Customer development: the science of acceleration for growth businesses | Bob Dorf | BoS USA 2012

This was another incredibly well-received talk at Business of Software Conference – for good reason. Bob Dorf has earned his startup chops, is entertaining, highly experienced and has seen a whole lot of successful and unsuccessful businesses. More importantly though, he has spent a huge amount of time thinking about why he was successful and unsuccessful and is totally passionate about helping others learn.

Video & Transcript below

Among other topics, Bob covers…

• Why startups don’t fail from technology.
• Why the hacker, hustler, artiste combination is so powerful in startups and how to use the three skills together.
• Business plans are truly the enemy of startups or of young, innovative software companies. Business plan writing is in the wrong division of every university in the world. It belongs in the creative writing department, not the business department.
• How to use the Business Model Canvas rather than a business plan to make proper progress in a business.
• Understanding the Customer Development Process.

Video

http://businessofsoftware.wistia.com/medias/64rgllvkdj?embedType=async&videoFoam=true&videoWidth=640

Transcript

Bob Dorf: With a welcome like that the best thing I could do is go home and send you back to coffee break. I would guess about half of you woke up this morning feeling kind of like crap. I can’t imagine how that would have happened.

I woke up feeling rather unusual. I woke up feeling like a hero sandwich, not something I usually wake up feeling like.

Why hero sandwich? Well, we had one of the great heroes of startup founder thinking speaking before me. We’ve got one of the heroes of thinking and innovation speaking after me, and I’m the baloney in the middle. Actually, if you knew me better you would say I’m the shit in the middle of the sandwich.

We’re going to talk about some appetizing stuff. We’re going to try to talk a little bit about a book Steve Blank and I spent the last three years of our lives wrestling with. It’s very hard to have a debate among two people who are absolutely certain they’re always right all the time.

One hundred thousand words and 21 months later your exercise program for the Business of Software arrived, a three pound lead weight in your carry bag. [sarcastically] I’m sure each of you has read it completely by now.

Let me just very briefly, a little about me. I left a fabulous job that paid about $240,000 in today’s dollars back in 1972 because I couldn’t see doing that for 40 years. I knew I had a start… I’d started businesses in high school and in college. I knew it was time, so I quit that wonderful job, started my first of seven startups: two home runs, two base hits, three magnificent tax losses.

Along the way I earned three of these vests, which is really what gives me the privilege of standing up and talking to all you incredible people, talented, innovative, entrepreneurial. My three vests, I think, get me up here this morning.

Remember when you were growing up. Mom and Dad went off to Europe and came back and brought you a little t-shirt. “Mommy and Daddy went to Paris and all I got was this lousy t-shirt.”

My three vests are that nice, high-quality, fluffy fleece. They have a beautiful color logo on the front of some tech company or another, and on the back of each one of them is embroidered, “I invested more than a million dollars of my personal cash in this company and all I got was this lousy fleece.” With all due respect to the alpaca farmers in the world, I believe that that’s the true etymology of the word fleece. [laughter]

If you really want to get ill I’ll take my shirt off and I will show you the scars on my back. In doing seven startups and investing in more than 20 others…by the way, seven startups, zero venture capital, all bootstrapped because it’s much more fun and much higher risk profile and what the hell. Seven IPOs in my investment pile, six total wipe-outs. In my spare time teach Entrepreneurship at Columbia Business School.

This book…Steve and I, between us, not apiece, have spent 75 years doing startups. Our running joke is we took the 500 stupid mistakes we made, shoved them between the covers of this book, and hopefully people can read it and not make them for themselves.

The reality is today’s technology can build just about anything other than anti-gravity and the startup I’m still waiting for, the instant weight-loss pill, you can build virtually anything today. The problem isn’t technology. The problem is getting a business to scale.

Startups don’t fail from that technology. They fail from a lack of customers and finding passionate customers at a cost of customer acquisition that’s affordable, it’s increasingly impossible with the clutter and the variety, try building company on six dollar Google AdWords or whatever.

The way I think about startups and founders is if you are at the very beginning of the exercise, you are signing on for 20,000 hours of work over five years if you’re really-really lucky. That’s basically five years of 80-hour weeks.

You don’t remember your wife, your husband. Your dog forgets who the hell you are. If you’re a passionate, serious, committed founder, 20,000 hours is a good, a lucky number, to have a one in four or one in eight chance of success.

Steve Blank’s premise back 12 years ago was, “Why, when you start your company on day one, do you put your head down and start writing code?” Code is only half of the job. The other half of the job is figuring out who your customers are going to be, what’s going to make them excited and ecstatic and come back again and again and again, and tell their friends and so forth.

The new jargon in the Valley is “every founding team ought to have three members: a hacker, a hustler, and an artiste.” The way I like to think of it is the ‘artiste’ is the UI sort of person dealing with the customer interface functionality of your product. The “hacker” and the “hustler” are sort of obvious.

Every morning the hacker and the hustler ought to be face-to-face, the hustler saying, “Here’s what I heard from our customers today. I think we ought to change this and do this and maybe think about this.”

The hacker says, “Got it.” Then they both about-face. One guy goes all day on product. The other – or lady – goes all day on customers. Without enthusiasm, rampant enthusiasm, from your customers, the odds are overwhelming that your startup is going to fail.

