Lean Startups

Graham, Blank and Ries on MVPs

Three very smart people talking about the same thing from three slightly different angles:

A Quantum of Utility

We advise startups to launch when they’ve added a quantum of utility: when there is at least some set of users who would be excited to hear about it, because they can now do something they couldn’t do before.

-Paul Graham

Minimum Feature Set

The reality is that the minimum feature set is 1) a tactic to reduce wasted engineering hours (code left on the floor) and 2) to get the product in the hands of early visionary customers as soon as possible.

-Steve Blank

Minimum Viable Product

…that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.

-Eric Ries

To riff on the above, Brant Cooper and I would only remind you that when thinking about Minimum Viable Products:

…viable is not limited by an external determination of success, but rather is framed by the entrepreneur’s objective (user scale, specific functionality, payment) as measured by specific “currency” (usage, problem solved, money).

-The Entrepreneur’s Guide to Customer Development

Minimal Viable Products and Gall’s Law

Swapping out the word “system” for “product” in Gall’s Law gets us:

A complex product that works is invariably found to have evolved from a simple product that worked. The inverse proposition also appears to be true: A complex product designed from scratch never works and cannot be made to work. You have to start over, beginning with a working minimal viable product.

Note:  A minimal viable product may or may not work.

Keep this mind as you build and design your MVPs.

If you are pre-Product/Market Fit and you aren’t actually “Getting out of the Building” (actually talking to your customers), you aren’t doing Customer Development, and your startup isn’t a Lean Startup.

Let me repeat that:  If you aren’t actually talking to your customers, you aren’t doing Customer Development.

And by talking, I mean speaking.  With your mouth.  Preferably in-person, but if not, telephone/VOIP works as well.  (For the record, IM-ing does not count as talking/speaking.)

In The Entrepreneur’s Guide to Customer Development, Brant and I write:

Note that analytics, surveys, and other automated user-facing testing tools are complementary to, but not substitutes for “Getting Out of the Building.

Which is another way of saying that pre-Product/Market Fit, one good customer development interview is better for learning about your customers/product/problem/solution/market than five surveys with 10,000 statistically significant responses.  Obviously I am being a bit glib here, but I don’t want to understate what I see as a real problem in the adoption of Customer Development processes.

Unfortunately, this point seems to have been missed by a more than a few Lean Startup enthusiasts, and it appears to me, very often by technical co-founders of bootstrapped startups.

I have been pondering this for sometime and recently had an great conversation with my friend, Rich Collins, founder and organizer of the Lean Startup Circle, a group that meets in San Francisco to hear Lean Startup founders and employees share their experiences.

Rich pointed out to me, something that is obvious in hindsight, about why people with engineering backgrounds may unconsciously avoid actually Getting out of the Building.  Given that I am not an engineer, and Rich is, I thought I would share his many insights with you about engineers and Customer Development via an instant message interview I conducted with Rich a few days ago.


Patrick: You said something to me at the last LSC event that was quite interesting.  When I asked about why it may be that technical co-founders don’t like to Get out of the Building, you responded along the lines of:  engineers don’t want to go and talk to customers, that is why they became engineers.  I think you hit the nail right on the head.  Can you expound on that some more?

Rich: Sure.  I can mostly speak to my own experience.  I can’t say that I became a software developer to avoid talking to people but I can say that it is really easy for me to sit down for a long time without talking to people. And that I’m more uncomfortable talking to people that I don’t know than most people.

So what attracted me to software is that I thought it was fun but what kept me away from talking to customers is that it was way outside of my comfort zone.

To summarize, it’s easy for some people to become engineers because it’s fun and it isn’t a problem that you don’t have to talk to people.

Patrick: What are the ramifications of that?

Rich: And if it were easy to talk to people, they might not have become engineers.  In addition, the more time you spend doing software, the less time you spend interacting with other people face to face, so your communication skills get rusty.  Scott Adams made a recent post about how you shouldn’t go out and socialize after a long day of programming as you use your brain in a fundamentally different way.

He described talking to people after doing a lot of development like talking to someone that was at the end of a long hallway.  You feel quite disconnected from the real world.  So that makes something that started out difficult even harder.

Patrick: Do you find that to be true?

Rich: Definitely.

Patrick: In the CustDev context, how would you hack that?

Rich: One thing that I do is batch things together; I don’t switch back and forth.  Right now I’m in interview mode, so I’m spending all of my time talking to people.

Patrick: So batching tasks and avoiding context-switching works well.  Do you carve out a full day for that?  Or a few hours?

Rich: It’s easier than spending the morning programming and then talking to people in the afternoon.

I try to fill my schedule if there is some downtime I try to spend it finding more people to interview, but sometimes there is development that has to get done. So I work that in, but it isn’t like a typical dev day where I might go 6 hours or more straight without using my voice.

I don’t have a hack to getting started.

Patrick: Maybe JFDI?

Rich: Yeah.  I just had to feel the pain of not doing it enough.  Spending years failing helps you get over those types of fears.  Eventually you just say f*ck it.

Patrick: LOL.

Rich: So I don’t have as much trouble with getting over the fear of talking to people, I still do when I have to talk in front of lots of people because I don’t do it that often.

Patrick: What would you advise to a technical co-founder at a bootstrapped startup who is into Lean Startups? But isn’t actually “getting out the building”?

