)

Babbel presents a much trickier decision in terms of should I buy the premium. In the case of Babbel, the premium is a subscription service, and, alas, the only Dutch course available through Babbel is the beginner course.

Babbel is structured a bit like Rosetta Stone, in that it is heavy on pictures, and in that it offers a voice recognition feature. It differs from RS in that it relies upon word/phrase equivalency with "native" language (I used English, I believe you can set your native language but I haven't tried it). There is a nice little chunk you can try before you buy. The pictures are clear. They used what sound like native speakers. The words I worked through were at least as plausible as RS's. The dictation/spelling portion is more incremental than RS, in that one part of it just asks you to pick the correct first letter and another section offers you tiles to spell the word/phrase with. Those are both valid pedagogical strategies to (demonstrate that you are) learn(ing) how to read/write/spell a language.

The voice recognition feature suffers from approximately the same set of issues as RS, altho they handle the first syllable/communicate when you can start speaking marginally better.

Babbel is potentially highly disruptive of the RS revenue stream altho it suffers from offering fewer languages/only lower level courses in many of those languages.

If I had not already bought RS, and was looking for an online method for learning/reviewing Dutch or another language, I would start with Babbel. Because you get unlimited access for a period of time, if you work for many hours diligently, you could save a ton of money over RS.

If I decide to learn another language after this summer, I will be looking even more carefully at where Babbel is at that point.

BYKI Dutch on iOS (not a native iPad app)

Before You Know It is a product of Transparent Languages, which I keep looking at and then not actually trying (I think this is because they are not a great Mac option).

BYKI is a flashcard study/quiz app. As near as I can tell, once you buy the (premium?) version, you get access to all the lists/card sets (I don't think they charge for the download more section, but maybe they do and I just don't realize because I haven't hit the free limit yet). You can decide whether you want to "study", use the cards in one direction or the other or go to the quiz, which is nice -- you can just go straight to the quiz without sitting through the "learning" portion. They use a native speaker. Some of their translations are a little iffy (Dank u wel is "Thank you"), but nothing like AccelaStudy.

If you want a flashcard/phrasebook study tool, BYKI is excellent. I've picked up a few words through it, but not enough to commit me to spending much time with it, because there are other ways for me to build vocab faster.

I think I got BYKI as a downloadable application for my laptop some years ago and had a roughly equivalent impression of it.

ETA: I'm going to retract some of the positive review because they screwed up the translation for sour cream (giving it as creme fraiche and no I'm not going to do the punctuation on that correctly sorry). That is incorrect (I mean duh), see here for detailed, correct explanation of the various kinds of cream ("room"):

http://kook.rubriek.nl/eten&drinken/room.php

I'm sensitive to this issue for two reasons. I'm allergic to milk products in general, and it can be very difficult to figure out how to cook in a new environment when translating the basic underlying foods is hard. I helped my cousin M. (from Mexico) figure this out (not for a Dutch word -- this was from Spanish) when we were hanging out with my Mennonite friends. It took a while, even tho M.'s English is fantastic, because we ultimately needed someone who had milked cows to figure out which kind of cream she remembered from her childhood.

New Strategy for Learning Dutch Vocab: Buy kids learning apps

Like all kids' apps, there is the slight problem of not knowing how much you are signing up for in terms of in-app to get all the pieces, but the good news is, there's no question about whether these are words that Dutch people use (they teach them to their kids!) or that the speakers are natives.

Expect this to be updated. A. has stolen the ipad from me to play with the app, which is a good sign.

The word for squirrel in Nederlands is eekhorn. Heh -- pronounced like "acorn".

ETA: It won't let me buy the premium stuff! I get "canceled"! This may be region enforcement, but serious bummer. I don't have restrictions enabled on this iPad so it isn't that.

The "sour grapes" in me sez it's just as well. Their cartoon of a "zeekoe" showed it getting out onto something white and ice-like. I looked up "zeekoe" on Dutch wikipedia. Zeekoeien (is the right plural -- I had the wrong one but B. kindly corrected it for me) don't do ice.

Clayton Christensen appearing on Bloomberg reminds me that he exists and annoys me

The biggest problem with Christensen is that some of his ideas are really interesting, useful, important and unusual. This makes him, if not sui generis in b-school types, at least a rara avis. So when I am reminded that he exists, I feel like I ought to hear him out, which then annoys me, partly because he's Mormon, partly because he has the most amazingly blinkered perspective on the world, and partly because he keeps buying those interesting, useful, important and unusual ideas under all the rest of that shit.

