1. Hardware by the Numbers: A Narrative

    Last year in July, I did some preliminary research on the hardware startup ecosystem. The goal was to examine the growth of the sector by looking at funding events, on both the venture and crowdfunding front.

    This year, for my keynote at O’Reilly’s Solid conference, I wanted to put together a much more in-depth narrative looking at the trends that are shaping hardware, and leading to that growth. I’m interested in the growth of a community that encourages founders; the new support ecosystem that facilitates building hardware businesses; the manufacturing trends that govern where and how things are made; and the types of businesses that flourish as a result.

    This presentation and talk summarize that research.

    Here’s an updated table of Kickstarter data* :

    Year 2011 2012 2013
    Hardware Projects 120 239 784
    Product Design Projects 649 1373 2453
    Total Projects 26,124 41,440 46,100
    HW as percentage of total 0.46% 0.58% 1.70%
    PD as percentage of total 2.48% 3.31% 5.32%
    Hardware project dollars raised $2,543,850 $9,768,304 $46,645,309
    All project dollars raised $104,625,478 $323,555,323 $469,452,086
    Hardware $$ as percentage overall 2.43% 3.02% 9.94%
    Hardware project number of backers 19,019 91,996 370,189
    All backers 1,409,190 4,365,213 6,369,255
    HW backers as percentage of overall 1.35% 2.11% 5.81%

    And an updated bubble chart showing funding events (Series A and earlier; this latest version includes medical devices):

    Botsourcing stats…

    My slides (though, really, view them on Presentate - speaker notes and all citations are right there next to the relevant slide, it’s a much better experience):

    I’m inspired by numbers. I look forward to Mary Meeker’s Internet Trends every year (and it’s coming out this week!). One of the most challenging aspects of transitioning from Wall Street to seed-stage VC was the relative absence of quantitative data for decision-making. The fact is, while I’ve made a vested effort to source and cite each and every piece of information in this presentation, the underlying data is often dirty or scattered.

    If you compare this data set with my post from last year, you’ll undoubtedly find things that are included in one but not the other - this is due to changes in the classification taxonomies of the source databases. Crunchbase 2.0 has made a lot of progress and is a valuable resource, but even it is dependent on users making edits as they discover errors or incomplete information (ie, Boston Dynamics being classified solely as a software company). If you discover any glaring errors, please point them out to me (ideally, kindly!).

    If you’d like to chat about other interesting facets of analysis to add going forward, get in touch!

    A million thanks to the teams at Crunchbase, Octopart, and my very talented husband, Justin Hileman, for going above and beyond in helping with this project.

    *with the same caveat as last year: Kickstarter campaigns are limited to one category, so some consumer electronics projects, like Pebble, are in Product Design.


  2. Long-overdue update: book & baby!

    So, I fell off blogging here for a while (though I still wrote occasionally at O’Reilly Radar). This is why.

    I’ve been working on a book with my friend Nick Pinkston. It’s called The Hardware Startup: Building Your Product, Business, and Brand. The first four chapters were just released at O’Reilly Solid. The pre-release stamp is pretty funny - Raw and Unedited! Makes it sound like a sordid tell-all. :)

    It’s been the most time-consuming project I’ve ever undertaken. Both of us want the book to be genuinely useful to hardware founders. Software startups have some great roadmaps, such as “Do More Faster” and “Lean Startup”…there isn’t much long-form material out there for founders working through manufacturing challenges on the product while simultaneously building a business. We’ve been packing it full of case studies with entrepreneurs who have worked through some of the really complex processes unique to starting a hardware business.

    Putting out an unfinished pre-release teaser was challenging for me, since I like things to be polished before they see the light of day. It’s kind of a minimum viable product for the book, though, and the goal is to get feedback (which includes negative feedback). If you’re a hardware founder and you have thoughts about what you’d like to see, feel free to comment.

    Anyway, tying this back to my last post…I had this crazy idea that I’d be able to get the book done while pregnant. That didn’t happen, but it was primarily because I decided to pack in a bunch of work-related travel right up until 36 weeks. After China, I did a bunch of stuff within the US and then went to a conference in Turkey. My flight back to SF was on the last day I was legally allowed to be on an airplane. :) (The Air Berlin folks weren’t pleased.)

