Thursday, April 27, 2006

On Cloning YCombinator in Europe

A YCombinator in Europe? Paul Graham seems to think it can't be done:
"Hackers are unruly. That is the essence of hacking. And it is also the essence of American-ness. It is no accident that Silicon Valley is in America, and not France, or Germany, or England, or Japan. In those countries, people color inside the lines."
Business Week chief economist Michael Mandel agrees:
"Where does the U.S. competitive advantage come from?

First, the ability to finance and reward high-risk startups. No other country has the same well-developed venture capital structure that we do.

Second, the willingness to work long hours is essential. You can’t make new technology work without immense inputs of time."
YCombinator is a seed venture capital firm. They recruit groups of founders straight from university, give them $6000 per person to work on their idea for a summer (or winter) and take 5-7% of the equity in exchange. In effect, their business model is much like that of a traditional VC, but with lower stakes since they get in at the seed phase. This has the potential to be a highly profitable business.

Can we clone YCombinator in Europe? I'm going to argue that it's hard, but not for the reasons that Paul Graham and Michael Mandel cite. Instead, it's hard because YCombinator has unique competitive advantages that are hard to replicate, even in the US.

This is a thinking exercise; I'm not planning on cloning YCombinator anytime soon.

First, let's defuse the Europeans-are-lazy myth. While the average US employee works 1777 hours per year, the average German puts in just 1362 – that's 24% less! But remember that these values are averaged over the entire working population. Contrary to popular belief, there are still hungry and highly educated people in Europe. At my school in Switzerland, it's almost impossible to find the CS lab unoccupied at any given time of the day (and night). There are also lots of bright, hungry people in Eastern Europe willing to work hard to score a big hit.

Similarly, it's possible to find hackers in Europe. DeCSS was written by a Norwegian. Ruby on Rails, to my knowledge, was composed on Danish soil. Germany has the CCC, one of the world's biggest hacking organizations, and c't, the country's mainstream computer mag, makes PC Magazine look like Cosmo Girl.

Still, the vast majority of innovative startups is in the US. If Europe has the hackers, why doesn't it have the startups? I believe that an important factor is that we are missing role models: There are no Googles or Yahoos that originate from Europe – imagine what Silicon Valley would be like had there been no Hewlett-Packard. Another factor is that Europe has a much more risk-averse culture. The YCombinator model works in this environment, too. It needs 10 to 20 daring founders, not tens of thousands.

What infrastructure is required to make young companies blossom in Europe?
  • Talented people
  • Availability of venture capital
  • Potential buyers
While Europe doesn’t have hotspots of concentrated talent such as Silicon Valley or Boston, I'm not worried about being able to get good people. They couldn't even move to the US: Visa restrictions are tough. Inside the EU, people are free to move around, and even in Switzerland, arranging for a working visa is a short and relatively painless process.

A large percentage of the ETH Zurich's CS students are going to end up writing backends for banks, as that seems to be one of few career options. Some of them, I'm sure, would love to work on exciting products with smart, unbureaucratic people. The financial risk is much smaller, too: Remember that in Europe, going to university is practically free. After getting your degree, there is no pressure to pay back loans. You can even continue living in your subsidized dorm room, as long as you can convincingly tell a story about having decided to get another degree, this time in philosophy.

Venture Capital is a different story. Last semester, I attended a Startup-School-like lecture series at ETH and met a couple of local VCs and business angels. Some of these guys were really good, but others has no understanding of the technology and markets or acted as if VC were an extension of the traditional small-company corporate loans business, with similar expectations in terms of collateral and returns. These qualitative differences between VC companies, however, exist in the US as well.

As a startup, it's much easier to get bought if your offices are in Silicon Valley, not Oklahoma City. After all, it's just a one-hour drive to Yahoo headquarters. It's hard to find a spot in Europe with similar qualities, but since everything is one hour away by plane, location should not be a huge problem.

So what's a good spot? I first thought that Eastern Europe – Budapest, Prague, or Tallinn – may be a good place. However, while that's a cheap option, it's hard to imagine getting West Europeans to move east. In addition, you may risk appealing to students in the respective country only. A good choice may be Berlin; very appealing to young people, inexpensive, and in the West. Munich may work better, as it is home to many IT companies – read: potential buyers. Zurich, too, seems like a good spot; the Swiss, in general, seem more enterprising, and both Google and Microsoft have large offices. London, on the other hand, already has a sizeable startup scene and is in many ways virtually indistinguishable from Boston. For my part, I'd pick Zurich or Munich, but that's based on personal preference, not data.

Let's reiterate: We have the talent, and we could arrange for the capital and the buyers. What's missing?

To replicate YCombinator, you need three things: fame, picking talent, and capital.

The last of these is probably the easiest to solve. Per founding round, the YCombinator model only requires a couple of hundred thousand dollars. The founding fathers of the European YCombinator could likely cough up this money by themselves.

