Tuesday, October 22, 2013

Why they failed: Zeo, Green Goose, 100Plus

QS13 Here's what happened
Zeo was, for many of us, our favorite QS company: great, consumer-oriented low-cost hardware that measured something useful (sleep) to give insights we couldn't have had otherwise. After $20M+ in funding over the better part of a decade in existence, they went bankrupt, abruptly -- so quickly in fact that many customers couldn't even get their data off the machines.

In one of my favorite sessions at last week’s Quantified Self Conference in San Francisco, Zeo co-founder Ben Rubin joined a few other co-founders to discuss why their original dreams didn’t quite pan out.

Zeo’s biggest success factor, says Ben, was "persistence" -- to build a great product, get retail distribution in places like Best Buy, expand internationally to the UK, build a portable Bluetooth version -- but the same persistence was also the seeds of their undoing, when they continued to push sales- and marketing-wise even when it was clear that the mass market didn't want the headband they were selling. Sleep measurement alternatives popped up from the fitness band companies, and although they weren't nearly as precise, it became harder to explain the Zeo advantage. Zeo needed a similar passive solution (maybe something you attach to your bed, like Beddit), and they were working on it, but persistence takes a toll on founders -- one left in 2010, Ben left in 2011 -- and the professional management team that remained found themselves confronting the worst of all worlds: a market that said not "yes" or "no", but a dreadfully-ambiguous "maybe".

Ben’s advice to anybody thinking of a similar venture is to fit, somehow, into the lives that people live right now. You may be trying to change behavior (isn’t that the point of measuring it?) but the mass market wants a straight line from where they are right now. You can’t ask them to uproot existing habits to use your product.

So what happened to Zeo's assets?  Will the products ever be revived? Ben obviously has to be discreet when discussing confidential information, so all he would say is that it was an asset sale by “a medical company interested in sleep." So who knows.  He also, sadly, confirmed my worst fears, that he knows of no way  -- even through reverse-engineering -- to get your Zeo data. The company failed so quickly that the people in charge of maintaining the servers had no time to close things gracefully. It's gone.

Brian Krejcarek, founder of Green Goose, had another exciting product that failed: $4 attachable "stickers" with embeddable sensors you can place throughout your house -- on your toothbrush, your bike, your dog. It was a brilliant idea, and they quickly attracted $1.3M of funding. But the direct-to-consumer company is super-tough – how do you get the word out? how do you handle the logistics? Meanwhile, technology marches brutally forward, and the wireless base station that originally made the stuff work at great (but elegant) engineering cost, is now available on Bluetooth 4.0 and your iPhone for next to nothing.

Chris Hogg, co-founder of personalized health prediction startup 100Plus, had a comparatively happy ending: his company was sold, to Practice Fusion (a $70M Series D funded medical records company). But Chris left when that happened, because he was always more interested in the “personal” side of health, not Big Enterprises, which is where the company ended up making its first money. In fact, that was one of his pieces of advice: “be careful where you get that first dollar, because that’s all next your investors will want to talk about.”

That’s the tension in every new business: on the one hand, you want to be flexible and listen to your customers; but on the other hand, you want to be true to your original mission.  When you find that consumers don’t bite, either because the product’s wrong or the technology has moved on or that enterprises turn out to be more interested, you may find that your original dream no longer applies.
All three entrepreneurs knew their original ideas were worthwhile and we’ll get there someday. But sometimes the future happens later than we’d like.

Monday, October 21, 2013

Are public schools better than private?

The Atlantic has a short interview with University of Illinois - Urbana-Champaign professors Sarah Theule Lubienski, and her husband Christopher A. Lubienski: “a new book argues that public schools are actually academically superior”.

I don’t have the book (The Public School Advantage- Why Public Schools Outperform Private Schools to be published in November 2013), but the authors wrote a 2006 paper that seems to conclude the same thing. Essentially they use a large dataset to argue that, while private schools do outperform public schools overall, the advantage disappears when you account for demographics of the students.

This would indeed be interesting if it were true. But a quick look at the paper makes me wonder about some obvious mistakes in methodology behind the  headline-grabbing “conclusion”.