If you go back to 1908 when Harvard issued the first MBA, it wasn’t for startups. It was for big business. It was cost accounting and divisional P&Ls and strategy and marketing and all these big stuff. It was all about rigor. It was all about this process for building a company where everything was tying to the box on the far right, ‘First Customer-ship.’

“We’re going to have the press party, the balloons, the champagne, on April 2. We could do April 1^st, probably not the best day.” Every date in that startup’s life backs up from April 2nd. “That means we’ve got to raise the money by January, December. That means we’ve got to be in product for this long.” “We’ve got to do this. We’ve got to do this. We’ve got to do this.”

On day one you don’t even have a clue of what you’re doing or what your customers are going to give a crap about, but you write this massive business plan that assures the investor that you know everything, that you know exactly what the customer problem is you’re going out to solve and exactly what the feature set is that will do that.

Sort of ironic when, as a startup, you have absolutely no customers. How could you possibly know all this stuff?

Those of you that are old enough to remember the days before agile development; perish the thought; if somewhere between the first two boxes your customers said, “All those features kind of suck. In fact, Adobe does them better,” or, “Oracle does them better.” “Why are you building a product around that? But gee, if your product could do this, I’d buy a hundred of them!”

If you walked into the development group and said, “Wait a minute, guys! We’re going to change. We’re going to drop this set of features. We’re going to add these,” they’d all say, “Wait, we can’t do that! Look at the business plan. It says we’re shipping on April 2nd. This is going to be a four-month delay.”

You were stuck in this lockstep march to oblivion most of the time. There wasn’t very much you could do about it because you wrote this thing called the business plan and some schmuck venture capitalist said, “Okay. This looks about as good as any other 40 cats and dogs in my portfolio. Here’s some money.”

You, as a starry-eyed founder said, “Wow, this guy’s got the money. Golden rules says if he thinks this is a good idea let’s go do exactly what this damn thing says.”

Steve Blank and I contend that business plan writing is in the wrong division of every university in the world. It belongs in the creative writing department, not the business department. How can you possibly tell me that in the second quarter of 2013 you’re going to have 82 customers? You’re going to lose 6? Four are going to reorder?

It’s just all Grimm’s Fairy Tales and there’s a hidden Easter egg somewhere in Excel. No matter what you plug in at the front end of your five-year cash flow forecast, the end of the five years, it always comes out to 100 million revenue, 15 million in profit. Those are the two numbers that get you the check.

Business plans are truly the enemy of startups or of young, innovative software companies like many of you are out there running. They put you into board meetings that are checklist conversation.

“Did you do this? Did this number match that number?”

“Of course that number didn’t match that number. I made the damn number up!”  [laughter]

“But you spoke about it with such certainty just before I gave you the check!”

“Yeah, I did, but you gave me the check. Who cares, right?”

A lot of this thinking is oriented to early early-stage startups. You folks are in the most competitive industry on the planet. Startups are painting targets on your backs every day.

Whether your company is two years old doing 20 million dollars or five years old doing 100 million dollars, if you aren’t continuing to have some of your old startup DNA hidden in some corner of your highly successful, growing – apparently – software company, we would submit that you’re out of your mind.

You’ve got to always be innovating. You’ve got to be constantly getting feedback from your customers, looking at new ways to change some aspect of your business model, not necessarily the product or the channel. “What can I change to make it even better?” If you’re going to invest those 20,000 hours, don’t you want to do something great?

If you’re talking about losing your best girlfriend, sometimes your spouse, the dog that doesn’t recognize you when you come – “What are you doing here?” – they’re barking at you when you come home. If you’re going to make that kind of an investment, don’t you want to build something that’s spectacular, that’s exciting, that’s got virility and viral growth to it and so forth?

Sure, you’ve got to do what the business plan says a little bit because you made that promise to your investors. But the reality is you don’t get to staple a copy of your business plan to every copy of your product.

Imagine the A&P. Every can of peas has a business plan stapled to it. It says, “You will buy three of these, one to wrap as a Christmas gift, one to give to your mother, and one to go in a stew. You’re going to come back in two weeks and buy two more…” and so on, so forth.

Webvan wrote this massive business plan. Louis Borders, a bunch of geniuses, 2.6 billion dollars raised before somebody realized that the business plan said that, “We were going to have about a 70 percent customer repeat rate every 15 or 18 days.”

When they delivered the groceries they didn’t staple that to the top of the box, so the customers are abandoning in droves. Three months after they launch their first pilot the repeat rate was closer to 17 percent than 70 percent. All of the economics of the business tumbled right into Chapter 11 of the history of Webvan.

Webvan was maybe a little more established than a startup, but what really is a startup? Startup isn’t a company at all, right? A startup is a band of pirates, usually on the insane side, who get together temporarily to see if they can make all of the elements of a business model come together and sing the ‘Hallelujah Chorus.’ They only sing the ‘Hallelujah Chorus’ with a very important backup band, their customers.

Startups are in search mode. They’re searching through the jungle to find that pile of gold or that magic karma of a business plan, a business model where all of the elements work together. By definition, there’s no such thing as a six- or eight-year-old startup. What that is, is a four- or six-year-old failure glued to a two-year-old startup.

Some startups stay in search mode for a year or longer, conserving cash, getting feedback, figuring out what really works well, what really sings the ‘Hallelujah Chorus.’ Then and only then do they have some facts they could use to write something that looks more like a business plan. It’s really, at that point, more of an operating plan because it’s based on facts, not fantasy.

