Live blogging Startup Lessons Learned Conference #sllconf

Good breakfast done. Eric Reis up on stageā€¦.

…great opening - Eric Reis dismantling all of the myths and misconceptions around lean startup

  • lean startup is about learning
  • lean startup doesn’t mean small
  • lean startup doesn’t mean cheap
  • lean startup doesn’t mean bootstrapped
  • lean startup does mean disruptive

…I suspect we’re going to see lots of tweets mixing the hash tags #kentbeck #sllconf and #goats

  • he likes scratching goats
  • the build-measure-learn cycle is the wrong way around
  • assumptions-metrics-experiment is the cycle in reverse
  • Team Vision & Discipline (and optimizing for team performance not individual performance) > Tools > Process
  • When it comes to measuring progress Validated Learning > working software > comprehensive documentation
  • Initiating change > Responding to change > Following a plan

case study time

Wired Reach is talking about continuous deployment… Why did head of engineering choose now for bathroom break?

  • moved to continuous deployment
  • release every time they commit code
  • commit » test » deploy » monitor
  • average deployment 25 lines of code
  • takes twenty minutes to release not all day
  • shorten QAcycle
  • code in small l batch sizes
  • deploy manually at first,then automate
  • functional tests better than unit tests
  • but watch the release cycle time - multiple qa boxes and isolated sandboxes - keep cycle time below 30mins
  • only deploy if ALL tests pass - it’s your only line of defense
  • requires solid automated update systems -both push & pull
  • don’t be a feature pusher 80% existing features 20% new features
  • constrain the features pipeline: backlog » in progress » done » validated learning (was the feature any god?)
  • close the loop with validated learning: qualitative first (contact the customer) then quantitative (verify with data)

Afraid we miss out on liveblogging notes from the Grockit agile case study and most of the IMVU story as the wifi was not working but it is up again now. Customer Development vs Agile tension was the focus of Grockit and there’s great notes on IMVU on the Startup Lessons Learned blog.

  • IMVU: hack week (work on any project you choose) works well only when people can only work on one product and time and must choose to ship it or kill it
  • culture of learning encourages accepting responsibility and accepting failure as okay so long as you learn

Cool. It’s time to hear Dave McClure in person after working through his slide decks and reading his blog. Everyone serious about a Lean Startup approach should go to Slideshare if you haven’t already and work through his slide decks. Solid ideas and methodologies.

  • lean design is about discovering an emotional reaction and a testable response among users and find a way to convert that into their happhiness and your business value.
  • What is design? Up first Andrew Chen (again read his blog - it’s awesome)
  • Slideshare CEO: sometimes our visual design can be crappy (not as crappy as eBay) but we think on design at a very deep level - focus on overall vision - as you optimize you’ll see the visual design suffer so the vision is important
  • Zynga guy: what is good design in games? Quotes Jobs “great design is about how things work, not how things look”
  • the bias is often to focus on things that are easy to measure and quick to measure
  • it’s an art not a science - too much small testing and focusing on metrics based optimization risks arriving at a local maximum
  • interaction design comes first then leverage visual design to that end - visual design has a big impact on emotional response to product
  • Dave McClure: “with Mint we were late to market so we built a blog and did SEM testing to measure results based on what we were blogging about and outranked competitors in some categories before even having a product”
  • qualitative doesn’t mean not measurable or unscientific - but when focus grouping random users may be less helpful than specific user profiles - e.g. facebook game users rather than general Internet users is far more helpful for Zynga
  • as you narrow down the screen of the users you can start really understanding who is going to like the product when it’s launched - iterate about the market and marketing as much as about the product

Time for lunch….

Randy Komisar is up on stage being interviewed and he is a great speaker. I do wish it wasn’t so impossible to get into Kleiner Perkins - he just comes across as one of those VCs you want to work with.

  • dash boarding - how do you want that information presented to you?
  • not a template - culturally relevant to your company - as relevant internally as the board
  • What are the alternatives if it doesn’t come back as you hope?
  • Analogs and antilogs - what are you solving and who for? Analogs are companies who have tried to answer similar leap of faith questions successfully and antilogs are companies that have tried to tackle similar problems and failed.
  • conversation ended nicely - audience question asked Randy, as he has so many answers what questions does he still have… Randy answered via anecdote of asking the same question of a Bhutanese monk who answered “why aren’t people more kind to each other?”. Yup, wonder that myself every day.