More important than deep pockets are the skills of the founding fathers: The ability to judge the promise of business plans and founders. Here, YCombinator is well positioned by bringing on a team that has already been successful at Viaweb. I'd be willing to bet that the partners - Trevor Blackwell, Robert Morris, Jessica Livingston, and Paul Graham - will achieve a higher batting average than the VC industry. How do you assemble a similar team? Even in the US, this would be hard to do.

Getting financed by YCombinator increases a company's chances of succeeding. You'll get more PR – on launch day, it's almost guaranteed your startup's website will be on top of Reddit, and the New York Times will be happy to mention you in their next article. The team also helps with advice on technical and strategy question. It never hurts to get input from startup veterans, as long their ideas aren't forced on you.

Still, of all these factors, Paul Graham's fame is probably hardest to reproduce. His hacker street cred greatly increases the quality and quantity of applications. How do you find a high-profile European who can get hundreds of hungry European CS students to apply for a founder's program? Who could set up a good clone of the Founder's Program and organize a Startup School with high-profile guests? I can't think of anyone who could fill this role.

I'm convinced that a European YCombinator would work. We need a skilled team and the famous guy.


Thanks to Douwe Osinga, Branimir Dolicki, Bálint Miklós, and Fabian Siegel for their insights and for reviewing earlier drafts of this.

Monday, April 24, 2006

Categorizing E-Mail Users

In the past weeks, I've talked to a number of friends about their e-mail habits and problems they have. I've even asked you, dear readers, to for input in an earlier post. It turns out there are a lot of different strategies for dealing with e-mail. I've tried to group them into pairs of opposites. In general, you'll find yourself somewhere between the extremes.

Deleters v. Keepers: Deleters are users who habitually delete e-mails from their inbox. One user reported deleting up to 50% of her incoming e-mail after having answered it. Others keep an even smaller percentage of e-mails: Some ETH students delete almost all of their e-mail after having processed it, a decision which I attribute to privacy concerns and the illusion of strictly limited storage space on student e-mail accounts: As freshmen, they learned that there was a 50 MB storage limit on e-mail. Space is now unlimited, but habits are hard to unlearn. Another motivation for this behavior is "keeping the inbox clean", which some feel is better done by deleting than by archiving or sorting e-mail into folders.

Sorters v. Inboxers: Some users keep almost all their e-mail in their inbox, and have no sophisticated hierarchies to file e-mail into. Others typically file away incoming e-mails – either into a multi-layer folder hierarchy. GMail users like to use archiving as a way to store e-mail without seeing it in their inbox.

Writing filters vs. manual sorting: Many reported being too lazy or dumb to write filters and prefer to sort their e-mail manually, especially when they receive just a few e-mails per day. Very few fine-tune their filters to catch more than just a couple of newsletters.

Folders vs. Search Folders vs. Labels: Most Sorters have a number of folders to put their e-mail into. This can range from a simple work / private categorization to a complex hierarchy of folders for each subproject the user is working on.

GMail introduced the concept of "Labels": While each e-mail ends up in the inbox, multiple tags can be attached to each message. Each label can be used as a view over the users inbox and archive – it is materialized when the user clicks on the label. Users don't seem to apply labels manually, but let filters do the work.

Thunderbird and other mail clients have "Search Folders", which look like normal folders in the user hierarchy, but are actually views on the user's total e-mail showing all messages matching a search query. No one seems to be using these.

Ways of Marking as TODO: Many have found a way to assign priorities to e-mails. One user sorted incoming e-mails into folders for 'now', 'daily', and 'weekly', and processed e-mails according to these intervals. That was, of course, before giving up on this scheme because it was too complicated.

A commenter on the previous post uses the keys 1-5 to assign Mozilla Thunderbird's labels to incoming e-mail.

I used to flag important e-mails in Thunderbird but gave up on it because I'd forget about them when they dropped "below the fold" of the first page. Instead, I now use the same as so many others: keeping messages unread to mark them as still needing attention.

Even this strategy doesn't help e-mails from going unprocessed. One user wrote: "I tend to think that I leave emails that I still need to answer/work on marked as unread. But my 'unread' pile has aggregated to 870 emails at the moment."

Search vs. QuickSearch: GMail users like their full-text search and use it often. Users of Outlook, Thunderbird, and Lotus Notes dread full-text search, because it takes so long to get results. However, the QuickSearch box seems to be quite popular with Thunderbird users.

The problem with all search approaches, however, is coming up with just the right keywords to find the e-mail you want.

There seems to be lots of room for improvement in organizing and viewing e-mail, as well as educating users. I'll be trying out some ideas in the next weeks and will keep you posted.


Thanks to all interviewees for their input plus Fabian Siegel and Bálint Miklós for reviewing drafts of this.

Tuesday, April 11, 2006

A Week in California

I'm sitting in the café of the Borders in Sunnyvale and following the great Silicon Valley tradition of sipping your latte and staring in your laptop. Almost everyone around me is staring into a screen. Are they all working on the next great idea? I doubt it: Some are studying for some sort of Cisco exam, others are reading marketing books.