A Swiss guy named Alexander Osterwalder about four years ago now developed the idea:” What do you do instead of a business plan?” Osterwalder developed the concept of the business model canvas, which suggests that you can summarize all of the key elements of any business, regardless of shape, size, industry, whatever, in nine boxes.

First one: value proposition. What is my product? What are the features and benefits? What is my uniqueness, my competitive distinction? Think of the value proposition as the contract I’m offering to my customers.

Number two: customer segments. Not “males over 40” or “people in finance,” but define it as tightly and as narrowly as you possibly can to see if you can connect, first and foremost, with the ones who are most excited about the problem you’re solving or the need you’re fulfilling.

Then Osterwalder, being a Swiss Ph.D. type, described customer relationships as, “When you get a customer how do you treat them right?” and all that sort of stuff. Steve and I looked at that and said, “Nope.”

We are this lifetime marketing-guy types. Customer relationship is about the three most critical activities of any business: getting customers, keeping them, and growing them by selling them more and by getting them to enthusiastically refer you to other customers.

I’m not going to go through all nine boxes because I know the time is short. Those four are really the key. Who is my customer as defined as tightly and as narrowly as I can, and what is my promise or pledge or contract I’m offering to that customer?

Give me money, give me page views, give me postings or whatever – slides are kind of having a mind of their own, but we’ll deal with that in a minute – when you put those two together you have what Marc Andreessen first called “Product Market Fit.”

Product, obviously, is the product. The market, the customer segment. When you have that, you can start to think about “How does money spill out of that?” and, “Does it spill out enough money to make this business worth doing?”

When you’re done with this elaborate nine-box model, all you really have is a bunch of guesses. When I teach this at Columbia where I could have never gotten in as an applicant, I could have only gotten in with an Uzi… [laughter] If you saw my transcript you would know I was doing something else and it was the ’60s, but it wasn’t that, it was starting businesses.

When you finish this business model, all you have is a bunch of guesses. “We think we can sell this through AdWords. We are sure people are going to click on this box and do this and do that and come back once a week”.

What do you do with your guesses? That’s where the customer development process picks up. It’s “How do I take those guesses and turn them into facts?” Since there are no facts whatsoever inside your office building you’ve got to get out and talk to your customers and understand whether they agree with your guesses or not.

What you see over here – hopefully it’ll stay up here for a minute but it won’t – is the four steps of the customer development process developed about 12 years ago by Steve who did eight startups.

Number eight, Epiphany, an incredible data mining software company where I actually met him when I was running One-to-One marketing. Epiphany, at a moment in time reached a market cap of eight billion dollars, something I wish on each and every one of you.

Was a little better performance than number seven: Steve Blank, CEO, Rocket Science Games. Great team, raised 35 million dollars. Everything was wonderful. The technology, the marketing, the sales were all fabulous. The game sucked and they blew through 35 million dollars in Silicon Valley record-setting time: from the cover of ‘Wired Magazine’ to selling the office furniture, under two years.  [scattered applause]

Steve took a couple of weeks off, kicked the dog. He doesn’t drink, otherwise I know what I would have done after blowing through…Then there it was. Do it again. In fact, when he called his Russian immigrant mother to tell her that they had just lost 35 million dollars for a bunch of investors, she said: “Oh, Stevie. Did they make you give it back?”  [laughter]

His answer, of course, was vintage Silicon Valley: “No, Mom, they want to give me more money to do it again.” [laughter]

Her answer: “Only in America. Their streets really are paved with gold.”

We’ve stopped writing about the third and fourth step, the second half of the customer development process. Whether you’re a startup or an established company trying to maintain or enhance your edge, the first two steps are just so powerful for you.

The other two steps, they teach that stuff in business school. If you have a business, go on like this, either making money or on the cusp of making money. You can usually find some pretty credible established people to keep it going.

In fact, as Dr. Wasserman said, “Very often your investors will go and find them as they show you to the door.”

Page 83 of the “How to Be a VC” handbook says, “You can hire anybody to be a founding CEO, but once it looks like he or she’s got a success on their hands, we’d better get a real CEO in there because we might actually make some money on this one.” It’s almost an axiom that the CEO goes, as Noam said earlier.

The first two steps of customer development, discovery and validation, work as well in a corner of your existing business as they do in the first days of your three- or five-person startup.

Take a couple of people. Give them some time to go and talk about the current business model and create other business models and see if any of them look any better.

If you think on instinct that, “Gee, maybe Freemium with aggressive free-to-paid conversion would be an interesting alternative,” or “Maybe we should sell this in supermarkets – after all, there’s lot of software sold in supermarkets.” “How about gas stations? We have yet to find any software in gas stations…”

Test it. Go out and talk to some gas stations or supermarket managers. Most important, talk to customers. “Hey, we’re just inventing this wonderful super-duper new menu planner. Would you like to find it where you find your groceries?” and so forth.

Those first two steps take you through the process of validating all of those hypotheses or guesses that you’ve made in your business model. You don’t expect every customer to answer or know about every one of the boxes. Some aren’t going to know about the channel they’d like to buy it in or they’re not going to understand how their company would pay for it and so forth.

