Lean Startup: A Founder's Guide

Editor's Notes

• #2: My name is Abby Fichtner, Hacker ChickI’ve worked with hundreds of startups, I’ve actually been called the cheerleader for Boston startups & the guardian angel for boston startupsI’m currently in the midst of opening a hacker space, which is a non-profit community space for people to come together and createI want to talk about Lean StartupThe thing to know about startups is that they’re not at all like big companiesAnd they can feel pretty crazy & chaotic, but there’s actually a method to the madness & that’s what I’m going to talk about today.
• #3: Let’s start off with – what do you think makes entrepreneurs entrepreneurial? Any thoughtsSaras “saraswathy” has done some research into this topic and what she found was thatNormal people employ what she calls Causal Reasoning. That is, they set an end goal & come up with a strategy for how best to reach that goalStraight forward –They have something they want so they come up with a plan to get itEntrepreneurs however, use Effectual Reasoning.In effectual reasoning we don’t even start with a specific goal in mind.Instead, we take a hard look at where we are TODAY – who we are, what we know, WHO we knowAnd we start to tinker – we start imagining & implementing possible things that can be created with all of thisAnd so we progress, allowing our goals to emerge over time.And so causal reasoning is all about the perfect plan & then following through on itBut effectual reasoning is more about exploring. The focus is on execution over planning -- trying stuff out & then adjusting our path based on what we learned from it
• #4: Has anyone ever read Where Good Ideas Come From? By Steven JohnsonSUCH a good bookAnd in it he talks about these patterns of innovation that happen over and over again… but not innovation that happens in startupsHe’s talking about innovation in evolution & how our planet has evolved & how we’ve evolvedAnd he then ties this to innovations we think of – like the lightbulb or the combustion engineOne of patterns is called The Adjacent PossibleAnd the idea is that at any given point in time – there is a set amount of what’s possible.Things that can be done with the technology and the knowledge that we possess todaySo all innovation is going to happen at the adjacent possible – meaning what’s just beyond what’s possible todaySo if we can look at what’s possible today and take it just a little bit further, we’ve actually expanded that set of what’s possibleSo we’ve made the world a little bit better for everybody, because there’s more that can be doneAnd so I believe this idea of effectual reasoning where you look at what’s around you & see how you can build upon it to take things to the next levelI believe THAT is how we change the world.And what I love about lean startup is that it actually gives us a method for doing this.
• #5: Eric Ries is the creator of the Lean Startup methodology
• #6: 9 out of 10 new products fail
• #8: As a founder, you start off with your Big Vision for how you’re going to change the worldAnd I know – you’re saying no, no – I’m doing effectual reasoning, I don’t have a specific goal in mindAnd while that may, you really can’t found a company without a vision of what you want to achieve, otherwise – what are you doing?It’s just that the trick is – to identify where your vision intersects with what reality can actually accommodateAnd that’s where the exploring comes in & the willingness to try different things out
• #9: Great example of this is a Startup called the Point.The Point’s idea or BIG VISION was to harness the collective power of all the people on the internet<Read Tagline: Whether you’re raising money, organizing people, or trying to influence change: if you can’t do it alone, you can do it on The Point>Fantastic vision! But totally unfocused. It could be used for anything from political activism to boycotting multinational corporations to getting a group discount at your favorite pizzeria.It wasn’t gaining traction, they had to lay off the co-founder. About to become one of those 9 out of 10 statistics…But the CEO looked harder and said is there anything in this vision that IS working?And actually – that getting people together for a group discount at a pizzeria, THAT was working.And if that sounds familiar, it should…
• #10: Because that CEO was Andrew Mason and he took that idea and turned it into Groupon, which became the fastest growing company EVER.Over 30MILLION Groupon subscribers in North America alone
• #11: So I want to share with you some of the key learnings that Andrew Mason took away from The Point -> Groupon experience<play video>
• #12: I don’t care that you’re MIT…<read>Build that very small thing & get it out there to the masses, and keep on trying different things. Eventually you'll get it right
• #13: The thing that you have to keep in mind as you’re starting your own startup… <read>Large organizations are already operating in a known market, solving a known problem. They’re focused on execution: better ways to solve that problem for their market & reaching more customers
• #14: Startups are starting from nothing.They haven’t even validated that there’s a need for this fabulous idea of their’sThey’re about learning and discovery, not execution.
• #15: So founder lesson #1 is….<read>And this is REALLY hard…
• #16: Because building stuff is FUNAnd you’re doing a startup. That’s really sexy. So your friends all want to know what you’re building.And you’re all – not much, but I’m learnin! Learnings sexy too right?Building stuff is fun.