  • 150K public students (96%) vs. 6500 private students (4%)
  • 2K Catholic, 2K charter, 1K “Christian”
  • Of 6,000 schools overall, only 150 are “other private”.

Since almost all of their data from private schools is from religious schools (mostly Catholic), shouldn’t that be the headline? This is NOT comparing your local school with the $20K/year highly-selective boarding school that many people imagine when they think “private school”.  As the study itself points out, there are many reasons parents might shell out extra money to send their kids to a school, but religion is a big one that, if anything, would trump "academics” in a lot of cases.

As always, my conclusion is to wonder how useful it is to know in aggregate whether something as variable as education is better done one way or another. What matters is what’s good for your kid. Trying to make a generalization about education systems based on a database of thousands of schools is like trying to predict the value of your home by looking at trends in US real estate. Who cares?

Friday, October 18, 2013

[book] Wealth and Power

I read this because, with all the recommendations I'd received, well, I had to. When I first moved to China and was looking to start reading up about the place, I followed the advice of David Moser on an early Sinica Podcast to "read anything by Orville Schell". When none other than Sinica host Kaiser Kuo mentioned that Schell had a new book coming in June 2013, I immediately put it into my reading queue and this week I finished it.

A nice overview of key people in China's past two hundred years, the book has chapters with the usual suspects like Mao, Chiang Kai Shek, and the Qing Empress Dowager Cixi, but you really should read it for its descriptions of less well-known (to the West) people like Wei Yuan (the guy who during the Opium Wars first worried about China's "humiliation"), Liang Qichao (the scholar/journalist whose New Citizen journal influenced everyone remaking China in the early 20th Century), Chen Duxiu (energetic founder of the CCP, who introduced Mao to communism but was later ostracized) and dissidents like Liu Xiaobo (winner of the 2010 Nobel Peace Prize).

After reading the devastating Jung/Halliday book Mao: The Unknown Story, I found Schell's relatively mainstream depiction of both Mao and Chiang Kai Shek to be bland. Unless you're not already familiar with Edgar Snow's cheerful picture of Mao as heroic fighter against Nationalist corruption, you can skip those chapters. I'm being dramatic, of course -- anyone whose byline is on the Tiananmen Papers is no lackey for the CCP -- but to a curious, outside observer like me, deep skepticism is justified about everything people assume is true about Mao, and this book largely repaints the standard picture in ways that ensure the authors won't have their China visas revoked.

The last chapters, with its late-breaking history of China between 1989 and Liu Xiaobo were the most interesting to me. China, having lately received the wealth and power it has been lacking for the past two hundred years, may have found in the CCP a powerful new form of adaptable government, one that is intensely pragmatic in ways that seamlessly replace the capriciousness of revolutionary Mao with the flexibility required to steer a huge country into modernity: "resilient authoritarianism", as Columbia professor Andrew Nathan calls it.

Another final anecdote I liked in the book quotes famous sinologist Douglas Fairbanks lecturing a student who wants to compare China too much to the West: "Mao didn't make the Chinese Communist revolution for you", he says. China is on its own timetable, and the interesting characters portrayed in Wealth and Power are reminders that you can't understand the future unless you understand what's happening in China.



Wednesday, October 02, 2013

Charts of RAW data

The Density Design Lab at Politecnico di Milano just released a cool new (free) web site, called RAW, for very simple data visualization. It’s roughly similar to products like Statwing, Tableau Public or DataHero, except this one gives you the entire d3.js source code on GitHub.

I made the following chart using some data downloaded from Charles Murray’s excellent book Coming Apart, that has income and college graduation information for every US zip code. I simply selected the spreadsheet data right in Excel, then copy/pasted it – as is – to the RAW web site. A few clicks later and I had this:


The rectangles are US cities with more than 15,000 families organized by state. The size of each rectangle correlates with the median family income, and the color represents the percentage of people with college degrees on a spectrum from blue (lower percentage) to red (higher percentage). That big brown square on the middle bottom is New York City, with a high median family income and a substantial percentage of college grads. The California cities on the far right are bluish-green because they have lower diploma percentages.

Pretty cool, huh? I have a long wishlist for additional features (additional chart types, labeling options, etc.) but you have to admit this is a pretty easy way to do impressive data visualization for free, with only a few mouse clicks.