If you’re not a startup, take a team. Give them a week off or five days off. Run this process within the bowels of your own company. Not because there’s something wrong, not because you have a problem, not because your VCs told you to do it, but because you want to see.

Is there a way to crank that needle up one more step? Let’s go out and see if our customers would like us better if we ‘blank’ or our channel partners would like us better if we ‘blank’, or whatever. Just take some time every couple of months and take a deep breath and lead it from the front.

If the founder isn’t engaged in this it’s a project. It’s a toy. If McKinsey does it for you, number one, all you hear are the last three pages of the presentation. Under the ‘Consultants Full Employment Act of 1996’, the last three say: “The next project will be this and it will accomplish this and will build on the great success of this project.”

Customer discovery begins in a funny place. Eric Ries writes about this. Eric was arguably the brightest student that ever set foot in Steve Blank’s classroom. Eric and his founding CEO, who had worked for Steve before, said, “We’d like you to fund our IMVU,” the poster child for lean startup.

Steve said, “I’ll fund it under two conditions. I just finished this book about customer development. The conditions are you’ll read the book attend my full-semester class at Berkley on the customer development process. Number two is you will use it to build your company.”

They build minimum viable products, got them up and got them tested, got feedback. To me one of the most magnificent pieces of feedback they ever got was, “I love it. I use it every day. Does it really have to crash my entire system every morning when I log on?”

“I know it’s going to crash, but I know I really want to use it.” That’s a devoted, passionate customer. Obviously there are a few bug fix reports that needed work and so on and so forth.

Whether you’re doing this within your company or in a raw startup, as quickly as possible, you want to have something that customers can play with and react to, even if it’s a non-working, dummy day-in-the-life demo, or it’s a simple three- or five-page website that doesn’t look much prettier than a bunch of wireframes.

The reactions you get from customers are so much more genuine and so much more helpful when they’re looking at something and playing with something, than when you’re telling them about it.

My favorite example of a minimum-viable product is a company called ‘Diapers.com’. Any ‘Diapers.com’ customers in the room? Any unhappy ones? I haven’t seen one yet.

Two entrepreneurs who had just cashed out modestly from a startup they did together both had young kids.

They said, “We’ve got to do something. This diaper business is messed up, even before they get used. Why don’t we try an incredible value proposition for customers? Cheapest price on earth, guaranteed next-day delivery, spectacular, elegant, customer service 24 hours a day, absolute, unconditional returns at our expense, no questions asked, and we have every size, configuration, shape, form, and aroma of diaper available for you any time.”

Pretty complex business model – warehouses, forklift trucks – that’s hard to do. How do you do a minimum-viable product for a warehouse delivery ecommerce company?

They did it without buying a single diaper. They built a little website in about a week, nothing fancy. “Here’s all the Pampers. Here’s all the Loves. Here’s all the generics. Pick your size. This is the price at Safeway, Wal-Mart, Costco, our price, delivery free.” Now they start taking orders. They have no diapers. You could go to jail for stuff like that.

Every day at four o’clock in the afternoon they got into SUVs and vans and one went to Costco, one went to Walgreens, and one went to Babies R Us. They picked and packed their orders off the retail shelves and shipped them out to customers. Brought them back to headquarters – the house – wrapped up the orders, put the labels on them, and at 6:29 every night they got them to the UPS loading dock for guaranteed next-day delivery.

They lost money on every single order because they were buying retail, selling at lowest price anywhere, and free shipping to make it even worse. They weren’t really testing the profitability of their model. They were testing whether customers give a shit. Customers and the end users…the end users certainly give a shit which is why the parents do.

They were deluged, DELUGED, with orders. It got so bad that they had to rent step-vans because they cleaned out Babies R Us and Wal-Mart every day, to the point where the Wal-Mart manager in town actually yelled at them. “How can you buy all these diapers? You’re not leaving any for anyone else.” “Well, it says ‘for sale.’ Am I doing something wrong?”

In a matter of a month or two, they had proven an incredibly powerful business model or contract that they were offering to their customers: “Shop with us. You’ll always get what you need. You’ll always get it tomorrow. You don’t have to carry these big bulky boxes around. Guaranteed best price. If you order the wrong one or whatever, send it back, no problem.”

“The only problem in our business model is every sale we make we lose money. But we can make that up on volume”. That’s what my Polish ancestors would say.

The reality was they were testing the model. The incidental loss per sale was a marketing investment to see if they really had a business. Within a few months they were getting warehouse space, forklift trucks, automatic labellers, and wholesale contracts with Proctor & Gamble and Kimberly-Clark and so on and so forth. Within just a little over two years they sold the business to Amazon for plus or minus a billion dollars.

Zero capital investment, a seven dollar a month Intuit website to start. The SUVs and stuff they already owned. They were testing their consumer promise or their value position, and before you know it, the consumers say, “Yeah, we really like this promise. We’re in.” Then it was, “Get out of the way.” Execute, execute, execute.

Google really did it exactly the same way. I had the great privilege of working with Sequoia people on a couple of other companies some years ago. It’s one of the most magnificent, insightful, venture funds on the planet. They were the first investor in a small company you may remember called Hewlett-Packard, Apple, Yahoo!, Google, the list just goes on and on.

Sergei and Larry walked into their office one day, showed them this algorithm. At that moment there were 14 search engines on the net, and these two young crazy guys had number 15.

