• Team-driven investors primarily care about the founders themselves: their relationship, what they’ve accomplished prior to starting the company, and why they’re the right mix to tackle the problem at hand. The incubator guys (e.g. Y Combinator, AngelPad) obsess a lot over this stuff, and they’re as well-positioned as anyone to understand, and avoid, the common causes of early death for companies.
• Problem/product-driven investors want to understand the urgency of the problem (“Is this a hair-on-fire problem?”), and how the team intends to solve it. Product orientation seems common among angel investors who made their money as engineers at Zynga/Facebook/Google. These guys believe (rightly in my opinion) that if the problem is sufficiently compelling, and customers get enough value from the solution, that investment and revenue will surely follow.
• Market-driven investors care about the numbers: how many customers, how much revenue from each of them, gross profit, acquisition cost, and churn. Market-driven investors tend to come from finance and banking, and play a large role at later-stage (Series B) funds, where the key question is scalability, rather than founder issues or technology risk.

When trying to raise seed capital, focus on the first two. Most angels and early-stage investors I’ve met believe startups fail because of team and product/market fit issues; they’re far more tolerant of market sizing issues, than, say, lack of a compelling problem, or bad founder chemistry.

I used to think Silicon Valley was crazy for betting on companies without business models, but now I’m not so sure.

(Note: It occurred to me, as I was writing this, that the investors’ concerns form a sort of blueprint for how to start a company. Start with good people, then find a problem, then concentrate on a market — in that order.)

1. Quit the company: burn professional bridges, and leave a black mark on your resume
2. Transfer to another group: kind of like quitting, but looks less bad because you’re at the same company, or
3. Stick it out: see whether things improve while collecting a paycheck and letting your career idle

Faced with this dilemma at Microsoft two years ago, I chose option #1 (quitting). When the ship started to sink, I didn’t pray, I jumped.

An individual’s choice seems highly linked to his risk appetite. Regardless of the long-term consequences, quitting is definitely riskier in the short run.

Idealists tell us to “follow our hearts”. Following one’s heart is romantic, but it doesn’t go over well with venture capitalists, or the parents of significant others.

Viewed one way, a watch is nothing more than a device for telling time. But this misses the point: go to any mall or jewelery store, and look inside the case: you’ll see many watches on display, of varying quality, and vastly different prices. From $5 to $5000, all watches share a basic feature: a set of hands, or a digital display, for indicating the time.

There’s more to product design than checking off lists of features. Other things to consider:

• What’s the overall feel of the product: cluttered and busy, or minimal and clean?
• Does it work?
• Is it easy to use?
• Is it fun?

Hard to measure? Absolutely, but so are reputation, brand value, and love–and those are pretty important too, no?

PCHRs increase record portability by (1) standardizing the way medical data is stored, and (2) putting it all in one place. Centralization comes with the additional benefit that the patient has full control over who can access his records.

Of course, centralization begets the question of who stores the records. Google, Microsoft, and Dossia are vying for pole position, with three different approaches to the storage and retrieval of health data.

This post recounts my experience getting started with HealthVault, Microsoft’s offering.

About me

I’m pretty average as software developers go: mid-twenties, primary computer is a Mac. I use git for revision control, unfuddle for project tracking, and mostly develop in Rails. I’ve been a professional developer for 3-4 years.

Note: A friend at Microsoft informed me that the SDK error message was due to a missing DLL. Installing the .NET Framework 2.0 Redistributable seems to have fixed the problem (25 June update).

Getting started

I clicked on the first link, and was taken to a pretty informative page on MSDN about getting stated with HealthVault. Before I begin, I needed to do five things:

I created an account to get familiar with the system. Easy enough. I also noted that the “Supported Browsers” noted support for “IE (latest two versions)”, Firefox, and Safari. Not a bad showing, especially considering two of them aren’t Microsoft browsers.

The fail started when I tried to download the SDK:

Hmm, “Double-click healthvaultsdk.exe to start the installer”? Yeah, not gonna work, I use a Mac. Time to find a workaround.

Documentation?

I poked around on MSDN for a while, hoping to find code samples, or some kind of walk-through for getting started on the platform. The best thing I could find was the Microsoft Health Schemas, a set of reference documents defining the methods and “things” available in the API.

I did manage to find some code, but it was in a .NET language (I think C#), and it just called a bunch of functions on classes defined in the SDK; it wasn’t instructive on what was going on behind the scenes. Since I was trying to develop my app in Rails, I was hoping to discover how things actually worked beyond seeing some opaque driver code.

The code to the SDK itself wasn’t posted or browsable anywhere on the site, nor could I find documentation on what the SDK methods did. I decided I needed to find some way to crack that damn SDK. Time to break out vmware.

Virtualizing Windows

Per the guide: The recommended development environment for HealthVault applications is Windows XP with Service Pack 2 and Visual Studio .NET 2005. Some of the security settings in this guide are specific to this environment.

40 minutes later, I had a working version of Windows XP installed, with the SDK executable downloaded. I tried to run it, but I got an error:

Really guys, this isn’t a very descriptive error message! I guess I’ll track down the error in some MSDN forum? Maybe?

Coding on Rails

Microsoft doesn’t officially support development on Rails, but there are two projects that claim to make it easier: HealthVaultRubyLib, and RubyHealthVault.

HealthVaultRubyLib

HealthVaultRubyLib is a standalone component hosted on Microsoft’s Codeplex. It isn’t documented very well, and following the link to the “live rails application using this library” gives the following:

RubyHealthVault

RubyHealthVault is a Ruby gem, but it hasn’t been updated since June 7, 2008. With no recent activity, zero associated developers, and no documentation, it appears dead.