    Then I had our baby - I’ve blogged about that a bit on my personal/DIY project blog, and will eventually post something here about the maternity leave experience. In brief, though, Xander was born on December 18th. He’s now 5 months old. He’s a real handful but getting more fun by the day. Obligatory proud mama photos below.

    In the hospital

    2.5 mo family photo shoot

    Yesterday at just past 5 mos.

    I also thought I would get the book done while on maternity leave, and that was even more laughably unrealistic. For me, writing requires sustained focus and a big chunk of time. I was able to keep up with email and meetings and discrete tasks, but with a baby who woke hourly and never napped, writing was pretty difficult. I was relieved when we started daycare; I felt productive again.

    So! Now that things have settled into a routine, we’ll be putting the finishing touches on the book and I’m looking forward to posting here more regularly.


  3. Pregnancy and competence

    I just got back from ten days in China - factory tours in Shenzhen! photos and video forthcoming! - and am getting caught up on news.

    Anti-Foreigner VC Also Supports Hiring Discrimination
    One advantage startups have over established companies is that there are no discrimination laws about starting businesses. For example, I would be reluctant to start a startup with a woman who had small children, or was likely to have them soon. But you’re not allowed to ask prospective employees if they plan to have kids soon. Believe it or not, under current US law, you’re not even allowed to discriminate on the basis of intelligence. Whereas when you’re starting a company, you can discriminate on any basis you want about who you start it with.

    It is amazing to me that this is news.

    It’s amazing because almost every woman I know in the tech industry has known about that Paul Graham quote for years. It’s occasionally still brought up at diversity dinners - when people ask why there aren’t more female founders, or more women VCs. It’s not causal, obviously, but it is indicative of an undertone of condoned discrimination. Occasionally a male colleague hears about it for the first time and doesn’t believe it until he’s read it for himself.

    I’m six months pregnant at the moment. Maybe that means I’m less capable than I was six months ago. Maybe it means that in another three months my productivity will suffer some sort of precipitous decline, from which it won’t recover for an indeterminate period of time. No one seems to wonder whether my husband’s work ethic will suffer the same fate.

    I don’t talk about gender issues very much on this blog because I prefer to focus on things I do, rather than things that happen to be. But during in my time in male-dominated fields, I’ve gotten the sense that people think differently about professional women during and after their transition to motherhood. That perception has long made me rather ambivalent about having a baby in the first place - what would I be doing to my career, exactly?

    It’s not really PC to talk about that kind of ambivalence; being concerned about that sort of thing makes a woman sound like a callous careerist with misplaced priorities. But as we’ve entered our late 20s and early 30s, most of my ambitious female friends agonize about “when to do it” and wonder about the economics of freezing eggs. When I was in finance, we joked about it in terms of amortization. Conventional wisdom says that there’s never a good time for anyone - but when one of the gurus of your industry is publicly on record talking about how he wouldn’t do a startup with a new mommy, and the joys of being able to discriminate, there’s a sense that your marketability somehow decreases once you’ve reproduced.

    This is one of the reasons why I had such a difficult time with “Lean In.” It’s a great message, but it falls short against the reality of cultural conditioning in many industries. Many women I’ve spoken with feel that after their kids, they’ve had to work even harder just to be taken as seriously as they were before. I’m trying to come up with parallels that male colleagues experience, but falling short. Maybe the best example would be a male employee having major surgery. Sure, there will be a period of adjustment for the company and the individual, but the most likely outcome is known and can be planned for. The person may be physically impaired during recovery, but it’s assumed that his professional capability will be back to normal in no time.

    So maybe it’s good that this PG story is news. Maybe it shows that attitudes are changing. But for the moment, I’m still wondering: Can motherhood ever be a biological rite of passage without being a professional one?


  4. Hardware, by the numbers

    Conventional wisdom around Silicon Valley is that hardware is hot. Influential incubator founders and VCs have been blogging about their interest in the space, and popular tech blogs are constantly writing about prototyping technologies and hardware startup fundraises. However, there’s been very little in the way of actual numbers or data documenting the growth of the ecosystem.