• #18: In lean we have a term for that – we call it muda.Muda is a japenese term because all of these lean concepts stem from Lean Manufacturing, which actually came from the Toyota Production SystemAnd we might translate it to waste, but it’s actually much broader than thatIt means anything that doesn’t add value or is unproductive.
• #19: So Startup Lesson #2 is eliminate Muda<click><read>In fact… Eric Ries said the the #1 most important thing for a startup is learning to tell the difference between value & waste.
• #20: So, to show a really big example of waste I had to go all the way back to the 90’s when we had really awesome phones.In the early 90’s, Motorola spun off a new company called Iridium to build satellite phonesAt the time, it was a HUGE engineering triumphThey spent $5 BILLION dollarsBuilt it to scale to MILLIONS of customersAnd when they finally released, 11 years later, nobody wanted itAnd it became one of the largest bankruptcies in US history
• #21: John Richardson, who stepped in as CEO after all of this happened said<read>
• #22: The truth is… <read>And so we need a different model then that which big companies follow
• #23: What do you think startups should do differently then big companies?<split into groups of 3-5 people, take 5 minutes to think about how you’d run a startup differently><Have volunteers share their best idea>
• #24: The typical way we think about companies operating is… you’ve got a vision of what you want to do… <click> and you go build it.But when you’re creating something brand new & you haven’t yet validated that there’s a market for the product you’re buildingThat approach tends to lead to… WHAT?<click> 9 out 10 new products failingWhich is bad, right?
• #25: The typical way we think about companies operating is… you’ve got a vision of what you want to do… <click> and you go build it.But when you’re creating something brand new & you haven’t yet validated that there’s a market for the product you’re buildingThat approach tends to lead to… WHAT?<click> 9 out 10 new products failingWhich is bad, right?
• #26: The typical way we think about companies operating is… you’ve got a vision of what you want to do… <click> and you go build it.But when you’re creating something brand new & you haven’t yet validated that there’s a market for the product you’re buildingThat approach tends to lead to… WHAT?<click> 9 out 10 new products failingWhich is bad, right?
• #27: So our goal instead is to do some exploring to figure out what piece of our vision reality can accommodateWhat is our group discounts for pizza?So what does this exploration look like?This is called the Build-Measure-Learn loop. And this represents the process we’ll be iterating over to move us towards a successful business modelThe idea is – you start with some assumptionsAnd then you think about how you might be able to LEARN whether these assumptions are right or not(I am assuming that if I give people a platform for group action, they’ll use it for worthy causes – like political activism)So now you step back and say, okay, what can I BUILD that will let me capture the right DATA in order to learn if my idea is true or not?(maybe I can build the platform for group action & then capture some analytics to see how people use it)But – here’s the catch, and this is what Andrew Mason realized after he’d spent 10 months building The Point.We need something faster than taking 10 months to build the platform and see how people use it.We need to get that 1 little thing out there & see what people do with it.So the key to iterating through your ideas to a successful business plan is to <click>GET THROUGH THE WHOLE LOOP AS FAST AS POSSIBLEThink days or a week or two, NOT 10 monthsSo now we need to ask the question again. I’m assuming that if I give people a platform for group action, they’ll use it for worthy causes – like political activismHow can I learn if that is true or not in just a couple days or maybe a week?So you guys might have some ideas – keep those in mind
• #28: Our strategy for doing this is something called Minimum Viable ProductWhich says – what’s the SMALLEST thing I can build that will take us to the next step?And remember – we’re focusing on LEARNING over execution, so what’s the smallest thing I can build that will let me LEARN something?Preferably whether the assumptions I’m making are correct or not – so I can learn if I’m moving in the right direction or not
• #29: So a story about this that I like is a blog post that Christian Wyglendowski wrote called The Minimum Viable Tree HouseAnd he’s talking about how he’s building a tree house for his kidsAnd it’s taking a while…And his oldest one says <read>This time, Christian had a WAY bigger vision…
• #30: This kind of ROCKS, right? But here’s the thing – if we built this, is this really want our target market wants?If we’ve got kids – they’re probably looking for something high up, away from the adults, that feels like a secret stealthy kids only placeThis – ain’t stealthyFurthermore – if we actually BUILT something like this – you just KNOW we’d be inviting ALL our friends over & be all “look at what I built for my kids”Do you think kids really want to be hanging out in the place that mom & dad and all their friends are?Kind of defeats the purpose, right?
• #31: And so kids might actually LOVE thisTHIS is our minimum viable feature set: It’s up high, away from the adults, and best of all, it’s available in a day so they can start playing NOWAs opposed to the other one where maybe it’ll be ready by the time they go off to college.You should NOT be building this (flip back)You should build THIS (slide)— something so crude that it embarrasses you to put it in front of your market.
