Live Blogging Rough Notes
http://theleanstartup.com/sxsw/
On Twitter: #leanstartup
Today is about the 5 minute montage period of movies about Startup.
Startup = Experiment
Goal of startup: Stop Wasting People’s Time.
Most Startups Fail (e.g., Web 2.0 logo slide)
Anything can be built with software, but the question is should it be built.
Unit of progress: Validated Learning.
Customer Development, Steve Blank.
Hypotheses, Experiments, Insights.
His startup had thousands of lines of code, thrown out because they couldn’t get anyone to download it. Perhaps alternative approach, create 1 page with photo shop version of app with “download” button. If no one clicked, you learned just as much.
Our experiements get more productive based on validated learning.
Ideas, build, code, measure, data, learn (repeat) …
Minimize TOTAL time through the loop … (not build faster)
(All this info is online, like cool circle visual of this with a few elements listed on the side)
Innovation Accounting (Not talked about today, you’re gonna have to read his forth coming book to hear more on that …. 
Pre-order here: http://theleanstartup.com/
* Opening Talk Ends *
Panel of 3 folks:
Case Studies on LEAN BUILDING:
- Pascal Louis-Perez, Wealthfront
- Ash Maurya, USERcycle
- Parker Thompson, Pivotal Labs
Wealth Front
- Ship to Product 30+/Day
USERcycle
- Backlog
- Mockup
- Demo
- CODE
- Partial Rollout
- Validate Qualitatively (user interviews)
- Full Rollout
- Verify Quantitatively
Goal: Acheive 60% activiation rate.
A feature only gets to live if in a short time it can prove to support the macro goal. If not, we remove it from the application. Its important to go fast, but it has to be limited by how fast you can learn.
Pivotal Labs
Started out talking about how he tested out his presentation and will skip the parts that people said sucked.
The Visionary (covered this) … narratives often created retrospectively. Usually does an entrepreneur a disservice.
Story about a company with a huge vision of changing education, everyone has something to teach … etc … 6 months of dev, pushed out the door, found what they built didn’t make sense to their users.
*END PRESO’S*
Panel Questions:
Telling an entrepreneur that their vision is wrong doesn’t help. I’m sure you are right, let’s find an empirical question to prove it.
How know what to test?
Its easy to get lost in the data. Try not to have more than 3-5 things you are testing. Pre its mostly qualitatively. Post its quantitatively like user activation, return users, etc.
(Eric says Running Lean book is a good source about this, I think Ash is the author)
Find metrics that are really relevant to your particular business. Like for a job site, measure the number of people that return back to your job site.
(Eric: “Vanity Metrics” is what you put in your Press Release, but when it comes to making decisions, etc. Don’t get caught drinking your own coolade. We all have our own private realities of what’s working.)
Actionable metrics that tell you things about your customers…
Metrics often go up and down based on the markets, who tweets on it … but other metrics can be less volatile (e.g., activation rates) … though there are some differences.
WHen does your startup end being a startup?
Eric. Some believe your company goes through phases. But in the future that is not what will happen. Fast followers are just getting faster and better. Managers of the future will have to manage a portfolio of innovation and disruptive innovation (didn’t catch exactly what he said).
Parker. Some of our customers are actually large companies (intraprenuers) and have done some really impressive things with this.
Other guy. You need to understand of who loves and hates your product. Then get out of the building and talk to them. Understand why. Answer the empirical questions.
Eric. You got to ask yourself what customers actually do in reality. Its not about solving lack of traction on the whiteboard.
Other guy. Our demo’s are screenshots.
*END PANEL*
The Lean Startup: King of the Apps Showdown
Who will win the King of the Apps Showdown? YOU DECIDE! Play judge alongside Robert Scoble, Eric Ries, Dave McClure, Bill Boebel, and Stacey Higginbotham for a startup showdown. The winner gets $10,000 in cash &prizes, provided by the awesome folks at Rackspace.
Moderator:
- Dave McClure, 500 Startups
Panelists:
- Robert Scoble, Rackspace
- Eric Ries, The Lean Startup
- Bill Boebel, Rackspace Hosting
- Stacey Higginbotham, GigaOM
SnapAppointments Quick Case Study
- Started, stopped, restarted a few years later
- Simple CRM, etc.
- Pays for itself even if you are just able to remind one person that makes their appointment when they otherwise wouldn’t have
- Attorneys, salons and spas, Dentists, etc..
Panel Reviews
- Scoble. I get it. I use Tungle.me, issue is adding buggers. How are you getting user feedback, watching them? (Yes. My mom is a photographer who is computer illiterate. We have several people like that. We give it to business owners who are not very computer literate).