    So, let’s have a look at hardware by the numbers: the number of companies started, the flow of venture and crowdfunding dollars, the overall growth of the hardware community, and the degree of broader public interest.

    It’s tough to track exactly how many hardware startups are out there, or whether that number has increased over time. One interesting data point to start with is the number that have gone out and attempted to fundraise. AngelList keeps track of this, and was generous enough to share their data. In 2011, 94 hardware companies went out and opened rounds on AngelList…25 of them successfully closed a round (though not necessarily via AngelList’s platform). In 2012, that number jumped to 267 opened, 96 closed. So far, as of mid-July 2013, 259 have already opened rounds, and there have been 46 closes. While fundraises aren’t an exact proxy for number of startups, those figures make it seem like there is an increasing number of hardware companies. (The slight dip in percentage closed is also an interesting thing to note.)

    Let’s have a look at the actual dollar amounts that angels and VCs are investing in hardware startups. Crunchbase and AngelList both track this information. Since we’re looking for indications of a hot market, I pulled data for angel, Series A, and seed rounds.* The information below comes from Crunchbase and is represented in an interactive chart form so you can dig in and see the underlying funding events. The chart is derived from Crunchbase’s funding data from all 12 months of 2011 and 2012; the data from 2013 stops at 30 June 2013.

    There were 52 early-funding events in 2012, totaling $209.5MM; in the first six months of 2013, there have already been 47, totaling $222.9MM.

    Since crowdfunding has had a profound impact on the growth of the hardware startup ecosystem, let’s have a look at Kickstarter data as well. Kickspy is a great resource for tracking Kickstarter project stats.

    Year 2011 2012 2013
    Hardware Projects 121 248 365
    Total Projects 26,319 42,113 23,274
    HW as percentage of total 0.46% 0.59% 1.57%
    Hardware project dollars raised $2,551,986 $9,750,168 $22,683,014
    Hardware project number of backers 19,181 93,870 195,449
    All backers 1,438,657 4,496,155 3,354,019
    HW backers as percentage of overall 1.33% 2.09% 5.83%

    Based on the data, in the first seven months of 2013 we can see an increase in the percentage of hardware** projects on the site, and an increasing percentage of the site’s backers are helping to fund hardware projects.

    Update 8/1/2013: Someone helpfully pointed out that HW projects on Kickstarter are split between the “Hardware” and the “Product Design” categories. So, the hardware numbers above are actually a conservative estimate. The Product Design category is a bit messy (any physical good can wind up there), but there were 1,809 projects launched (419 successful) that raised $19MM USD from 1/1-6/30/2013. In 2012, there were 570 successful projects that raised $41.1M. Check out Kickspy for more details.

    To gauge community interest, Nick Pinkston recently looked at Hardware Startup Meetup data. The growth in the San Francisco membership roster is interesting (see graph below). Nick has some additional stats about community on his blog, including a map of hardware meetups around the country.

    But what’s happening outside the tech and early-adopter communities of Silicon Valley? Checking Google Trends for “hardware startup” was completely inconclusive (seems like a lot of people search for that term when they’re having a difficult time getting their computers to boot up). However, looking at specific hardware-focused technology terms shows that interest in topics such as the “Internet of Things” and “3D printing” is at an all-time high.

    The examples of growth cited in this post are based on potentially incomplete data. There are some confounding factors (ie, AngelList and Kickstarter are themselves startups that have experienced high growth in their own audience between 2011-2013). There are also other indications of growth that I didn’t touch on here - the arrival of top-notch hardware-specific incubators, and the new crop of logistics and distribution startups designed to help the hardware startups take their products to market. However, overall, all of the data seems to validate the conventional wisdom that investors are paying an increased amount of attention to hardware companies these days. It’s great to see such a thriving ecosystem.

    If you know of any other interesting data or touchpoints around the growth of hardware startups, please share in the comments below!

