Several years into my Silicon Valley startup experience, I’ve gathered a few tidbits that are worth remembering.  Most of them seem obvious on the surface, but as always, the execution is the hard part.  I’ll dive more into how all of this is accomplished in future posts.

1. Focus.

You have limited resources. The most costly mistake any company can make is focusing on the wrong (or too many) things. When you notice you’re on the wrong track, pivot immediately. The money and effort you’ve put into the wrongthing are not coming back. That’s what economists call a sunk cost - treat it like one. Google just demonstrated the importance of this with wave.  They have money to burn, so they can afford to take a few product risks. Often startups can’t.  This is also important from a product design perspective. Many of the most successful products, especially those oriented toward consumers, are neat and simple. Don’t spend time building complicated features that only a small fraction of your users will use.  That only wastes your time and dilutes your product.

2. Deliver.

Take a little biblical wisdom from Ecclesiastes: “A living dog is better than a dead lion.” We’d all rather have a lion, but we can’t all be as bad ass as Mike Tyson. If we’re going to make it though, we have to work our way up to that level of greatness.  It’s way better to deliver a small-but-good tool that successfully does its job than to be overextended and running out of money when that pesky 2nd 90% of the product needs doing.

3. Iterate.

Often, you won’t know how the product will be received.  Get potential customers involved early. Take their needs into consideration, but don’t let them drive you completely. “Dogfood” the product internally so you have a fair sense of it’s quality. Release it to the wild so you can conquer this unknown as quickly as possible.  ”Done” is a mythical word in most product development.  It truly is a cycle.  Your customer’s needs will grow and evolve, your product should too.

4. Jockey for position. 

Depending on what your product is, it’s always advantageous to have a solid position as a leader in your market.  ”Leader” is a pretty flexible word in this context.  Maybe you were the first to market, or you have the most users, or the best customer service.  Whatever it is, know it and stress it in the relevant media.  Whenever something happens in your sector, you want to have your name mentioned as the people to use/beat.  This is huge for fundraising and selling purposes, but it’s hard to manage in the long term without a great product.

5. Don’t get stuck in the local maximum.

This applies to so many things within a startup beyond AB testing.  It’s easy to be sucked into short term contracts that bring revenue but don’t move the needle closer to your goal.  It’s easy to save all your money and not hire the talent you need to be great.  It’s easy to maintain the status quo when a project really needs new leadership.  Keep your eye on the end goal and your mind on systematically removing the obstacles in your way.

It’s a new year in a new decade and that means it’s definitely time for some reflection.

2009 was my first full year in San Francisco.  Looking back, it seems like it went by in a flash.  I ran my first marathon, made tons of friends, found myself a wonderful new girlfriend, inherited the senior engineering position at YouNoodle, moved into a new neighborhood, celebrated my 23rd birthday and generally had a blast.

It seems like the longer I stay in the bay area the more people come to join me.  I can see how the “hella” locals might take for granted what a wonderful place northern California is, but I certainly don’t.  I’ve taken serious pleasure in exploring the bay area.  Camping at Trefethen, running courses every Saturday morning with Team in Training, hiking through the wilderness in the many state parks, hanging out at ocean/baker beaches, bumming around the Stanford campus, and dining at some of San Francisco’s finest restaurants (which are stellar).  I’m sure more people will end up out here before everyone settles down.

2009 was a year of foundation building professionally and socially.  In 2010 I’d like to continue both.  Socially with a focus on finding a “sidekick.”  Professionally I want to consolidate my position at YN, but I also want to spend more time learning.  I want to establish a working knowledge of statistics and statistical computing.  I want to drive less (possibly bike instead) and travel more.  There’s more to be done, but I think that’s enough resolutions.  Under promise and over deliver, as they say.

May the tens be even better than the ‘oughts.

This year, I received my domain, mikeihbe.com for my birthday.

I wasn’t really sure what to do with it, so it will probably be an evolving project.  I definitely wanted it to reflect who I am and what I do, but also have the potential to open new opportunities.

Whenever I start a personal project I typically try to learn something new while I’m doing it.  Lately, I’ve been playing around with Rich Hickey’s Clojure, and this is a great chance to build something real with it.  There’s tons of new aspects to this project for me: building something non-academic with a Lisp, a focus on functional programming, learning a new language, and more experience with the JVM.

I used java for my primary project while I was at Amazon, but that was really my only experience with it.  Their highly evolved build tools left me unprepared to properly build a java project from scratch, so I opted to develop/build in the Clojure Eclipse extention Enclojure.  Enclojure is in alpha, and definitely feels like it, but they have made some decent first steps.  Those first steps did not lead to a reasonable method of deployment, however, so I hacked together (it’s really ugly) a deployment script that seems to work.

While building the application (the code is here), I borrowed significantly from a bunch of other open source contributors:

http://clojure.org
http://github.com/weavejester/compojure
http://jonraasch.com/blog/a-simple-jquery-slideshow
http://www.cssplay.co.uk/menus/flyout_horizontal.html
http://github.com/drewr/postal

So I want to give a big “Thanks!” to them.  Though the code is available, and is a fair example of a Compojure application, there are a few things I’d like to improve: the amount of purely functional code, a greater appreciation within the application for Java’s pre-compiled nature, and how the application is built and deployed (probably using leiningen).  And of course, the aesthetics.  After that, I’ll start adding more nifty features.

Also, I think Compojure will undergo some pretty serious changes too.  It’s a very lightweight framework, but it still could use some polish (and documentation…).  It looks like it’s also going to be separated into constituent parts, which is a good thing, but probably means significant application changes in my future.

Regardless, it works, and it’s successfully deployed on a virtual Linode server.  I’ve integrated Google email for the domain (it wont forward a catchall, which is unfortunate, but a well documented bug, so I used POP instead).

So there you have it.  Unveiling  ——>Mikeihbe.com<——

RT @nallyo “Periodic table of beer styles: http://bit.ly/4vaIlp.” I’m a fan of columns 4, 6, 9 and 20.

Having gone to my local pub’s trivia night, I have determined that I know nothing trivial.

Cybernetic hand successfully integrated with human nervous system: http://news.cnet.com/8301-17938_105-10408139-1.html

USPTO takes first step, admits they have a problem: http://tr.im/HCxP

Quantum propulsion proposed in MIT tech review: http://www.technologyreview.com/blog/arxiv/24499/

Great book, well written, humorous author, and good explanations of some familiar econ 101 concepts. Well worth the read.

The author’s explicit purpose is to make the economic sciences interesting to the masses, because most people take Econ101 and their eyes glaze over at all the charts and then they think econ isn’t interesting, but really it is. And he goes through how governments can use economics to incentivise (or disincentivise, as the case may be) certain behaviors to try to overcome macro obstacles like global warming. Of course, he also covers the basics of money handling from the perspective of the Fed, all the normal 101 stuff too, but he skips the math and keeps things light. There’s a nice pitch in there supporting globalization too, for anyone that might still be against it, I recommend you pick this up for that alone.