• #32: So Founder Lesson #3 is <read>
• #33: In a startup it’s obviously essential that you’re able to BUILD a productBut it’s just as, if not MORE important to BUILD your customers.So Lean Startup combines agile development (for how to dev your product)With Customer Development (for how to develop your customers)
• #34: So Startup Lesson #4… (just point & pause), OK.
• #35: The way we develop our customers is with a process called Customer Development (created by Steve Blank)It has 4 steps, as Steve describes in his book: 4 Steps to the EpiphanyIn the 1st step: Customer Discover – we’re discovering who our customers are & what problems they haveIn the 2nd, we’re validating this knowledge. This is our test step and it’s important because… we’re dealing with people hereAnd people are notoriously complex & misunderstood & tend to say one thing (wow, that’s an awesome idea, I would totally buy that!) and do another.So you see there’s this arrow going back here. This is a very iterative process and you need to iterate on this until you get it right.And ONLY once you’ve been able to validate who your customer is & what they’re willing to pay do you get to start scalingScaling up your customer baseScaling up your company and your product
• #36: And so it’s really these 1st 2 phases that we’re talking about when we talk about Lean StartupAnd what we’re trying to achieve with these phases before we start scaling is...
• #37: Something called Product/Market FitWho your market is (and what their problem is) and what product you’re going to build that’s going to solve that problem for them.Requires:Customer is willing to pay for the productCost of acquiring the customer is less then what they pay for the productThere’s sufficient evidence indicating the market is large enough to support the business(not mom & roomate)So how do we do this? How do we do Customer Development….?
• #38: We get outside the buildingSteve Blank likes to say NO FACTS live inside the building, only guesseshttp://www.flickr.com/photos/jenniferyellow/2203389749/sizes/l/in/photostream/
• #39: So let’s say our BIG VISION is that we want to make flying cars.So when we’re doing customer discovery, we get out of the building, we talk to customers & we try to learn…
• #40: So let’s say our BIG VISION is that we want to make flying cars.So when we’re doing customer discovery, we get out of the building, we talk to customers & we try to learn…
• #41: Once you’ve gathered this knowledge, you bring it back to your development teamWho creates not some 10 month long full blown product, but the Minimum Viable ProductWhat’s the MINIMUM thing we can build to take us to the next step?They give this Minimum Viable Product back to the Customer Development team who sets out to validate it.
• #42: Once you’ve gathered this knowledge, you bring it back to your development teamWho creates not some 10 month long full blown product, but the Minimum Viable ProductWhat’s the MINIMUM thing we can build to take us to the next step?They give this Minimum Viable Product back to the Customer Development team who sets out to validate it.
• #43: How do they validate it?They give it to a customer and see if they’re willing to pay.Because UNTIL you have customers that are actually willing to give you money, you don’t have a viable business.You just have guesses.
• #45: So go back to the drawing board.Think about what assumptions you’re making, build a little something to validate if those assumptions are right, put it in front of the customers to learn from it, and keep iterating.
• #47: So – if we’re not moving towards a set goal
• #49: Here are our lessons so far – going to give you a minute to read themAnd then let’s stop & do some questions before we move onto the last part… questions …So now that you guys know all about startups, let’s dive into some concrete examples of what startups are doing& then I’m going to turn it back to you guys to give it a try with some of the big ideas that you have been coming up with
• #50: So let’s start with a really basic example
• #51: So let’s start with a really basic example
• #52: Oops – this feature isn’t available yetIf we had some analytics that could show us the count of 0 people that clicked through, that would have been just as effective.We would have learned just as much without having to actually build the product.And so a totally viable lean startup tactic is to build a landing page.
• #53: A great example of this is Buffer. They do social media sharing.This is how they look today. But this is not how they started…
• #54: 1st, they created what’s called a landing page to just see if anyone was interested.In other words, 1st off, could even get people to come to the page?And 2nd, will any of those actually click through to Plans & PricingIf they clicked plans & pricing…
• #55: Ooops, you caught us before we’re ready.But if you’d like to be notified when we’re ready, give us your email address.They tweeted it out, a few people gave their email addresses. That was good enough – we don’t need a billion people, because we don’t want to wait months for this – we just need some initial validation.So now, let’s move to the next MVPIs the next MVP to build the product? <ask>We haven’t validated our business model unless… what? <ask>People are giving us money!