    Many thanks to the folks at AngelList, Crunchbase, and Kickspy for their help in pulling together various data sets for this post, and to Justin Hileman for his help with the viz. Also: the folks at Hizook have been keeping tabs on the robotics subset of hardware startups - check out the funding data they’ve compiled!

    *You’ll notice some outliers; while most people traditionally think of a Series A as being under $10MM, Crunchbase applies the term more liberally; for example, Anki and Lytro’s $50MM “Series A” rounds. I didn’t include B or later because I felt that those rounds were a function of the specific successes of individual companies than a reflection of investors making early bets on a hot vertical. Crunchbase has it, though, so you can get it if you’d like.

    **There is a potential for misclassification to be skewing results here – some projects are classified only as “technology,” and lack the more specific ‘hardware’ tag.


  5. Taking VC funding, and tracking metrics

    Two more Dear Abby posts are up! It’s been really interesting to see what people are asking. Last week, an entrepreneur asked me whether taking VC funding is necessary at all. This is an increasingly hot topic, particularly as equity crowdfunding rule changes have the potential to impact the early-funding landscape. Here’s an excerpt:

    Does every startup need venture funding? Sometimes I wonder if founders think they need funding because it is said so often that they simply start believing it. Do you think if a startup has a true solution and an effective story, venture is needed? - Female founder, health startup

    There are definitely situations in which venture funding is not needed. The primary things to consider are how competitive your market is, and how quickly you are likely to reach a point where your business is bringing in cash. If the market is competitive and a first-mover has a clear advantage, outside funding may be a good idea.

    If your true solution and effective story yields results – meaning you begin to gain widespread adoption during the early bootstrapped phase – the dilution you’d take from venture funding may not be worth it. Venture capital is for scaling more quickly than otherwise possible and not every company needs (or wants) to do that. GitHub is one example of a company that became successful without venture funding; they provided a service that their customers loved, and they grew organically. While they did eventually take VC funding, their first round – $100 million - happened when the company was four years old and needed some cash reserves and capital to continue to scale.

    This week’s question was about metrics. Specifically, which metrics are most important to monitor following seed capital, while looking ahead to a future Series A raise. It’s a question that reflects the entrepreneur’s desire to avoid the “Series A crunch.”

    Once a promising seed-stage company is funded, what are the business milestones that stage-A VCs like to see before investing? -Female technologist, San Francisco

    Meaningful ones! That answer isn’t a cop-out; it really varies according to the type of business the company is in. Generally speaking, the metrics that matter are the ones that track meaningful customer engagement with your product. B2B metrics are very different from B2C, and investors evaluating a company will compare your metrics to comparable companies in the same sector.

    Seed stage capital is for finding product-market fit. You want to show trends over time which demonstrate that fit: user numbers continuing to increase, revenue (if you have it) picking up, a growing number of enterprise clients moving through your pipeline. The specifics may vary, but ultimately you’re looking to demonstrate that your market is excited about what you’re building.

    Keep the questions coming!


  6. Dear Abby for Tech: due diligence

    I’ve recently started writing a series of posts for Women 2.0. Instead of picking a topic each week, we’ve been soliciting questions from the community to get at the areas that people feel most uncertain about. We kicked off the new column with a question about diligence:

    How do you (and the early-stage VC community in general) like to conduct due diligence on the companies you are interested in funding?

    The purpose of diligence is to answer Who, What, When, Where, and Why questions about a prospective deal. At a seed-stage fund, this typically involves four things: market research, competitive research, customer feedback, and founder reference checks. Taken together, these present an investor with as comprehensive a view of a deal as possible. A fifth element, much more common in later stage investing, is an examination of financials.


    Market research involves looking into trends in a given sector. This type of research is often done well in advance of a specific opportunity coming in, for the purpose of understanding a hot market or with an eye toward crafting an investment strategy. This is because it’s time-consuming, and may be rushed if done reactively around a particular deal. However, sometimes it does happen in conjunction with diligencing a particular deal. Some common questions that investors try to answer include:

    • What are the drivers in that segment of the economy?
    • Is the addressable market of customers growing or shrinking?
    • Is there an underlying regulatory framework that dictates rates, prices, or behavior?
    • Are there changes happening that make this market particularly prone to disruption in the near future?