I just wanted to rehash my last week in California. I met a lot of old friends, visited Google, and managed to sneak in a couple of more touristy activities.

For the first couple of days I stayed with a friend who is a longtime Googler and one of the smartest guys I know. He was kind enough to loan we his old car for the week. I was happy to see that while he had upgraded his lifestyle a bit – a nice apartment in San Francisco, flatscreen TV, etc. – he hadn't gone on a crazy spending spree. In fact, all of Google still seemed very sane. I had feared that the company might have gone off the rocker, but nothing I saw seemed more excessive than what already had existed during my internship in 2004.

So what do you do when you meet all these smart people? You ask them for ideas on your Master's thesis, of course, so that's what I did. I'm still processing all their input but will likely post on it later.

The weather wasn't that great – it was raining for 4 days out of 7 – but on Saturday, the only day that felt like summer, I went on a trip to Point Lobos, a small peninsula south of Monterey. I was accompanied by my former roommate Peter, who I tend to get into long-winded discussions with. We usually cover almost all imaginable aspects – this time, it was everything from Larry Summers to the business models of cell phone providers.

The immensity of the Pacific Ocean keeps surprising me – Lake Zurich just doesn't compare.

I also went to two museums: First, Herzog & de Meuron's new de Young Museum in Golden Gate Park. I was a bit disappointed: While the copper skin of the building is a nice idea, the building looks a bit like a warehouse with an attached observation tower. Also, the architects wasted a great opportunity of using natural light to illuminate the art. I guess they didn't want to get accused of copying from Renzo Piano's fantastic Fondation Beyeler building, just a few miles from their office in Basel, Switzerland.

In addition, I dropped by the SF MOMA, for what was probably my 5th visit: They always have really good special exhibits. This time, they showed a collection of photos from the 1906 Earthquake. Apparently, handheld cameras by Kodak were ubiquitous by the early 1900s, and several thousand of photos documenting the damage exist. I was most impressed by San Francisco in Ruins, an aerial photograph by George R. Lawrence, showing all of downtown SF and the rolling hills in the background. It was taken from a kite.

The earthquake seems to be a hot topic in the Bay Area right now. If the hype went on for long enough, I'm sure it would drive housing prices down. Human fear of the incontrollable and media hype are strong forces.

I kept bumping into people I know. Across from the Apple Store in Palo Alto, I bumped into Marc Tobias, an former coworker from Yahoo Germany. I hadn't seen him in 5 years. It's a small world after all.

Update: Check out these pictures.

Wednesday, April 05, 2006

How do You Organize Your E-Mail?

Last Friday, a long-awaited event finally occurred: After 8 semesters of toiling away at ETH Zurich, I finally had all the credit points to start my working on my MS thesis.

The topic: How can we better prioritize, categorize, filter and search your e-mail?

Thanks to my fantastic professor and advisor, we were able to keep the topic a bit vague and leave lots of room for exploration. I have plenty of ideas, but, of course, I won't give them all away yet.

As this blog has built up a small but respectable audience (30000+ pageviews / month, without counting feeds), I was hoping to get some input from you guys:

How do you organize your e-mail?

  • Do you keep all your e-mail in your inbox, or meticulously sort it into folders?

  • Do you use flagging, labeling or keep mails unread to make sure you get to them?

  • Do you prioritize reading and answering e-mails by person or by topic?

  • Do you have carefully fine-tuned inbox rules like Bill Gates, or do follow the GMail philosophy or "search, don't sort?"
Write a comment and let me know, especially if you're doing something extravagant in any of these categories! Readers have been a bit shy in the past, so here's the threat: If no one responds, I'll set up an online poll.

Getting Real

I'm on a United flight somewhere over the Atlantic Ocean and just finished reading my copy of Getting Real, the latest book by 37signals.

This book is a must-read for web developers. It explains the 37signals philosophy of building web apps: Stay lean, keep it simple, and start with designing the interface.

Nothing in the book will be surprising or new to people who have been following Signal vs. Noise or have read the team's other books. It just packs their already-known ideas into a single entity and explains them in more detail.

I think many of the companies that have sprung up during Web 2.0 have borrowed from the 37 signals philosophy: Lots of these sites have few features and a simple interface.

For some time, I've been wondering: So if the "Getting Real Method" is, actually, special secret sauce, why is 37signals giving out the recipe? Don't they worry about competitors that may spring up, stealing their customers?

After questioning their motives, I came up with the following explanation. The 37signals wanted opinion leadership, and publishing their ideas has helped them achieved that. In addition, they love the sound of their cash register: This post says the book has already generated $120'000 in revenues.

(Side note: While self-publishing as an e-book is highly profitable, I wasn't really comfortable with the format. With a book designed for screen reading in Acrobat, why do the pages have so much whitespace all around?)

While I didn't agree with all their thoughts, reading the book certainly gave me some good ideas for YourGMap. Recently, I've been very busy with school, so there was little time to tweak the product. "Getting Real" was very energizing, so expect some cool changes soon.