• #56: So the next MVP added a page to check for just that.When people clicked Plans & Pricing they got 3 plans, 1 free, 2 pay.If anyone clicked any of the plans, they just got brought back to the Oops/give us your email address pagePeople continued to click the Plans & Pricing button, and now they clicked through on one of the plans, a small # even clicked a paid plan.Great, enough validation – they were ready to put something out there.They thought it was only going to be 1 week to build a really minimal version – I mean it’s just queuing up tweets to send out, right?As it turned out, it took 7 weeks and they had to leave out some pretty vital functionality like they didn’t even have a sign up wizard(think: couple of pallets & a ladder)But they put it out there & very shortly thereafter had 500 users, and they were generating revenue through their monthly plans
• #57: They’re validating their assumptions. They’re starting to bring in revenue to help pay to build out their startupThey have the validation of paying users which VCs love to see if they want to raise more fundingAnd they’re able to start getting feedback from real users so they’re evolving the system to what makes sense for the people using it. Not just what they THINK or guess might be right.
• #58: Another example is something called Concierge ServiceOccasionally called Wizard of Oz approach for the man behind the curtainBecause instead of BUILDING something you say, hmm, what could we accomplish MANUALLY for our users?And this is really weird if you’re a tech startup where you feel like your whole reason for being is using technology to solve problems.
• #59: A great example of this is AardvarkAardvark does social search. The idea is there are some things that you need to ask a real person, not a search engine.Like – what’s the most awesome startup to work for in Boston?Okay, how do you automate that? I don’t know – why don’t we see if anybody even WANTS it before we worry about how we’re going to build it.Which is exactly what Aardvark did.They put up a simple web form where people could ask questions.When a user asked, it would email someone at Aardvark -- probably one of the founders so they could understand what types of questions were being asked – and then a person at Aardvark would go track down an appropriate expert in the area, ask them via email, and then email the answer back to the initial user.Once they HIT the point where doing this manually wouldn’t scale any more.THAT’s when they realized, hey, it looks like we have a viable business now.Now we can go start building it.So Aardvark built it out. It took them a year. I imagine they did it very iteratively, probably prioritized to handle the most manually intensive parts of this first.And then they got acquired by Google for $50M. So it sounds totally ghetto. But it works.And they didn’t have to waste ANY time building things nobody wanted.
• #60: Last example is from Task Rabbit.And this is for when you’re a little further along, you actually have a functioning product & you want to add a new feature or a new design to it.There’s no reason you can’t continue using the MVP approach even once you’ve got users.So this was Task Rabbit’s initial design, they were getting something like a 5% conversion rate (not real #)They had a GUESS (or an assumption) that a better design would improve their conversion rate.
• #61: So they brought in a designer to do this new design.EVERYBODY loved it. They were SURE this was going to do it and increase the conversion rate, so the designer said let’s update the whole site with it.But Task Rabbit said NO – it’s just our ASSUMPTION that this is going to improve conversion.We need to validate that before we take the time to update the entire site.The designer was horrified.Because you’ll have the front page with one look & set of navigation, and then you click in and everything else has a totally different look & different navigation.My god, you can’t put this in front of customers.So they put the new page up, and sure enough, they had a higher click through rate on the Post a Task buttonBUT, the overall conversion dropped because people were abandoning the post a task page.And then they did something even WORSE from a design perspective and put up this totally ghetto 90’s style pick list when they clicked the button to make it easier for people to post a task.And, the end result was that the overall conversion rate nearly doubled – even with this awful UI in the middle
• #62: So they brought in a designer to do this new design.EVERYBODY loved it. They were SURE this was going to do it and increase the conversion rate, so the designer said let’s update the whole site with it.But Task Rabbit said NO – it’s just our ASSUMPTION that this is going to improve conversion.We need to validate that before we take the time to update the entire site.The designer was horrified.Because you’ll have the front page with one look & set of navigation, and then you click in and everything else has a totally different look & different navigation.My god, you can’t put this in front of customers.So they put the new page up, and sure enough, they had a higher click through rate on the Post a Task buttonBUT, the overall conversion dropped because people were abandoning the post a task page.And then they did something even WORSE from a design perspective and put up this totally ghetto 90’s style pick list when they clicked the button to make it easier for people to post a task.And, the end result was that the overall conversion rate nearly doubled – even with this awful UI in the middle
• #63: So you might think – this is a horrible practice. WHY would you want to put up this hacked together design and risk alienating customers?But – say it takes 1 month to refresh the site with the new design.If they’re getting 10,000 views a day, at an increased conversion of 4.75% - that’s 13,500 additional customers posting tasksIf each conversion is worth $2, that’s $26,000 in revenue by not waiting.That’s more than most startups pay for design services, total. So right there – boom, the ghetto design paid for the nice designAnd Task Rabbit is doing pretty well today!