    Competitive landscape research involves examining a sector to determine who the component players are. The first step is to identify the large incumbents – the 800-lb gorillas who serve the majority of customer needs in the space. Then, there is a survey of smaller players and other startups. Who has entered the space recently? Are those startups funded or showing evidence of traction? Is this a “land grab” situation, where dozens of startups are trying to differentiate with highly-specific niche features, or are there few entrants (few entrants isn’t necessarily more ideal; that may mean there’s no real demand)? The competitive companies are then compared to the startup under consideration, typically on pricing, the offering, and/or the technology.

    Read the rest here. And if you’d like to send in a question, contact askavc@women2.com

  7. Revisiting bitcon price vs. search volume

    On March 19th, The Economist posted an interesting chart and analysis entitled “Bitcoin’s record price looks like a bubble.”


    I reworked that chart in light of the market correction (or crash, if you prefer) over the past couple of days


    Note that the old peak search, around June of 2011, has been completely eclipsed by the volume of searches for “bitcoin” now.

    Interestingly, “bubble” is not among the top 20 searches for “bitcoin” in the Trends report for the overall period (chart below). It was also not in the top 20 during the thirty days surrounding the previous peak in June of 2011. However, it is showing up as a “breakout search” in data from the past 30 days.


  8. The curious economics of the “sharing economy”

    Venkatesh Rao published a thought-provoking piece on Ribbonfarm today, comparing the deal-seeking masses who use Groupon and Airbnb with a swarm of locusts. It’s a bit of an intense analogy, but the basic premise is that certain aggregator business models are terrible for the underlying businesses, and that platform middlemen are capitalizing on the predatory instincts of deal-seekers:

    They draw in nomadic deal hunters from a vast surrounding region who are unlikely to ever return; … most deal-hunters carefully ensure that they spend just the deal amount or slightly more; … a badly designed offer can bankrupt a small business.

    Locust economies are built around 3-way markets: a swarming platform “organizer” player who efficiently disseminates information about transient, local resource surpluses, a locust species in dormant grasshopper mode, and a base for predation that exhibits a scarcity-abundance cycle.

    (You should probably just go read it.)

    Rao’s piece paints collaborative consumption and daily deals with the same broad brush, but he’s getting at something interesting: despite the idealism underlying the sharing economy, the attempt to turn sharing into big business hasn’t meaningfully transformed self-serving consumer behavior. Most people aren’t there for the sharing; they’re there for the savings.

    Building a marketplace requires incentivizing both buyers and sellers. In the idealized sharing economy, the buyer is looking for a unique and authentic experience (community), and values access over ownership (sustainability). In reality, he is most often looking for a deal. The economic incentive is more honestly articulated on the other side of the market: the seller is looking to capitalize on a resource that is otherwise sitting around unused.

    The reality of the “sharing economy” is that most of the p2p startups born of the movement aren’t going to survive. They’ve taken venture funding, but they aren’t venture-fundable businesses. The economics just don’t work.

    Airbnb has become a household name. It’s the X in “We’re X-for-Y” in pitch decks across Silicon Valley. It’s a success story. And this is because it’s one of the very few exceptions to the locust phenomena described by Rao1. Even assuming bargain-hunting buyers and a cut taken by middlemen, Airbnb generates material value for the seller. The average rental on Airbnb varies quite a bit by city, but in NYC and SF a seller can earn ~$150/night renting her place. Even with Airbnb’s 6-12% cut, renting an underutilized apartment for a weekend nets ~$275. It’s also a low-friction experience for the seller: the apartment is in a fixed location (the buyer comes to you), lockboxes and entry systems are largely a solved problem, and the platform enables cleaning fees and security deposits. It’s tough to do enough damage to render a place uninhabitable (low rate of asset depreciation). There is a meaningful rate of return relative to the time necessary to manage a place. And on the buyer side, Airbnb yields significant savings over staying in most hotels (it is also, often, a genuinely unique experience). Low friction, high rate of return, benefits for all involved…a thriving marketplace.

    No locusts.

