10 Startup Facts Panel at the University of Washington

If you’re in Seattle and want to learn more about what it’s like to work at a startup, you should attend a panel I’m in on May 2nd at 3:30 in the UW CSE building.  Glenn Kelman (Redfin CEO), Christophe Bisciglia (WibiData CEO), Oren Etzioni (UW CSE professor and entrepreneur/investor), and Dan Weld (UW CSE professor and entrepreneur/investor) will be on the panel with me.  More details about location/time here.  Below I’ll tell the story about why I’m organizing the panel.

I’ve been writing this blog for almost 5 years now, and BY FAR the most popular post I’ve ever had was 10 Facts About Working at a Startup vs. a Big Company.  About 30k people have read the post, and I’ve received a huge amount of interaction in comments, Twitter, etc.  The whole process has been a wild experience for me.  But most importantly, I’ve learned that there needs to be a lot more education about startups.  People don’t really know how amazing working at a startup can be.

Startup education and evangelism is especially lacking at the University of Washington’s CSE program.  The CS program at UW is world-class.  The quality of graduating students is incredibly high.  Yet the HUGE majority of them take jobs at mega-large companies such as Google, Microsoft, and Amazon.  I get so sad envisioning these awesome engineers working in a cubicle at some gigantic company, where they may make some small impact if they get lucky with a good team.

I originally wrote the “10 facts” blog post to try and show my friends why they should work at a startup, in particular at a startup I’m doing right now :).  My original blog post inspired me to put together the panel I mentioned in the intro to this blog post.  My hope is that Glenn, Christophe, Dan, Oren, and myself can inspire the UW CSE community to consider working at a startup.

I hope to see you on May 2nd!  Bring your friends, invite others that may be interested.  Even if we don’t convince you to work at a startup, at least you’ll understand what they’re all about.  See here for more details about the panel — location, time, etc.

Here We Go Again – Introducing MemCachier

The last nine months have been a practice in self discovery — I’ve been keeping myself busy trying out startup ideas, consulting, and talking to investors and other entrepreneurs.  I’m pleased to say that the search has been postponed for now.  I’m working on MemCachier, a better memcache for Heroku.  And I’m working on it for the long run.  We’re going to turn MemCachier into a thriving business, and I’m stoked for the opportunity!

MemCachier provides a caching service to web developers that gives them an easier, cheaper way to scale their website.  Currently it’s available as an addon in Heroku.  Nerd stuff, I know.  But the reality is that I’m a nerd and I’m good at nerd stuff.  I couldn’t be more excited.  I’m working with Amit and his lab mate, David — the three of us make a great founding team.

Amit started building MemCachier about six months ago.  He launched it in private, then public beta on Heroku.  And earlier this week we launched into general availability.  Times are very exciting, and I’m stoked to have been invited by Amit to work on this with him.  My primary responsibilities will be on the business/company side of things, although I expect I’ll write a little code here and there as well.

MemCachier isn’t an entirely new concept — memcached has been around for a long time for websites who manage their own servers.  But memcached hasn’t yet been offered as an easy, managed solution in the cloud.  Enter MemCachier: the primary goal is to make caching easy for developers.

Anyway, I thought I’d update you guys with what I’ve been working on.  If any of you use a cache in Heroku, EC2, or anywhere, I’d LOVE to get your feedback.

What I Learned from Making a Fake Product

I originally envisioned BreakStreak helping me and my friends be better about getting into a routine such as cooking more, working out, running, etc.  Rather than build a full fledged product, with a mobile app, complex reminders and check ins, and payment processing, I thought it’d be great to run an experiment first.  BreakStreak.com makes its users think they’re using a real product, complete with all sorts of fancy features.  But really they’re using a totally empty, fake product that’s tracking their usages in the background.  Cruel, I know, but this experiment let me disprove my original hypothesis that BreakStreak is a good implementation of my original vision.

The Product

BreakStreak is essentially a wizard that takes the user through a few steps.  The user is asked what routine they want to get into, along with a desired per-week frequency.  Then they’re prompted with a description of the product, and a few mechanisms for tuning their incentives — reminders, donations to charity if they fail, and guilt from friends.  Finally, they’re asked to create an account.  After creating an account, they’re directed to a page that explains this experiment, apologizing that BreakStreak isn’t actually a real service.

The Experiment

I wanted to see how far along in the wizard customers would get.  So I built the wizard such that only the user’s progress was stored, not the actual data associated with their routine.  For every single user I know exactly what part of the wizard they stopped at, what data fields they filled out (but not the values), and when they gave up with the wizard.

The Results

The results were very surprising to me.  Of 271 page visits, only 4 (1.5%) actually created a real routine and a real account (another 8 created test accounts).  Keep in mind, too, that those 271 page visits came from friends who found BreakStreak through my Facebook and Twitter posts — they weren’t a fair representation of the average consumer.

Furthermore, of those 271 page visits, 119 (43%) actually filled out the name of a routine (running, gym, etc.).  And of those 119, many were test routines such as “make sexy time,” “test,” and “masturbate.”  I built a little dashboard that models the funnel and the choice of incentive:

(You’ll notice in the chart that 12, not 4 finished.  8 of those finishes were tests, not real routines.)


My BreakStreak experiment proves two things: people want to be better about getting into a routine, and BreakStreak isn’t the solution.  The 43% conversion rate when prompted for a routine is very high.  However, the 1.5% final conversion rate shows that the product has fallen short of its original vision.

I built the BreakStreak experiment in about two days, whereas the real app would have taken me many weeks.  I was able to save tons of time otherwise spent on development by running an incredibly meaningful survey.  Instead of asking my users if they’d pay for something, I gave them a product experience that was real and let the results dictate their buying patterns.

I’m incredibly happy that I made the decision to run this experiment instead of wasting my time building the whole thing.  I’m stoked about the results, even with with (mostly) negative outcome.  Now I can rethink the original BreakStreak intention and either build a new/improved product or move on to my next idea.  (For now I’m choosing the latter.)

I’ve always been a fan of the Lean Startup movement, and this experiment drives home the importance of experimentation, at least for someone like me who’s relatively inexperienced building consumer apps.  Not to mention my sub-par design skills ;).

Introducing BreakStreak

I’d like to introduce my newest project: BreakStreak.  BreakStreak helps you get into a routine.  It does this in a few key ways:

  1. You sign up to perform a routine, say cooking or working out, along with a specified number per week.  Every time you accomplish your goal, progress gets added to your “streak.”  Eventually your streak will get so long that you won’t ever want to break it and start over.
  2. You can create incentives for keeping your streak — although incentives are optional.  For example, if you break your streak, you can have a SMS and/or email be sent to friends.  You won’t break the streak if you know your friends will find out.  You can also donate $1 to charity for each activity you don’t perform for a given routine.  30% of that $1 is kept by BreakStreak to help operate the website.  Again, these incentives are optional.

I built BreakStreak because there are so many things I wish I spent more time doing, in particular cooking and meditating.  And many of my friends have the same problem, too.  BreakStreak is designed to help us set a goal and stick to it.  My hope is that BreakStreak can be a common and useful tool for helping each of us hack our lives better, with the sole goal of making us happier and more proud of our accomplishments.  A friend of mine, Brian, has helped me with much of the brainstorming and product ideas.

Give it a try at breakstreak.com!  And let me know if you have any thoughts or feedback.