They said, “We’re not sure how we’re going to make money, but look at the number of people that are coming to use our search engine every day and this dramatic increase. Every time we open another language the pattern repeats. Customers are loving our search engine.”

For those of you who are too young to remember, the big difference was it presented results based on what customers wanted to see and what they wanted to click on as opposed to what advertisers wanted them to click on and so forth.

The Sequoia guys said, “Okay, couple of things. Close the door. Three things we’ve got to tell you. Number one: you cannot leave our office today without a check. Number two: if you guys would like to put some money in your pockets because this is going to be a long haul, we’ll buy some of your personal stocks. You can each have a couple million bucks to buy stuff with while you’re working hard. Number three: we don’t want to think about how you’re going to make money for two years. We want you to be the overwhelmingly dominant search engine, because at that point, the dog sleeping over there in the corner of the office can figure out how to make money.” I think the dog did a pretty good job if you ask me.

How do I find a test for a new unit of my business, a new idea or a different way to market my business? Isolated from the rest of the business, because you’ve got numbers to hit and people to pay and bills to pay and stuff like that, and do just this kind of testing within your existing successful organization?

If you’re a startup, even better. Most of what I’m talking about can be done with tiny amounts of money, and the further you can take your idea along the spectrum from day one to customer discovery to customer validation, the more equity you’ll preserve for yourselves and your founding team.

On day one it’s, “Hello Mr. Investor. We’re a bunch of really smart people with fancy degrees and lots of experience and we have a great idea.” Sounds like the golden rule applies. The guy with the gold makes the rules on valuation.

At the end of customer validation: “Hello Mr. Investor. It’s us again. You didn’t fund us two years ago when we were bright people with a great idea, but now we’ve invented this thing called a money machine. You put a dollar in the top, it kind of bounces around like a Pachinko machine, makes a few turns, clink, clink, clink. Customer gets a product and we’re left with 17 cents or 14 cents or 12 cents. We have now tested that in the customer validation process three, four, five hundred times. You know what? On average, every dollar goes in, 17 cents comes out of the bottom. So Mr. Investor, we’d like you to give us some more dollars to put through this machine.”

Think of the valuation dynamics in that conversation versus in the conversation, “I have bright people with a great idea. Line forms over here.” The more you can do to prove and validate, not with yourselves, not with your investors and board members, but with your customers.

This is where they want to buy it. This is the feature set they want. This is how often they want to use it. This is what will encourage them to refer you to their friends. They’re going to find you on Google or at trade shows or at street fairs or on the side of a blimp or whatever.

You want to get to this point of minimum-viable product as fast as you possibly can so you get honest feedback along the way. That’s what the customer discovery process, the first step, is all about.

You go out and talk to customers and say, “Look at this thing. I know it’s made of Styrofoam and there’s a fork stuck in the side and a light bulb over here and a whole lot of duct tape, but this is what it’s going to do for you. What do you think?”

When the customer says, “It sucks,” you’ve got to resist your natural genetic temptation as an entrepreneur to say, “Oh, no! That’s a grade A fork and that’s a 100-watt light bulb and that’s the best duct tape money can buy!” Your natural instinct is to sell and to correct them and to tell them about all the great features.

What you’ve got to do – for me to do this, to stop selling and start listening – is to take a pill every time I go out on a customer discovery call.

“It sucks.” “Why? What other product does a better job? Is there anything we can do to make it not suck? Are there other ways you solve this problem now? If it didn’t suck so much how would you find it? Would you find it at a trade show, at a street fair,” or whatever.

You are doing this on a continuous basis. It’s most powerful in the early days of your business or your testbed within your company because you can really effect dramatic change. It really never stops in good-quality companies that trying to shoot for the moon.

If you’re thinking, “This guy’s talking about startups. My company’s seven years old. We’re doing great.” If you’re happy, good for you, congratulations. If you’re complacent you’re probably really not an entrepreneur, so find a way to apply this same thinking within your successful company. Invariably, along the way, it will lead you to a pivot.

‘Pivot’ has become the word du jour in ‘startup land’. In fact, this year it was enshrined in the perfect place. ‘The New Yorker’ magazine had this wonderful cartoon. Very well-dressed man and woman sitting at a table for two in a very elegant cocktail lounge and she says, “No, I’m not leaving you Henry. I’m just pivoting to another man.”

You know that the pivot is an established, important part of our lexicon. What a pivot is, is not when two or three customers tell you, “This sucks,” or “I’d only buy it if it was blue,” or, “I really need this feature.”

If you start to hear that consistently from 20 or 30 or 40 customers, then you say, “Okay. Full stop. Let’s go back to the business model. What do we need to change to get customers excited?” Then you go back out, talk to those customers again, and see if your better idea is voted as better by those customers.

Pivots happen in big companies all the time, but they are usually consummated with a firing. Usually VP of Sales goes first. The new VP of Sales comes in, has to change everything because whatever the last guy was doing got him fired. Then if the new VP of Sales starts to struggle his finger points to the VP of Marketing. She gets fired. A new one of them comes in.

Every one of them is a crisis that doubles the consumption of coffee in the company because everybody’s standing around, “What do we do? Who’s the new guy? Where are they going to come from? How are we going to change? What are we going to do different?” and so forth. It’s a crisis.