    But even within the same economic climate that ostensibly helped Airbnb succeed, there has been no comparable runaway success in p2p sharing of cars, bikes, or household goods. And that’s because renting out those items is a high-friction experience with a low rate of return. Example: The average cost to rent a car on RelayRides or GetAround in San Francisco is $8/hr. The car owner keeps 60%, so he’s making $4.80 per hour before taxes. And for that, he’s putting more miles on his car (a rapidly depreciating asset), risking moral hazard on the part of the borrower, and taking a chance on an accident rendering the asset potentially unusable for a long period of time. Sure, the middleman platforms offer insurance, but most people who have endured the hassle of an insurance claim - and managing vehicle repairs - will tell you that $4.80/hr simply isn’t worth it.

    As the seller is earning $4.80/hr, the platform is collecting $3.20…and paying for insurance, marketing, salaries, etc. Given how low revenue per transaction is, this is a model that needs to hit massive scale in order to generate meaningful returns for the venture-backed middleman. Because the locust problem is an equilibrium problem2. Incentivizing enough sellers to participate in the market at rates low enough to compete with Zipcar3 (and with each other) is tough. But if the rate is pushed high enough to make it worthwhile for the seller - and to generate a sufficient cut for the middleman - it is a far less exciting prospect for the buyer4. Buyers will only sacrifice convenience and quality if the price is compelling.


    Unless, of course, the buyer and seller are participating in the market because they’re truly motivated, first and foremost, by community and connection. Those are powerful forces. They’re the reason that co-op businesses and credit unions thrive in many neighborhoods. At a smaller scale, typically within a discrete community with strong ties, p2p collaborative consumption works.

    No locusts.

    But also: no venture-backed middlemen.

    1. Rao thinks it is but I disagree.
    2. Which Airbnb may also encounter at some point
    3. Zipcar, which rents in SF at ~$10/hr - $2 more than RelayRides’ average - is a known quantity. Clean cars, effortless entry, convenient to park and return. At this time, it seems they’re still operating at a loss.
    4. The exception to this may be markets in which regulatory arbitrage enables P2P businesses to thrive (ie, Lyft, Sidecar, Uber confronting the false scarcity of the existing medallion system). It remains to be seen if this is temporary success; it’s somewhat dependent on legislation and enforcement.


  9. Everest Base Camp

    It took me a couple of months to get these photos up, but from Dec 22-Jan 6th, Justin and I did the Everest Base Camp Trek. It was incredible. The days were strenuous, the nights were freezing, but we saw some of the most breathtaking mountains in the world, made new friends, and experienced a fascinating culture that thrives despite the inhospitable landscape.


    We started the trip via a flight into Kathmandu by way of Hong Kong. My luggage stayed behind in San Francisco. Pro tip: always wear a complete set of layers and hiking boots on the plane…I would have been so screwed if I hadn’t. The streets in Kathmandu are lined with shops selling knockoff North Face, so I was able to fill in the gaps by renting cheapies.


    The first day was mostly sightseeing around the Buddhist and Hindu sites in Kathmandu. Early on the second, we woke up and boarded the first flight to Tenzing-Hillary Airport in Lukla. The airport is one of the most dangerous in the world because of its very short, sloped runway, infamous winds, and the ring of mountains surrounding it. The landing was exhilarating - immediately upon touching down, our pilot slammed on the brakes and we screeched to a halt. The 70-mile trek begins right on the runway, so we set off on a three-hour hike to the town of Phakding.


    We were traveling in the winter (off season), and it was very cold at night (5 degrees F). I thought that since we were staying in tea houses, we’d still be warm. Turns out, they aren’t insulated and only the common areas have (intermittent) heat and electricity. Most have single-pane windows and a squat toilet, because Western plumbing is more prone to freezing. The beds were single-size cots; you’d throw your sleeping bag on, and then borrow a quilt. On most days, we would hike from 10am-3pm while the sun was out, then spend the afternoon reading and playing cards in the main room, which was heated by a wood or yak-dung-burning stove.