- Ries. I get it, get it pitched several times for various service providers. Its a clear problem I like. In terms of presentation, I’d love to hear what they’ve learned from customers. (30-60 seconds of questions: “we learned its fun to use. very simple. don’t want functionality. Dave: what problems did they tell you and what you did to solve it. One complaint, not on functionality today but features. Ask for things like buffer times, etc. So we built in buffer times. Booking multiple services at a time, like haircut and wash.)
- Eric. 2nd question: what’s one element of your vision that was wrong? (We thought the business would be more excited. We found that its the client getting excited about it and demanding it. This is the way we do business and manage our schedules.)
- Bill. Clear what it does. Has a few questions. Its hard to sell to people who are late on the adoption curve of things technical. (Usually been successful at this. But most of them I know. Even to people we don’t know have been good. After talking 5 minutes people willing to sign up and give it a whirl. Small town in Montana.) You guys are going to have a way to do that at school. You can’t walk down the road to do that to everyone. Focusing on one type of service provider might be better. Nail a vertical (We think this will be client driven, so don’t focus on just one vertical).
2nd Company to be reviewed by Panel:
Snap Web Manager
www.snapws.com
Allows you to manage your websites, back ‘em up to the cloud.
- Stacey. Something about this is like echosign meets time machine. Wished you started out with who you are trying to sell it to. Later it became clear but I had to figure it out.
- Boebel. Get a better more update interface. Spent a huge time on features. What is the most important thing to focus on. Perhaps a good UI guy could really help your product and pitch.
- Eric. You have the most features I’ve ever seen. Do your customers use all of them? …
- (Lots of questions and answers, I started downloading apps and got distracted wtih my new ipad2 … sorry … didn’t miss much though)
Lightening pitch, 3 minutes from guy from crowd … something about pulling in all your data and showing it on a map. mapomatic.me (sign up for beta)
- Scoble. I’m interested, but I’m a location based whore. Like where will I go.
- Boebel. I love the idea of putting context on the map. But it is a separate app I have to run or is it on top of GoogleMaps.
- Eric. What do your customers think? You guys are my customers, its not launched yet. Just did a build. Eric. Can you launch it now. CHEERS FROM THE CROWD. Hesitation and then the 1 week approval issue of the apple store kicks in. (Fun moment).
2nd 3 minute pitch:
Planzai. Started out asking who has ever managed a project. Got married. Has kids. Etc. Very cool accent and good start. Pulls out site.
Judges ask questions.
Eric. So what does your customers think. “Um, they like it. Obviously.” (laughs, a judge, “wrong answer”).
Mccure. Next pitch, do you have customers and shipped product? Yes.
Icebreaker. A new way to meet people on Facebook and Twitter.
Its global, uk users …
Went to wrong url, corrected, Icebreaker without the “e”. (Right. Got it.)
icebreakcrapp.com
Ries. What do you do if you pull up the app and no one is around you? (we can pull in twitter people.) If Robert Scoble can’t find people …
DONE: then awarded $10k to SnapAppointments.
* LUNCH BREAK *
And We are back …
The Lean Startup: Pirate Metrics for Design & Distribution
Dave McClure debuts “Startup Metrics for Pirates: The Sequel” — it’s your chance to better understand design, distribution, and proper use of expletives. Following Dave’s talk, you’ll get to hear case studies from some of the most effective “Lean UX” designers and developers this side of the Mississippi.
Opening Speaker & Moderator:
- Dave McClure, 500 Startups
Dave McCure - Starts out with criticism of his own slides. Then flies into them …
Colorful fonts. Fast talking. “You can do these in any Fucking order …”
Started to have a lot of good quotes, lots of laughs … but I got distracted redeeming the free Lean Startup Bundle for SXSW courtesy of Microsoft BizSpark ($2,740 in value, sold for $99 special, got for free) because I “pitched” my app and how I’m using #leanstartup principles. Told her about my sxsw.inward.me iteration … Was fun …
Case Studies on LEAN UX:
- Laura Klein, Users Know
- Janice Fraser, LUXr
- Ian McFarland, Pivotal Labs
- Dan Martell, Flowtown
DAN MARTELL, FLOW TOWN
1:24 pm CST
Design metrics and distribution.
Who wants to learn not to fail? Who wants to learn that now? Alright, you are in the right place …
DESIGN: wireframe
DASHBOARD: look at metrics internally (weekly retention, you’ll adjust as often as you track, so track weekly)
Try doing an activity stream. Color code based on what they are doing. Then drill down into specific ones to see where they had troule.
Experiments
Feature X -> Metric Y.