He apparently had a stroke, lost a bunch of language and regained it with RS (but that is _not_ how I stumbled across this: I really was watching Bloomberg and it was the least productive interview I've ever seen on that channel. No exceptions). In an effort to figure out what he's really on about these days, I found this piece in Wired from last month:

http://www.wired.com/business/2013/02/mf-clayton-christensen-wants-to-transform-capitalism/all/

Christensen is correct in noting that there's a lot of capital that is idle and/or not being deployed in longer-term efforts. He does not mention in this piece, vs. in the Bloomberg interview, his belief that disruptive innovations create big change that then result in more jobs -- an aspect of his whole approach to thinking about innovation that I would find funny except _he really fucking believes it_. He is _that_ kind of nutball (duh, he's Mormon and thinks he was raised to ask questions, as opposed to understanding that he was raised to be comfortable with multiple paradoxical and ludicrous beliefs that many people around him do not hold -- which is actually tremendously useful if you're going to disrupt an academic field).

I imagine you could find a disruptive innovation that initially created more jobs than it destroyed, but they are pretty rare. Generally speaking, the "additional" jobs are the duplicative ones (that is, the old way of doing things takes a while to go away and its jobs with it), altho whenever a disruptive innovation creates entirely new economic activity (read: consumption? somewhere, anyway) it can create jobs (altho trying to calculate the net effect of horses vs. cars on jobs, say, is a really non-trivial exercise, never mind calculating how many jobs in digital photographic processing were created to compensate for the lost of all those while-u-wait photo development kiosks).

He thinks that "journalism ... and publishing broadly. Anything supported by advertising. That all of this is being disrupted is now beyond question. And then I think education is just on the edge of the crevasse. Generally, universities are doing very well financially, so they don't feel from the data that their world is going to collapse. But I think even five years from now these enterprises are going to be in real trouble."

I see one item to disagree with in this assessment: I think five years sounds like a real short time frame -- but I could be wrong, so I'm going to try to remember to check this one (and likely fail -- I'm still struggling with tickler reminders more than a year out).

I think that he formulated what I think of as the p-disruption (p for physical but often paper) as journalism/publishing/supported by advertising says a lot about how he sees (and fails to see) the world. And I honestly cannot believe he didn't say a damn thing about the massive disruption in retail over the last decade-ish. Does he think it somehow is over?

Also, poor IT departments. That's some _serious_ disruption.

The Bloomberg interview was frustrating because the interviewer brought up Amazon, but said, but you have to make money! And Christensen just emitted his placeholder "You're exactly right" and went back to something else, rather than pointing out that complaining about "failing to make money" by choosing to re-invest it in the business is _exactly_ the kind of capitalist behavior he thinks the tax code should be modified to incentivize.

I'm going to quit thinking about this. It's too irritating.

The Mall of the Future/Today

Photos of Amazon's Scottish facility:

http://www.flickr.com/photos/scottishgovernment/6352123585/
http://www.flickr.com/photos/scottishgovernment/6352123451/

h/t http://www.wired.com/business/2013/03/yuppies-same-day-delivery/

There's been a bunch of rumors about Google and/or Amazon setting up same day delivery in some urban areas, particularly since sales tax collection online is now more or less a done deal so the incentive to locate in places to avoid sales tax collection responsibilities is reduced.

Somebody did a survey and found out only a small number of people would pay for this service.

"The outliers in the results were what the pollsters called “affluent millennials” — urban dwellers ages 18 to 34 with an annual household income of at least $150,000. Members of this demographic said they’d be willing to pay up to $10 for same-day delivery, while others said they’d only pay up to $6."

But earlier in the article, the author says:

"Rather, outside of a few self-important yuppies, no one else seems to care that much about getting their orders that quickly."

All very stupid. Same day will only happen in certain big cities for logistical reasons, not because "no one cares". You need a whole lot of people _close_ to your facility to pay that extra for same day to justify the network to make it happen. Even if _everyone_ in Wyoming wanted to pay double for the same, it wouldn't make any sense to try.

There are things that could be said about certain people are more likely to order if they could get same day, or that you could get a whole different category of orders if you did same day, but why?

Big ups for pointing to the picture, tho -- those are great photos.

BCG on the subject of same day delivery (with addenda on groceries)

This is _so_ bad it's _almost_ funny.

http://www.bcg.com/media/PressReleaseDetails.aspx?id=tcm:12-129157

I'm going to pick my favorite sentence where "favorite" means "most easily proven to be unbelievably wrong".

"Economics dictate that retailers should offer same-day delivery for only a select number of products that are small and light and that carry high margins. Electronics, office supplies, and apparel are likely candidates."