Conclusion

As an independent Rails developer, I’m having a very hard time getting started with HealthVault. If Microsoft wants to encourage development on HealthVault, they have to do a better job of making it easy to get started on the platform. The key factor missing from the experience is knowledge of the audience. As developers, we don’t need step-by-step directions on how to download and install SDKs—and there’s no reason these can’t be packaged as ZIP files. (I used to be a Setup PM at Microsoft; get LCA approval to have the EULA displayed on the download page, don’t make the user “install” the SDK.)

Also, write more example code, and explain how the system works at a higher level. The current examples are too heavily tied to .NET, and don’t illustrate the principles of how the system works. Google Health has much better supporting documentation that presents its content in a language-neutral, principles-first way; I much prefer this style of explanation, versus a low-level step-by-step tutorial that will be obsolete by the next release of the SDK.

The rate of information exchange in a place like Silicon Valley can be hard for outsiders to fathom. On a recent trip, I serendipitously ran into a Berkeley Ph.D. while waiting for the Caltrain. After a few minutes of conversation, we discovered that he’d known my academic adviser while the two were students at Berkeley. We got into a lengthy, and fascinating, discussion of database semantics which continued on the train, all the way from San Jose to San Francisco. I can’t imagine this kind of thing happens routinely in many other places.

However, Silicon Valley’s hyper-connected, homogeneous population breeds the worst kind of groupthink. So many smart people clustered together, and the best business ideas they can come up with is to make iPhone games? Foursquare just made waves by passing the 2-million-user mark, but let’s not forget, that’s less than 1% of the US population. Adrian Holovaty’s remark a while ago on Signal vs. Noise caught my attention:

Another big plus for working from a non-web-2.0-echo-chamber city is the diversity of people and professions. My friends here in Chicago are furniture makers, architecture students, journalists, professional musicians, philanthropists, cops, lawyers.

While I’d agree that it’d be a great motivator to live in a place saturated with other Web developers, it’s also helpful, from a product focus standpoint, to be around people who don’t know what the f—k RSS is.

• I’m building an internet service with Dr. Lyle Berkowitz and Aleem Zafar. The service, which we haven’t yet named, will help medical care coordinators manage their caseload. We’re actively researching our customer base and haven’t yet incorporated the company, but I’ll be sure to post details as they develop.
• I’m contracting through Prefiat, my new company created for this purpose. We provide technology implementation services—hardware, software, and websites—to customers of all sizes. The company’s primary place of business is Office Nomads, located at 1617 Boylston Ave in Seattle.
• I’m organizing Barcamp Seattle, a technology/lifestyle/2.0 event to be held in June.
• I’m doing pro bono math tutoring to help struggling high school students succeed.

Suffice to say, I’m definitely keeping busy. There are so many great things to do in the world, and so little time; I can hardly imagine ever being bored.

In only 52 seconds, we see: automatic search completion, spelling correction (in French!), maps, translation, definitions, flight times, and consistently high-quality search results. All useful features, tied together with a compelling storyline about a guy falling in love in Paris.

Photo sharing, eBooks, email, web browsing, movies, and calendaring, in just 30 seconds.

AT&T vs. Verizon 3G coverage (image courtesy of Engadget). AT&T wanted this ad taken down so badly, they actually sued Verizon, claiming it misrepresented Verizon’s coverage (it didn’t).

“The New Busy”. What’s the product? Does it make me busier? Do I even want to be busy? (Personally, I consider busy a four-letter word.)

Good advertising explains how a product satisfies a customer need. The first two ads are powerful because they show nothing but the product. The Verizon ad visually depicts the carrier’s coverage, a major concern to high-end data customers, who tend to be quite mobile.

The Hotmail ad, by comparison, feels like a beer / soda commercial; it tries to be catchy because there’s nothing unique or special about the product. I can’t believe Hotmail is really that bad of a product. This ad campaign is terrible. Someone should get fired for this.

This job won’t pay cash, so I decided to start another company to do consulting and contracting while I build the startup: Prefiat LLC. The company is a boutique hardware/software agency and is available for hire immediately. I just put up the website today, and I have to say, I’m starting not to suck as badly as I used to at web design.

I received the completed boards from the cerebral palsy project from BatchPCB today. After soldering down all the components, I plugged it in, and it worked right away. As an unexpected bonus, I got eight free copies of the board—I’m guessing they had some free space on the production run, and decided to stick a few small ones (mine) in along the edges.

Next comes experiments with switching and software. I spoke about this project at Ignite last Thursday; I’ll post the talk here when it’s available. In the mean time, check out the media section for the slides.

To the extent that one can control who one does business with, there’s a lot of wisdom in that. The average American is overweight, and in debt. Most marriages end in divorce. Entire industries (video rental and credit cards, for example) derive significant revenue from peoples’ inability to follow through on their commitments.

My dad’s not an elitist, he just doesn’t take people at their word, unless he has reason to.

It’s easy for my dad to make a statement like that, though, because he’s spent the majority of his working life dealing with a small number of highly-qualified professional colleagues inside corporate America. When you’re on the outside, selling, things get a lot murkier. To what authority can one appeal, when a customer is being unreasonable? What to do, when someone refuses to pay? And how can one know another’s character, prior to doing business with them?

Doing business with great people is a privilege, something to fight for, not a right.