“FEature X will improve Metric Y.”
Is this feature that is just something cool, or is it something that will move the needle.
DISTRIBUTION: Invite Flow.
Figure out how you can get something to your users that they can share. Collaboration software inherently has this. Figure it out for your product. Could be a report you can share with people. Turns every user into 6. Helps you avoid paid acquisition of users.
* Laura Klein *
Use Case: Food On the Table.
Integrating Quantitative and Qualitative. What and Why.
Read the blog to find out what the problem was.
*Janice *
Starts are like bunnies. Who has a startup. Its like this. Then there is this guy, Eric Ries, makes you want to make it look like this.
Case Study: Taskrabbit.
Had a home page conversion problem.
Redesigned. Everyone loved the new design. Wanted to do an A/B test on it. UI guy wanted to do is design for all pages. He was terrified. Speaks to the cultural issues that must be addressed. Why expect designers to get it right the first time.
* Ian *
Our best Waterfall Project Ever … and why it sucked.
Huge PRD. Did with a great design firm, solved the brief.
Lots of rework re: nav bar.
Bad design decision … time elapses … discovery of design decision. Everything in between is mostly Wasted Effort.
Waterfall is a great system if you knew EXACTLY what you are going to build. And even if you do, change happens.
Tested it, things went down. Did small change, then it doubled. So, test if directoinally going right, then double down. Saves money not just in wasted development but also the lost business of 5% vs. 10% conversoin.
* Panel *
Fraser. You can’t A/B test your vision. The key to MVP is testing your empathy (when not building specifically for yourself). Its about how well you listened and understand the lives of the people you are building for.
More good commets, but I got sucked into a few links …
DONE.
30 minute break before …
Moderator:
- Eric Ries, The Lean Startup
The Lean Startup: Your Startup Sucks. Pivot to Win.
The difference between success and failure is your ability to pivot — fast. Learn how to more quickly recognize the signs of a pivot point, and how to pull the trigger when it counts.
Case Studies on PIVOTS:
- David Binetti, Votizen
- David Cancel, Performable
- Sean Ellis, FREEjit
- Manuel Rosso, Food on the Table
David Binetti
State busines model assumptiosn and iterate (good slide, didn’t get it)
Pivot is not synonymous with “change”. Its about your business model.
Uses CDM and AARRR (wrong accredmin)
3 pivots, 4 versions.
CDM / Hypothesis
Product / (try get his slides)
Problem
CHannel
Demand
Market
Competitive
Metric / Description
Acquistion
Activation
Referrals
Retention
REview
Improve results with A&A.
Metric / Version
Huge bump in conversation, especially in activation.
Improve R&R
Didn’t get the improvements wanted. So did a Pivot. Involved customers, made it easy for them to get involved. Then took that info and brought the produt forward.
Product: “Simple involvement tool.” Restated assumptions. People could use Twitter to contact senators.
Version I vs. II was not a pivot.
When made changes based on the “vision”, not metrics. Vision is a requirement. Not going to build a public company a/b testing.
Huge jumps. Still no revenue. That’s next. But heading in the right directoin.
Improve Revenue.
Dontation model, wasn’t working.
Pivoted again Version III. Market: Campaign Management.
Nobody bought. Nobody paid. WORSE, they told me they wanted it. That they loved it. Complete Fail. Spent a lot of money.
At the end of the day, “I’m just not that sure about this twitter thing. A little afraid of social media.” No contract, no sale. You want to believe. But show yourself the money. There is nothing more concrete that you can do to prove that you have something that people are willing to pay for. Your enthusiasm can kill you. False positives is when you throw your money and life away on things that you could have learned about a lot earlier.
Pivot. Freemium. Special delivery on a per message basis. Only pay for adde value. Initial sale: $200 first day. Much smaller but wille xpand.
0% to 11% revenue jump.
No PR. Period. It distorts what you need to learn. Especially if you start to believe your own stuff. We are pushing away New York Times, Life, etc.
Iterations loops got smaller, faster, fewer. Learning per dollar spent (repeated it like Alec Baldwin and his ABC speech, nice touch).
You are building a business, not a product.
Eric Ries. If you build to see what happens. You succeed. You see what happens.
* David Cancel *
Combine insights with actions …
High converting landing pages made easy.
First iteration: no markertor gave a crap about automatic optimization. Marketers want to customize. WYSIWYG experience not gratifying.
Iterated. Made it form base rather than WYSIWYG. Made it so simple.
Pivot 1: lowered price. Simple. Learning SMB pain to deal with. Supporting customization was hard. Landing pages were “nice-to-have”.
Pivot 2: Started to focus on Metrics-Driven Marketers.