In a startup, it’s almost a celebration. “Wow, we just found a way to get more customers excited about our business, more customers referring us to others. We found a better way to reach them, more cost-effectively or faster, or a better way to renew them or convert them from ‘free’ to ‘paid’.” The only way you get this kind of feedback is by talking to your customers.

How many CEOs do we have in the room? Probably half? This is not something you can delegate. If you are not doing this yourself and leading it from the front, you’re not changing the culture of your company to state, “Customer feedback is vital to our business. It’s as vital to our business as oxygen.”

You are also the only one that’s empowered to go out there and hear two or three customers say, “If it had the auto-pay and store feature I’d like it so much better…” You’re the one who can come back to the office and say, “Hey, guys, I know you got a long build planned and you’re really busy for the next three months, but we’re changing it. Timeout. Take these off. Build auto pay and file because I’ve heard that from enough customers.”

If you send 25 year olds out to do this kind of discovery, which is your natural tendency as an optimizing kind of CEO, they come back and talked to 12 customers. Eleven of them say, “Eh, kind of sucks,” and the twelfth one says, “Not for me.”

How are they going to present that to you? First, they’re going to stop, dress up their resume. Then they’re going to find some nuggets of good stuff somewhere there. They’re going to polish those turds until they glow. You’re not going to get the truth.

Same problem with McKinsey or Acme Market Research or any of the online research tools. You’re going to get lots of threes and fours and you’ll say, “Wow, we got a 3.2 this time. It was only 2.8 last time. Isn’t that a great improvement? That’s a 20 percent improvement. It went from ‘sort of sucks’ to ‘sort of sucks.'”

You don’t get good answers. You don’t get honest feedback. Most important, you don’t get the sort of outlier, corner-case ideas that sometimes fuel your entire business.

There’s one problem with pivots, and that is ‘premature evacuation’.  ‘Premature evacuation’ happens. I saw it two weeks ago in Bogota where one of my startup teams said, “We talked to three customers. They all said the idea sucks, so we’re going to do something else.”

I said, “Three customers? You call yourself an entrepreneur? Three people told you it was bad so you’re giving up? You must do wonderful picking up men or women in a bar. Go talk to 20 more, and if 20 more people tell you it sucks, then we’ve got something to really talk about.”

Are you pivoting based on enough data, consistent and valid from your key, core customer segment, the people who you most want to be most excited about your product? You have to do this quickly and with agility. These pivots have created billions and billions of dollars in market capitalization.

Two quick examples: Steve Blank’s pivot that took Epiphany from basically worthless to eight billion dollars in about two and a half years. He and his partner, Ben Webwright. Ben is the introverted engineer, but Steve said, “Hey, you’re a founder. You’ve got to go on some of these customer discovery calls with me, otherwise you’re not going to hear it from the horse’s mouth.”

You all know the difference between an introverted engineer and an extrovert, right? Introverted engineer looks at their shoes when they’re talking to you, and an extrovert looks at your shoes when they’re talking to you.

They’re going to Charles Schwab for the eighth meeting to try to sell them this Epiphany software which is this incredibly robust data-mining tool that could reach into every data silo in a company and roll all those disparate data sets up into one precise profile of the customer.

What do they buy? What do they sell? Do they make money with us? Don’t they? What percentage of their assets do they have with us? How much is their house worth? What’s their equity in their house? To get a complete picture and then build a one-to-one marketing campaign to that individual customer based on that comprehensive aggregation.

Back in 1995, you can do this with a computer you can buy at Wal-Mart in 1995. I assure you it wasn’t anywhere near that easy.

They’re at their eighth meeting with Charles Schwab, one of the more innovative online brokers first to move into the online world ambitiously and so forth. Epiphany knew that the financial services segment was a huge piece of their business, and since one of the people from Schwab was on the Epiphany advisory board with David Pottruck, the CEO’s, blessing, they were sure they had this deal done.

They’re telling J.P. Morgan and E-Trade and everybody else, “We’re signing a Schwab deal next week. In fact, Ben and Steve are going there Thursday to sign it.” They go through the demo and Mary Kelley’s brilliant database marketer says, “Guys, this is our eighth meeting. It’s the eighth and last time I’m going to tell you that we’re not buying your product unless you have data householding.”

‘Householding’ is a very common term, which is, if I live in the same house as my wife odds are we make investment decisions together. Therefore if you send her a stream of emails and postal mails saying, “Do this,” and you send me a stream saying do something else because our patterns and behaviors are different, you, the marketer, look like a total asshole and why would I trust you with my finances if you don’t know that we both live in the same house, sleep in the same bed?

You take the two piles of data, integrate them, merge them into one and make one conclusion and one marketing campaign. Ben and Steve leave the meeting. Get in their car to drive 45 minutes to get to Steve’s living room, which was world headquarters for the six non-developer employees.

Halfway back, Ben doesn’t say a word. Halfway back Steve says, “What are we going to do?”

Ben says, “What do you mean what are we going to do? What are we going to do about what?”

“Ben, we just blew the whole company. We’ve got nine other clients assured that today we’re selling Schwab and they said if Schwab buys we’re in. We would have to have it. That’s our whole business, it’s our whole future.”

Ben says, “Didn’t they read page six of the product spec?”

Steve says, “No. It’s a five-page product spec. How could they read page six?”

Ben says, “It’s page six. There’s page six now.” Page six, data householding, got stapled to the product spec.