I live in the Boston area. My Favorite Grocery Store, the one I shop at all the time and where the employees ask about the kids when they aren't with me, offers Grocery Delivery. They charge $9.95 per delivery, and they don't mark up the groceries separately (right down to you get the promos current when they do the shop). While you can't actually get same day today, partly that's because you have to pick your delivery window at least 4 hours in advance, and some of the windows get fully booked out. But you can pick the earliest window tomorrow in the Natick store's delivery area as I'm writing this, which is less than 24 hours from now.

$10.

For groceries delivered to your house. At the markup I pay when I shop in the store, and I shop at _this_ grocery store (not the Natick location, but you know what I mean -- Acton is booked out till Saturday, so there's no guarantee).

What people who try to do this analysis fail to take into consideration is the cost tradeoff between having to house customers and their vehicles while they do the picking (for "free") and then have to clean up after they decide not to buy that ice cream after all and leave it in the cereal aisle and/or after their small child vomits in the bakery department VS. the cost of paying employees to pick, supply refrigerated vans to do the delivery, and paying employees to drive the vans (which the customers did for themselves for free in the previous scenario). In dense areas, the cost of supplying parking adds up, and the delivery area shrinks, so grocery delivery makes more sense in dense areas _where families still live_. Families with two incomes AND children (the gratuitously slammed affluent millenials) will pay for not having to get to the store while it is open to buy the thing their kid needs really very soon if not right now.

But hey, keep being stupid about this and I'll just laugh while the companies that understand this make money off of your surprise. Good news here: this is a disruptive innovation that actually _creates jobs_, which we need more of, right?

ETA: I should note that this particular topic tends to bring out the stupid; BCG is acting like everyone else in this category.

http://www.investopedia.com/financial-edge/0711/10-reasons-why-online-grocery-shopping-is-failing.aspx

Best sentence:

"In Europe, shoppers typically purchase ingredients for their meals the day they prepare them. While this is a less common practice in the United States, customers often do shop for food at the last minute."

I used to get suspicious whenever I saw a sentence that started with "In Europe". Now, however, I don't get suspicious. I know that the rest of the sentence will be high-grade bull pucky. I have family in the Netherlands, and not because they moved there for a job after growing up in the States: these would be the ones who didn't leave when my grandfather did. They shop for food about once a week, same as us. I think sentences like this come from people who do a capital-every-day-or-so and delude themselves into thinking that the people they meet in capital cities are anything like the rest of the country, but I'm not really sure.

Still, it's just not true, and so freaking many people in the US shop for food on a weekly basis that this argument is ridiculous. The 4 hour delivery window is a much more serious issue, and that's probably why most of the people I know who do online grocery shopping pick up the results at the store on their way home from work.

ETAYA: This is a little better:

http://www.internetretailer.com/2011/03/18/online-grocery-shoppers-home-delivery

And this is _way_ better.

http://therobinreport.com/online-based-grocery-delivery-thrives-at-last/

Alas, that website is as uneven as everything out there. Compare this:

http://therobinreport.com/why-shoppers-wont-do-supermarkets-work/

to this:

http://whatthedutch.wordpress.com/2011/02/26/grocery-shopping-of-the-future/

Shoppers will do the supermarkets work, if you structure that work correctly. Albert Heijn never gets rid of all the checkers, because they don't want to lose elderly customers who are perhaps not as enamored of the technology.

"Along with all these factors, a legal challenge has arisen in California. That state has a new law prohibiting the use of only customer-operated checkouts in stores that sell alcohol. The law was union backed and is clearly aimed at Fresh & Easy, the only chain in the nation that uses only self-checkout lanes. The chain is still contemplating how to conform to the new law."

Interestingly, the Robin Report coverage does mention Dutch supermarkets -- in the US.

"Supermarket chains Stop & Shop and Giant, both units of Ahold, issue hand-held devices to interested shoppers allowing them to scan as they fill their baskets. Fresh & Easy is experimenting with such a system too." Of course that's exactly the AH scheme, but not mention that it's up and running and successful elsewhere.

It is nice when someone else is the guinea pig, otoh, that does mean the disintermediation is that much more shocking when it finally happens. And checker jobs, honestly, are some of the better paid, better conditions jobs out there these days.

While I'm on the subject of supermarket cashiers and also on the subject of disintermediation, here's what BLS has to say about that job category:

http://www.bls.gov/ooh/sales/cashiers.htm

3ish million of them in the US, you can often get the job without even a high school diploma and make more than minimum wage (which is what I meant above). Grocery stores have more of them than any other category and even BLS expects slower than average (about half) growth in the category. Category averages young and even BLS expects tech to continue to impact the need.