Friday, May 1, 2009

Lean Startup webcast post-game

Wow, what an incredible turnout for webcast this morning, right on the heels of a phenomenal crowd at startup2startup last night. It's been a great twelve hours for the lean startup!

We'll have a full audio recording posted soon, but I did want to share the slides with everyone. Much of the content is shared with the Web 2.0 Expo talk, but there are a few new slides and in the discussion/Q&A we were able to go into much more implementation detail. One big idea that I haven't had a chance to write about on the blog yet is answering the question "What is a startup?" Here's the working definition I've started using:

A startup is a human institution designed to deliver a new product or service under conditions of extreme uncertainty.

In a future post, I'll attempt to unpack this definition in detail. For now, I just want to call attention to the idea that it has nothing to do with size of company, sector of the economy, or industry.



And now for some instant twitter reactions:
KentBeck: @ericries thank you for the insights. "unknown problem" was an eye opener for me. i'm running junit max according to lean startup principles
OK, so I'm showing off. One of my personal heroes, Kent Beck, the creator of Extreme Programming and a truly wonderful writer was logged in and twittering, too. What a thrill.
OReillyUG: For more info on The Lean Startup by Eric Ries, register for our upcoming courses in SF May 29 and June 18 http://training.oreilly.com/theleanstartup/ #leanstartup
Very excited to announce that I've teamed up with O'Reilly to produce the upcoming Lean Startup Workshop as part of their Master Class series. As a result, we should be able to accommodate more people and more venues, maybe even more cities, if there is interest.

Now for some real feedback:
@rotkapchen #leanstartup Love this: Need both problem team & solution team working simultan, exchanging hypotheses, findings & solutions.
I'm continuing to push this idea, that departments can incentivize sub-optimal behavior. What makes a department (or individual) "more efficient" doesn't necessarily advance the company's goals.
farazq: Awesome session @ericries. LIKED: Y startups fail, cont deployment vs waterfall vs agile, small batches & learn fr biz metrics #leanstartup
As concise a summary as I've ever seen.
hwijaya: #leanstartup It's not the release fast dat matter. It's getting the evident dat u're on the right track (wat u learn frm making the release)

katrynharris: @ericries: "go as fast as we can work reliably to make progress & no faster" - absolutely & thanks for a super webcast! #leanstartup

Really trying to emphasize the importance of speed - but speed with regard to actual progress. That's what gives startups their disruptive innovation edge. Going "too fast" is not actually helpful. Learning to tell the difference is the hard part, and creating processes that act as speed regulators is the payoff.
clynetic: #leanstartup - All are ready to do this: Whether starting new company, have startup but need to iterate faster, or established company.

kylemaxwell: Behind every technical problem lies a human problem. #leanstartup
Always glad to see people pick up on these themes. It's never too late, you can always get started, and since all problems are human problems, there's no hiding behind a technical excuse.

Of course, not everybody had a great time.
wogsland: Still not understanding the "lean" aspect. Techniques sound fairly expensive to do correctly. #leanstartup
This is an objection I hear regularly, so I'm grateful to have the chance to address it. Lean is not about cheap, it's about speed. All lean transformation techniques, including the ones I advocate, involve making some part of the organization "less efficient" in order to speed up the whole. So they require investments to pay off. However, the level of waste in most organizations is so high that the investments pay off immediately, and therefore create new resources that can be reinvested in further improvements.

For example, it takes a few days of effort to build a simple split-testing system along the lines I advocate. Some companies tell me they "cannot afford" to do it. But if that split-testing system allows you avoid building just one major feature that would have been a waste of time, it's paid for itself many times over. If you don't have the confidence or ability to make trade-offs like that, you can't get lean.
kylemaxwell: you have known knowns, and known unknowns, and unknown unknowns. #leanstartup #rumsfeld
There's no way to avoid this Rumsfeld quote. Then again, "just because you're paranoid doesn't mean they're not out to get you." What can I say? He had a point.
jimmurphy: @ericries bug is your waterfall slide: Problem: Know Solution: *Known* #leanstartup
Thanks! This is fixed in the posted slides.
wogsland: uck, buzzword vomit #leanstartup
Well, I don't know what to say to this. I welcome suggestions for ways to get the same content across with fewer buzzwords. I don't think I can be very helpful without a discussion of the theory behind the lean startup, but I also don't know how else to teach that content. Please feel free to post your thoughts.


And then there's some questions that came up during the webcast that we didn't have time to discuss. I covered a few of them on twitter already:

  • Q: "Do you reccomend removing features that you've launched but don't movethe needle on engagement or revenue?" A: yes, absolutely.

  • Q: Does the Lean Startup framework work efficiently within a company of 2-3 people (think small)?" A: yes, yes, yes

  • Q: Does the Lean Startup framework work efficiently within a company of 2-3 people (think small)?" A: yes, yes, yes

  • Q: "Seems like the biggest issue is changing people's mentality. How did you deal with that on a new recruit (specificaly seasoned ones)?"

    A: have them deploy code to production on their first day as an employee. made a big impression.

  • Q "In CD you spend a lot of time on phone with users. How does this map to the web? Surveys? Trial and Error + Analytics? What else?"

    A: honestly, all of the above. Phone/in-person is still critical, but have to remember that is for idea generation, not idea validation

  • Good Q on webcast: "when measuring the results of your test, how do you avoid the post hocergo propter hoc fallacy?" A: you must split-test
If you had another questions you wanted to see answered (or do after you see the slides or audio), please post them here! I'll do my best to answer.

Lastly, I was really excited to announce on the webcast that we've opened up two dates for the Lean Startup Workshop - now that it's in O'Reilly's Master Class series, I have a lot more capacity to do these events. If you're interested, you can find out more here.


Reblog this post [with Zemanta]

5 comments:

  1. You mentioned as many as 50 deployments a day to production after successful local continuous integration.

    With so many deployments how do you deal with database schema changes and data conversion for these schema changes?

    OODBs and ORMs like NHibernate are great but you don't want users using the system while you're making database schema changes. Don't you need a maintenance window?

    ReplyDelete
  2. What's the role of a Product Owner in a Lean Startup (mentioned as "Clear Product Owner" practice in your slides)?

    BTW great webinar, thanks for sharing your work :-)

    ReplyDelete
  3. I work for a startup and I always try to do all the things you suggest - automatic deployment, continuous integration, A/B tests... - but I'm always short in hands. How many people - coders - do you think a small startup with a 6mm pages/month site should have?

    ReplyDelete
  4. Recording is now available as a YouTube video:

    http://www.youtube.com/watch?v=EnaLQiQL9ec

    ReplyDelete
  5. One problem I have with releasing as soon as possible, is that we are releasing in to a space with competitors. We have certain advantages that we are confident will differentiate us, but these may not be apparent if we just released as soon as the basic functionality of our app was ready.

    The danger then is that you alert your competitors to your existance, they can tell the advantages you have and because of their greater scale (despite their lack of speed) they start developing to make up ground and by the time your advantage is apparent, your competitor has closed the gap.

    As you quoted, "Just because you're paranoid, doesn't mean they're not out to get you"!

    ReplyDelete