Appreciated our platform. Willing to pay up front. Heavy daily users. We are pat landing pages, part metics and part automation.
Rants:
Customers = Paid
Solve Customer Problems
The Real Competitor is Indifference NOT Startup (someone not willing to pay you, thats a competitor, not some other startup, I don’t give a shit about techcrunch).
* Sean *
Pivots to IPO. (He didn’t mention company, but I believe its Logmein).
< missed some notes, giving peopel next to me my alpha url sxsw.inward.me >
Problem:
Unintended use case of personal remote access.
Customers overwhelming customer support
Solutions:
Formed new company to address the consumer opportunity.
Make easy to use and no need support.
Pivot 2:
Problem:
Competitor pivoted into this and got bought by Citrix. They started spending $100M’s in marketing.
This was a business model pivot. How do we come in with the freemium model. e.g., eFax, AVG. New it was viable. Self funded and boot strapped we had no way to compete head to head. Required tech breakthrough on costs.
Year 1: US vs. Them in google searches … went from 1/2 to 2x in searches despite their budgets. Continued to do well. 4 years grew, Citris flat lined. IPO filing 4th year.
* Manuel Rosso *
FoodOnTheTable
Why Concierge MVP?
If you can’t get them to adopt your idea high touch, face to face, you can’t do it online or cold …
Q: how can we support a small number of people without any code?
Once we figured that out, then we started to automate based on the initial learning.
We used a form that kind of looked like a website. We went to a starbucks, things on sale at the grocery store across the streeet. Asked them lots of things and how they did this process. We learned we were wrong about a lot of it.
e.g., how much you spend a week on groceries. Most moms didn’t know, didn’t care or didn’t want to share.
On the right, we had a Weekly Meal Planner. One would be talking, the other would be filling out the meal plan for this mom. Then print it out when she is talking.
e.g., learned that you don’t put Monday, Tues, Wed, etc. it was stressful. However, if you put numbers rather than days, it reduced stressed.
We’d then call them on the phone the next week. Would do the same, but remotely. Testing what it was like when you were remote and not infront of them. At end, would say wait 5 minutes as we put it together and email it to you. The next week when we tried to make it more automated, we made google docs. We then used the collaboration features that would allow us to cut and paste things that they’d then see. They’d say, wait, no, not that one, we’d remove and then add it elsewhere.
We started to arrive at what a remote user experience would be that ugided our development.
GET OVER YOUR VISION AND THINK SIMPLE.
Content Category // Vision // Starting Point
Groceries store 35k+ 1
Sale items per store 150-300 5-10
Recipes 50k+ 3 per sale items
CODE IS YOUR ENEMY. FUTURE CUSTOMERS ARE YOUR FRIENDS. THE MORE YOU CAN PROTOTYPE WITH PAPER THE MORE YOU’LL LEARN. TRULY THING WHAT IS THE MINIMUM YOU CAN DO.
END.
* Ries *
One good way to work on MVP and landing pages is just putting up the benefits. The “magic” product. If you can’t get people to buy your magical product, then don’t build it. (paraphrased).
Q: how apply this to network effect companies?
A: (panel) apply to small network to get critical mass (e.g., Twitter at sxsw or Facebook at Harvard). You’ll probably have to raise more money.
STEVE BLANK IS NEXT!
————–
The Lean Startup: Steve Blank on NextGen Entrepreneurs
Steve Blank is one of the world’s leading experts on entrepreneurship — nab a seat if you’re lucky and get the inside track on what it takes to be a NextGen Entrepreneur, how company-building is changing, and how you need to increase your competitive edge. Following Steve, you’ll hear case studies on customer development from the venerable KISSmetrics and Meetup.
Opening Speaker:
- Steve Blank, Stanford University
A mentor and investor of Eric’s from IMVU. Great book, The Four Steps to the Epiphany.
* NEW RULES FOR THE NEW BUBBLE *
Only a few of you out there will make 100M’s. A few others a few million. You are crazy because all of you are looking around feeling sorry for the people sitting next to you. (paraphrased).
1970-1995
Building a Business (acted like startups were small versions of big companies)
Building a Startup th eHard Way.
Playbook was a business pan. Product introduction model.
Two implicit Assumptions:
- Customer problem: Known
- Product Features: Known
There is no such thing as a 10 year startup. You have a two year startup attached to 8 years of failure.
More startups fail from a lack of customers than from a failure of product development.
1995-200: Flipping A Business (The Dot.com Bubble)
Nestscape showed market for immature companies going public.
$10’s millions to sart, long roduct dev cycle, millions customers, liquidity/ipo, repeatable methodology …
It was great for some people, you make millions, retire and then get up on stage as a professor and look brilliant.