Then they had the tough conversation with the 14 developers in the rented house next door: “Hey, guys, we know you just worked around the clock for three weeks to get ready for the Schwab presentation. We need three more around-the-clock weeks because we’ve got to do page six so we can get the damn order.”

Pivot in the business. The customer needs these features and in that case will give us a million dollar check for them that unlocks 20 million dollar revenue potential for us. Who cares what the build plan says? Who cares what we told the investors? The investors don’t care if we do what it says on page 14 of the plan. They care if we’re successful and knock down big orders. Stop everything, big pivot, and go on.

The sexiest one is probably Groupon which was founded, as you may know, as The Point some years ago. The Point was kind of getting fairly run down. It was more like a nub. The Point was basically about bringing groups of people together to do good work, go paint a church, go clean up a park, go help the homeless, do good noble deeds in Chicago.

They had a huge email list and they thought they’d be able to sell advertising and ads in the email and sponsored events and so forth. That was their revenue model.

About six months into it, yes they had lots of names, but they had no revenue. They’re sitting in the pizza joint in the ground floor of their elegant slum headquarters in Chicago saying, “What are we going to do? We’re almost out of money and this business is clearly not going anywhere. What are we going to do?”

While they’re having this conversation, one of the guys wanders over to Joe, who owns the pizza joint, and says, “Hey, Joe. It’s Thursday lunchtime and we’re the only people in here and you know we don’t spend a lot of money with you. We even bring our own drinks from the fridge upstairs. How’d you like it if we sent an email to everybody in this area of Chicago and said, ‘Next Thursday is two for one pizza day at Joe’s pizza? Come in with at least two people. Buy two pizzas, get one free.'”

Joe says, “What the hell. The place is empty and you guys don’t spend any money. I’ll try anything.” Groupon, as we know it today, was born. It’s a shame that the much more noble Groupon faded away, but congratulations to a bunch of passionate entrepreneurs.

I think that 12 billion dollar number needs a rather dramatic adjustment based on its behavior of late, but it’s a company that’s lost a lot of its startup mojo and passion. It is struggling to innovate and test new things and so forth while its customer base gets increasingly bored and inactive. Can happen to anyone in this room.

As long as you are careful about ‘premature evacuation’, you want to pivot as quickly as you can because in a startup and in most of your companies you are chasing the ticking time bomb of the declining bank balance. It’s declining at a steady, often increasing rate. The faster you can make these pivots, the more pivots you will have money in the bank to fund.

I don’t believe there is a company on this earth that hasn’t pivoted at least a half a dozen times on its way to success.

Facebook was sex for geeks. It’s a lot more than that today. Some of the features remain the same. Lots of them are very different. IBM, I think, started as a meat-slicing manufacturer. NCR, owner of Teradata, started as the manual electro-mechanical cash register and on and on and on. As long as you let your customers drive the bus.

You do all this customer discovery stuff – I’m not going to spend much time on this – but how do you ever know when you’re done? The real answer is you are never done, but you’ve got to take a pause and execute and make sure you’re hitting your numbers and doing your stuff.

We do a modest amount of consulting work with big companies trying to figure this out. We had the privilege over the last two years to work with a raw startup within General Electric, number three on the global Fortune 1000, 100 billion dollars in revenue, founded 100 and some years ago by Thomas Alva Edison, six sigma black belt. Execution, execution.

This startup was 18 people in a slum with some new technology that GE had purchased in the industrial battery area, an environmentally friendly, high-performance battery. They had convinced their senior management that they needed to do this crazy search thing without the six-sigma Gantt list and punch list and charts and timelines that are GE’s hallmark.

We’re working through the business model over a month or so, and the CEO said, “the one difference between that particular startup and virtually any other is they had 100 million dollars in seed capital and a one billion dollar revenue target for 2016. Other than that, it was just like a startup.”

I said, “How do we put a little six-sigma in this wandering business model search thing?” I drew these little circles. You can’t see what it says in the circle, but what it basically is, is a proving ground for your business model. If you are absolutely dead sure that you are right on your value proposition, the center box, the light turns green.

If you’re sure you don’t have a clue yet, the light stays red.

In the middle of the circle is a number you can’t see. In that case it’s 26. My certainty is based on 26 face-to-face customer conversations where I have consistently heard the same feedback, that this is why they’ll buy it, this is how they’ll buy it, this is why it’s better than a lead-acid or a lithium-ion battery, and so on and so forth.

All the lights don’t have to turn green before you push the ‘go’ button, but certainly the ones on the right side, the customer facing. What’s my product market fit promise, who are these customers, and how the hell am I going to find them? As you move through this process of customer development, you use Osterwalder’s canvas as your scorecard to know how close you’re getting to done.

When you feel like you’ve got it, when you feel like you’ve tested the business model to the point where you have some reasonable confidence that this could actually work, you come to the end of the first step, customer discovery. You’ve talked to at least 50, if not 100, customers face to face, whether you’re a physical channel, physical-goods business, or a web mobile business.

If you’re a web mobile business you talk to another 500 or 1000 through the medium they’re going to learn about you or buy you from because you can do it. You can do it inexpensively through chat lines and surveys, online focus groups and things like that to increase your ability to reach out to them.