    Tea houses all have the same menu. Locals eat a dish called dal baat several times a day - it’s a bowl of lentil soup, and a plate of rice with potato, cabbage, and carrot curry on top. Everything on the menu is white carbs (noodles, potatoes, rice). Stick with the native food; the attempts at pizza and pasta are a nice gesture, but they taste pretty bad. There are almost no animal products, because livestock are more valuable as beasts of burden…you’re better off not eating meat past Namche anyway, since there’s no refrigeration. All provisions are flown in from Kathmandu and then carried on the backs of donkeys, yaks, or yak-cow hybrids called dzoom. Provisions get progressively more expensive as you approach Everest, because it’s increasingly more difficult to carry them. Snickers bars and bottles of spring water are luxuries.


    Trekkers learn pretty quickly that the donkey and dzoom trains have the right of way. You stay to the ‘mountain side’ of the trail as they pass, or run the risk of being shoved off a cliff or into the river. On Day 3 we crossed our first high suspension bridges. The bridges are safe, but they’re still scarily high and pretty long. There is a particularly windy bridge that marks the entrance to Namche Bazaar. On the far side is a very rocky, steep set of stairs. We happened to be behind a donkey train on that bridge, and they didn’t like the steps so they kept backing up onto the bridge and making it sway. The locals tie prayer flags to the bridge ropes…they were swirling in the wind all around us, and waaay down below were the bright blue rapids of the glacial runoff. It was simultaneously beautiful and scary. I always let the livestock get completely off the bridge after that.


    Part of hiking to high altitude is acclimatizing properly, so we spent two nights at Namche (11,482 ft). On our rest day, we went shopping for candy bars, paid for hot showers (the only ones we would get until we were back in Namche again), and watched “Into Thin Air” over popcorn at a local bar. Since it was off-season, a lot of the businesses were shuttered. This made it easier to make friends, since we saw the same familiar faces at the few lodges that were open. One friendly American had packed a Santa suit, and he passed out little gifts. Christmas dinner was Ramen soup.


    After Namche the weather got markedly colder. We walked to Tengboche, site of the beautiful ancient monastery. Most of the monks had gone to Kathmandu to escape the cold, but two remained and we watched them chant. We walked through rhododendron forests, had our first clear views of Everest, and saw some giant eagles and our first yaks.


    After a miserably cold night, it was on to Dengboche, which is in the shadow of Ama Dablam and is absolutely gorgeous. Dengboche is at 14,862 feet, so we spent two nights there as well, with an acclimatization hike on the second day. We both felt fine and still had great appetites and were sleeping well, and hit the road to Lobuche with a lot of energy. Along the way is a series of stupas and rock piles with prayer flags commemorating those who have died on Everest. It’s a sobering place; the piles we saw were as recent as May 2012. Famous climbers, such as Scott Fisher and Anatoli Boukreev, are memorialized there.


    This is often the point where people start to get sick. My appetite vanished. The last half of the six-hour hike was on a vast open plain and the wind was brutal. We’d been lucky enough to have perfectly clear skies over the first seven days, but the eighth brought clouds and snow. It gets very difficult to sleep above 16,000 feet because the air is so thin. Fatigue + reduced caloric intake + the beginnings of mild altitude sickness made me pretty grouchy. I was also taking Diamox at this point, which is a great drug for reducing altitude symptoms but makes your ears ring, gives you pins and needles, and it’s a diuretic (so you pee a lot). Still, I was lucky. Some of our traveling companions had severe altitude sickness. One woman we’d spent several nights playing cards with was medevac’d out because she developed cerebral edema.


    Base Camp day is a very long day. We left Lobuche early, hiked three hours (3 miles) to Gorak Shep at 17,000 feet, and had a quick lunch. Gorak Shep is the “old” base camp. The name means “Dead Ravens,” and there’s really nothing more than a few trailers there. It sits on a dry lake bed. We changed into our warmest gear and set out for EBC, which sits at 17,600ft and a four hour hike over the Khombu Glacier. I actually hadn’t realized that Base Camp is on the glacier itself. It’s a bit of a slog to get there, because the glacier is covered with pebbles and it’s tough to get a foothold. But we made it, and I was thrilled to be standing there. It’s gorgeous and imposing and awe-inspiring. Nothing is alive, but the rocks themselves are constantly moving – the glacier is shifting and avalanches are happening, and the environment is in a constant state of flux. The mountains are so unbelievably vast and tall. There’s nothing in the United States on that scale.