Playbook was the S1.
2001 - 201
Why startups are not small version of a large company.
Build with less than 1/5M to start (open source, commodity hardware),
SHort product dev cycles (agile development)
100’s millions Customers (social networks, consumer)
Liquidity - M&A (revenue and network of users)
Repeatable methodology (agile + cust dev)
What’s a Startup?
SEARCH (scalable startup) to EXECUTE (Large company)
- Startups: A temp org to search for a repeatable and scalable business model.
- Large Companies: Execute found business models.
AND THERE IS A BOX IN THE MIDDLE.
“Transition” — build.
You’d think the day you put up a slide that says, “I think we’ve found a repeatable model” … you think they’d give you a pin. Actually, on the way to the car they are calling a recruiter.
World class entrepreneurs are great at search. The same people who would fly here, two days, go back, etc. The same people who would hate to do the same thing over and over again.
Founders. Survivors of dysfunctional families are great at shutting out noise and deal with chaos. Its the first time its a benefit to them. It all falls apart though, when you get to repeatable model. They then throw grenades at own people.
How can you teach this? We are teaching artists. We can teach art, but we can’t teach them to be michleangelo …
Metrics Vs. Accounting.
- Customer …
Tall VP of Sales from large companies melt down. Cause you are searching for a business model. Its different.
2001 - 2010 (The lean startup)
Were still using same business plans …
No business plan survives first contact with customers ….
4 step process:
customer discovery, customer validation, customer creation, company building.
Get out of the building, stop selling and stop listening. You can’t hand this to someone else. Its done by founders.
Test all hypothesis.
MVP and the customer is the first two of the 9 of your business model to nail.
This “pivot” is the heart of customer development. It simply means you are wrong, and its a normal part of the learning and discovery process.
Stanford Lean LaunchPad Class …
What students did in 8 weeks …
Hypothesis , fake emails, interviews, updated things what they learned….
8 weeks Business Canvas (pretty cool) … and you can see all their learning.
(all stanford slide decks of the students will be on Steve’s blog)
2011 - 2014
New Bubble (Hypotheses)
- We’re in one
- Rules are different in 2011 then 2006Exits will be IPO and M&A (real ipo’s, qulaity will diminish, acquires will targe)?)
Rules for The New Bubble
Founders/VC’s can engineer financial transactions (talking too fast …)
2011-2014 (Order of Battle)
Lists the limited number of people looking to fill holes via M&.
Wide adoption, monitize later, visibility, look larger than life, (new rule: if we are in a bubble, you might want to play this game: conferences, blogs, pr, etc may be your best friend again).
You can ignore this. I’m not suggesting you have to follow the new rules. I’m just suggesting. If I am right, you might want to consider.
Case Studies on LEARNING:
- Hiten Shah, KISSmetrics
- Andres Glusman, Meetup
Hiten (KISSinsights)
Lean startups are build to…learn.
Get outside and talk to people.
Start with a hypothesis. These types of people are having a problem doing …
Talk about their problems, not your solutions.
Find the severe pain. The deep problem. Its really important. When you talk to people. To understand. To get to the point of this is the hardest part of my day. Etc.
Test your riskiest assumptions first.
(e.g., people wanted templates of what to ask of their customers).
< great presentation, but missed some in my notes because I was listening … and my head is about to explode
Andres Glusman
Beyound 2 guys in a garage. VP at Meetup.
Does customer development, but doesn’t call it that. 7 years at Meetup.
“What not to do with Steve’s book…”
< another good one, notes are slim …>
Watching a user go through the beta (to get inline for it)
“I’m trying to …”
“That was humbling” - A developer after the session.
“watching people use the stuff you build, is the best way to comfort the malconovich problem”
Doing 600-700 use tests a year now. Start with projecting a user’s screen on the wall.
Test 2-3 days a week (t,w,th).
Set up tests prior to knowing what will be tested.
Shorten the window of “I wonder how a customer would respond to this” with slotting them into testing …
Goal is to have as many of these sessions as possible. We use go to meeting and other stuff to track.
No reports or big concepts sent up. We discuss it.
- give team access to customers (directly)
- Look for boulders in the road (doesn’t have to be super scientific(
- Substitute frequency for precision
- Strip out all costs wherever possible … but be willing to pay
- Have discussions
- Launch and learn
- Do as much as you can in low tech way, then scale
Q&A
Included:
Don’t ask someone like me to be a mentor. Its not a job. Its a relationship. I mentor those that I learn from in the first 10 minutes. etc.
{ 0 comments… add one now }