At the end of discovery it’s time to say, “Okay, we have proven that for every ten sales calls we make we get four expressions of interest, and two of them turn into orders within 90 days.” Ten sales calls for two orders short-term. That math works for us. We can afford to make 10 sales calls to get two orders when we’re doing 20 sales calls or 40 sales calls in the customer discovery early stage.

Now what happens when you start to turn on the gas a little bit. All of a sudden you have sales people who aren’t founders, who aren’t as good as you, who aren’t as passionate, aren’t as knowledgeable. Those percentages change. If they change, your ka-chunk, ka-chunk, ka-chunk money machine doesn’t perform as well as you thought.

Customer validation. Think of it as putting gas in the tank of a race car with a turkey baster or an eye dropper. A little bit at a time. Let’s see, when we did 250 dollars’ worth of Google AdWords what happened? Let’s try a 1000. Let’s try 2500. Let’s try 5000 and see if the same percentages hold up, if the same customer quality and stick rate happen. Let’s make sure.

Again, try and conserve your most precious asset, your cash. Make sure that all your guesses continue to get validated by customers. The kinds of customers you want to find early are these most passionate customers who are desperate to have your product.

If you’re selling Band-Aids find bleeding people on the street. They’re not going to care about the packaging or the logo. They’re not even really going to care about the absorption. If it’s a Band-Aid and it will stick, that’ll do.

Find customers. The customer in the finance department who’s got two feet of paper at her desk, and is the accounts receivable person. You have a way to automate collections and accelerate or reduce average days outstanding. That lady could be one of your most exciting early customers.

More important, she’ll take your product with the bugs and the flaws because she’s desperate to solve her problem. Even more important than that, she will be on the phone to you regularly saying, “You know, what would work really so much better would be if you did this or dropped that feature or you moved this one forward or you redid the UI.”

She wants that product to be better just as badly as you and your investors do. This is a great exercise to conduct no matter how happy you are with your business today. Go out and talk, not just to the buyers, but to the end users.

“What do you like? What don’t you like? I just want to sit here for a half hour and watch what you’re doing, and I’m not going to save you. No, you should have hit that button or clicked over there.”

I’m going to sit there with my notepad and say, “Why did you hit that button instead of that button? Oh, I see, we have a UI problem. We know how to fix that,” or, “Why’d you click over there instead of here? Everybody’s supposed to click over here.” “Well, you didn’t tell me that.” So you just put the words ‘Click here’ on the screen.

That kind of feedback never stops fuelling your incremental success to your existing business.

We have any accountants in the room? I hope not. This is not a talk for accountants. This is a talk for crazy entrepreneurs, passionate entrepreneurs who if business is great, if the money is looking great, still say, “I’m here to make this fabulous.”

What’s the delta between today and fabulous? The only one who can tell you what that delta is, is your customer, your abandoner, or your prospect who have different views of what’s great or not great or whatever about your product.

The reason why startups aren’t run by accountants is illustrated by this rather fun story. I suspect most of you know who Alfred P. Sloan is. He’s thought of as the founder of General Motors. He’s really the father of modern business today. In fact, he was the first hired-gun CEO.

This guy was the founder, Billy Durant. Billy Durant was the number one buggy maker in Flint, Michigan. He had high-end buggies, middle-of-road. He bought up every buggy company he could find. He’s selling buggies until he went buggy. One day he’s sitting in the bar with his buggy-making competitors and they hear this noisy, rattletrap thing coming down the street in Flint, Michigan.

He goes out and he looks at it and he goes, “Holy Crap.” He very quietly disappears from the bar, sells all his buggy companies, starts buying transmission companies, auto parts companies, metal-stamping companies, engine companies, and he founded this little company known as General Motors.

He grew General Motors so fast that in a small handful of years it reached 2.6 billion dollars in revenue in today’s dollars, at which point the board of directors that had hired him and funded him — didn’t hire him because he founded it, but they funded him and championed him along the way – fired him.

We can’t have this wacky, crazy, unpredictable entrepreneur with a silly hat running a 2.6 billion dollar company. Let’s get an accountant, an Alfred P. Sloan, in here, and let him run the company. They brought in Alfred P. Sloan.

What does Billy Durant do? Does he go off and count his money and hang out and have a good time? No. He takes some of his money, finds another guy – automotive engineer by the name of Louis Chevrolet, you may have heard the name – buys Chevrolet, uses it to collect another collection of auto companies, parts and assembly and so forth, builds a company called Chevrolet, takes it public, uses the IPO proceeds to buy General Motors and fire the board of directors that fired him. [laughter]

It gets better. They bring in the DuPonts as lead investors. The DuPonts invented, among other things, gun power, fairly important back then. They put a whole lot of money in the company and they say, “Wait a minute. There’s one thing wrong. We’ve got this wacky, crazy, unpredictable entrepreneur and in a silly hat,” and they fire him again.

Alfred P. Sloan dies rich, honored, and famous. Sloan School at MIT, Sloan Kettering Cancer Center – I don’t want to tell you how many times I’ve visited there but I’m still here talking about it – generous philanthropist, honored, books written about him, and so forth. The accountant: rich, honored, and famous.

Billy Durant died managing a bowling alley in Flint, Michigan.

Ladies and gentlemen, that’s the accountant. [circles Sloan’s picture] This is us. [circles Durant’s picture] Ballsy, crazy, wacky, passionate entrepreneurs. Thank you very much.