Aardvark guys talking about their process. Interesting…

  • you will usually fail at everything you want to succeed at so: minimize risk » maximize trials » be patient
  • conception - they spent six months playing with ideas - picked problems, built prototypes. And abandoned
  • rekkit, ninjapa, The Webb, web macros, Internet button (none of them good ideas that users wanted)
  • implementation took a year - raised seed funding, did “wizard od oz” testing, recruited core team
  • assume you are wrong
  • make continuousimprovement your goal
  • Experiment with process as much as product
  • hire and indoctrinate to support the process

Hiten Shah of Kissmetrics talks about the pivot:

  • it’s okay to pivot - everyone should be pivoting
  • YouTube did a customer need pivot - from dating site to video sharing
  • Paypal did a feature pivot - from palm-palm money beaming to emailing money
  • Facebook did a platform pivot - open up so others can build on top of it
  • Kissmetrics pivoted three times to get to the current service

The brilliant Steve blank is talking, after a break for ice cream and to see our lead developer demo our latest product to the team - it is awesome!

  • Customer development lean startup is the first business methodology that has been crowd sourced.
  • corporate VPs of sales fill up all available space in a room and have the lowest handicap of anyone you know - could have another career as pro golfer - Nothing to do with a startup
  • 90% of founders never see the promised land - founders very rarely run large companies
  • we do customer development, continuous learning, self organizing teams, pivots
  • startups are a search for a business model
  • this search is called customer development
  • using processes to find profits
  • you fail if you remain a startup

Steve recounted the story of Billy Durant and Alfred P. Sloan (find it via Google) to segway into how Business School churns out Alfred Sloans not entrepreneurs, not Billy Durants. entrepreneur school will emerge. Personally I am not sure you can teach entrepreneurial spirit - I was doing entrepreneurial activities aged 11 - but you can teach winning methodologies and that is what Customer Development and Lean Startup is all about. And the Durant School merchandise is available on Cafe Press right now!

One of my favorite startups out there is Dropbox - we use their service. Their CEO is up talking. They have a fantastic product. The startup I founded in 1999 that brought me to Silicon Valley in the first place, called InfoVault, had almost the same product but WAY too far ahead of the adoption curve.

  • everything we did with Dropbox in marketing was failing yet we were doing well - what was happening?
  • our users were spreading the service around so we made it easier - referral links with two sided incentive program
  • inspired by Paypal’s early $5 referral program
  • always think about what people who’ve grow quickly have done - split testing, landing page optimization etc.
  • their users have sent 2.8milllion referral invites!
  • vote box - vote for features you want in vote box - a bit like User Voice, I guess
  • recommended Dave McClure’s startup metrics for pirates - I have to concur, this deck is a must read

David Weekly closes the talks re-emphasizing the ideas we’ve heard today.

  • Second person to recommend UserTesting.com - I need to try this out - user testing is very important obviously
  • study your users for hypothesis + inspiration
  • pivot: wysiwyg turned out to be essential because users weren’t sharing the product not because they couldn’t use it but they thought their colleagues weren’t so smart
  • pivot: demographics - hunch 5% educators; interstitial survey showed actually 45% educators!

The final panel q&a is led by Sean Ellis

  • q: tips for getting customer feedback.
  • a: use your network to find early adopters; ask them to tell you what they’re doing today
  • q: what about getting data from competitors?
  • a: ignore competitors initially - focus on when customers see it as a must have which tends to mean they don’t know of a competitor that meets their need
  • my thought on a panel comment: um… does anyone still fall for the I’m a student and want to learn about your business guff?
  • another panel answer: buy keywords around [competitor name sucks] put up a landing page and see what people are saying
  • q: should you ever do pr before p/m fit?
  • a: how do you know what to push?; absolutely not; pr can be a bad track and get sucked in before you have found a sustainable business model; until you have found a point where you are validated then you pending zero; counter view - don’t do it too early but maybe do small amount to raise credibility and expertise by getting tech press coverage, but generally in agreement (not much of a counter view)
  • q: what about and ad driven business?
  • a: challenge with ad supported is you can drive away users by optimizing for advertisers and vice versa; advertising not big unless you hit scale which means satisfying many people
  • q: what if your product/market fit is in a segment that turnout too small?
  • a: one tough thing about entrepreneurs is being able to give upon opportunities; probably better to build a big market with p/m fit than throwing it away and searching for a new market and fit
  • q: how to grow out of minimum viable product?
  • at least one panelist took issue with idea that minimum viable product can be just a landing page; a MVP does not mean you are not building a product just that you are testing your assumptions; don’t spend money unless you have a goal you are going to achieve with it
  • q: are there any differences to validating a product with strong network effects?
  • a: (Sean) with one company we found the value proposition was a moving target like a market place it is very hard having to build both sides of the market; you can build products that test network effects
  • finally missed question but thought was that paid channels tend to get saturated and organic virality is better

…and we are done. Eric is up closing out. What was the purpose of getting together today? It was about having a conversation, and it has been a great conversation. Not everyone even agrees, when you look at the folks on stage today. We only have the beginning of a theory and a beginning of a community of practise. It has been a great day

[I’ll edit to tidy up tonight]