    Anyway: mission accomplished! After some celebratory pictures with friends, we turned around and hiked the four hours back to Gorak Shep. It was New Years’ Eve, and we celebrated in style with a fried Snickers pie and were all passed out by 9pm. The next morning Justin hiked up to Kala Pattar and took the gorgeous panoramic photos below. I made it ¾ of the way to the summit before a total lack of energy made me turn around. We still had a seven hour descent ahead of us, and it’s good to respect your limits. I wish I could find some way to keep my appetite strong when I’m that high up, it really makes a difference.


    I know this is a very long post, so if you made it this far, thanks for reading. The rest of the trip involved several more days of hiking back the way we came, a bumpy flight out of Lukla, and a celebratory steak in Kathmandu. I’ve got some more photos up here. It was a bucket-list trip for me, and it was just incredible. If you’re considering doing a trek or visiting Nepal, I can’t recommend it enough.



  10. Visualizing the future of government

    I believe that the future of government is one of increased transparency and collaborative problem-solving. What gets us there is creating a culture of participation, in which citizens across industries contribute their expertise to help solve our shared difficult problems.

    In November I got to participate in furthering open government at Code for America’s “Data Deathmatch” hackathon. California’s Fair Practice Political Commission provided hackers with a set of recently-digitized Statement of Economic Interest (Form 700) judicial gift disclosure and campaign filings data. They asked us to help them come up with tools that would generate insights and increase transparency.

    The team I joined — Team OpenJudge — consisted of seven talented people with diverse backgrounds and skill sets. Three of us spent hours working on a system for cleaning the data; although the digitization done by Captricity was phenomenal, the “human factor” (judges not following directions when filling out the form) caused significant messiness and the subsequent need for manual review.

    During the cleaning process we found some interesting activity: one judge, impressively, had managed to receive two different “1/2 iPad” gifts. One donor had given the same painting of a courthouse to nine different judges. To present the information, we made “leaderboards” showing top donors, judges who had received the most gifts, and the highest-value gifts. We also created a word cloud to highlight the relative frequency of different gift types: lunches, dinners, gift certificates, baseball tickets,etc.

    But the really important insights emerged when we made a network graph*. In the viz below, red dots are judges, and blue are donors*.

    Even though I’d spent hours looking at that data set, the pockets of interconnectedness that you see above didn’t pop out. And that’s the power of visualization: it can make those relationships immediately obvious. It enables a degree of transparency far beyond simply making a database of judges and donors available to the public; it makes the data comprehensible to citizens, not just motivated analysts and engineers.

    The FPPC invited us to Sacramento to present our work, and we spoke at their Commission meeting]this past Wednesday. They were excited about what we’d managed to accomplish in one weekend, and spoke of their vision of making meaningful information available to the people of California. When we got to the network visualization, they immediately grasped the potential, and asked a lot of questions about how it would handle new data, and what it could display. One of the commissioners happened to know a donor I clicked on during the demo, and approached us afterward to check out the details.

    I share this not because it’s a great visualization — there are important factors in the judge/donor relationship that aren’t encoded here — but because I am struck by how new and potentially game-changing data visualization is for public servants. As they consider what kind of tools to commission to make data more open and transparent, we’ve helped inspire them to make something far more interactive and compelling than a data table on a static site. We’ve hopefully also demonstrated that with the right partners, transparency and usability can be achieved both rapidly and inexpensively.

    Though implementing more transparent data visualizations won’t revolutionize government transparency and problem-solving on its own, I do think this is a meaningful first step toward bringing tech tools into government. I’m proud to have spent a weekend with a talented team who contributed valuable time and expertise to realizing a scaled-down — but nonetheless effective — culture of participation which successfully generated insights and enhanced transparency. And I’m looking forward to the next Code for America hackathon!

    *identity information has been removed; it will be available on their site, but I didn’t want to post it on my blog.