Rich: They really need to be convinced of its value. I would say to try it out for a week and see if they don’t generate a lot of insights that they wouldn’t have otherwise.

Patrick: What is the value specifically for an engineer?

Rich: I think you need to be aware of the value of the product that you are building if you don’t have those conversations first hand, you might have a hard time understanding what your goal is, unless you have the need specifically.

Patrick: Well-put.

Rich: That is why so many developers build developer tools. It isn’t a very good market, but at least they understand it.  The GitHub guys did succeed building something for other developers.  They didn’t really need to push themselves to get outside the building, since they were already going to dev meetups and working with devs, but there are only so many opportunities like that.

Most of the opportunities go unfulfilled because the people that can actually build the solution don’t know about the problem.

Patrick: I have heard that observation before.  Going back to the CustDev tactical level, what is most discomforting or fear-inducing part of getting out of the building?  And does it get easier for you?

Rich: The act of trying to convince people that you don’t know to share time and information you have to sell them on meeting with you and giving you valuable information.

Patrick: How do you accomplish that?

Rich: Many people have trouble talking to other people at a party, let alone asking people that you have very little connection to you for their time. I’m currently working on a tool for attorneys and everyone knows an attorney, so I just asked friends and associates if I could interview any attorneys that they know. I described to them what I’m doing, told them that I wanted to interview attorneys so that we could build the best product possible for them and let them know that we aren’t selling anything and that it would only take 15 minutes.

Patrick: And what has the response been?

Rich: Most people are quite willing to help out.  I’m not sure why at this point.  I think if you ask people for a favor, they will generally help you out at least once.

Patrick: Even if they are strangers?

Rich: I don’t solicit total strangers, I ask friends and associates for referrals, the people I end up interviewing are strangers, but I have some social proof by going through a friend.

I typically find that the people that I interview like having their opinion valued. They seem to enjoy it.

Patrick: Right. Do you do your interviews in-person? Over the phone?  Over IM?

Rich: I do phone and in person.  It’s harder to gauge emotion over IM and its easier for people to focus their attention elsewhere. I always try to do it in person.  I do phone if that isn’t possible for some reason.

Patrick: Are you actually telling me emoticons aren’t emotive enough for you for CustDev interviews?

Rich: Heh. Yeah it’s hard to tell what people are thinking over IM. Even though I prefer IM for most communication.

Patrick: Do you do try to get people in person?

Rich: I just ask if I can have 15 minutes of their time to show them our product and get their advice.

Patrick: At their offices?  Coffee shop?  Homes?  Local hookah bar?

Rich: I tell them I’ll meet them wherever is convenient for them.  So far it’s been office or lunch. When I was working on Stylous it was coffee bars, since that is a consumer product.

Patrick: Would you say that there is value in observing the emotional state and body language of a person during a custdev interview?

Rich: Definitely! I was interviewing a guy on Clientite and he was trying to tell me that it solved his problem but it was obvious that it didn’t excite him at all.

Patrick: What gave it away?

Rich: I’ve mostly been interviewing corporate attorneys.  It doesn’t seem to excite any of them so now we’re focusing on firms that have a high caseload where each case doesn’t have a high value.

Patrick: So he was verbally saying one thing, but body language/voice-inflection was telling another story?

Rich: Yeah his voice didn’t reveal any real interest.  Same with issues that we had with Stylous.  You can tell when people aren’t interested in something.

Patrick: Were you able to segue into anything that interested him?

Rich: Yes, but instead of pivoting we’re going to keep looking for people with the problem that were solving since were early on in the process.  So I now know more about what corporate attorneys that bill hourly are looking for and it isn’t our product.

Patrick: That is valuable information.

Rich: So we’re going to interview other types of attorneys.

Patrick: Maybe not what you wanted to hear – but valuable nonetheless

Rich: Yes, that would have been nice if all attorneys wanted our product but we had no reason to think that corporate attorneys would specifically like it I just took interviews with every attorney that I could, in hopes that they would point me in the right direction.

Patrick: I like that. Hustle.

Rich: So now I can start biasing my search.  Yeah, it helps to interview as many people as you can IMO. Spending that time doing more development is a waste in our case as I built an MVP that my co-founders firm is using.

Patrick: Do you go into your interview with a game-plan of what you want to know? Or do you go more free-form?

Rich: A basic game plan.  I identified the reasons that I thought that our product would be of value I came up with some signs that it would be valuable and asked them of those signs (symptoms?) manifested themselves at their firm. It wasn’t that easy for Stylous.  People that shop don’t have a “problem” per se that they are trying to solve.

Patrick: For Clientite, have you asked about pricing at all?

Rich: Vaguely, mostly about what we would charge for, not how much.

Patrick: When you do an interview – how do you judge whether or not it added value to your startup?

Rich: I don’t have a specific metric or anything like that.  Mostly gut. Did I learn something that will help me make decisions in the future? Will I be able to ask different questions in the next interview based on the results of this one?

Basically can I use this to refine my search for the best customer type.

Patrick: Do you think a technical co-founder can or should rely on a generalist co-founder to do custdev interviews for him?

Rich: I think it depends on the specific situation. I don’t think that technical co-founders should ever entirely opt out of the process.  They should at least see the product being discussed first hand even if they aren’t asking the questions.  Ideally, you could be both the problem team and the solution team but those opportunities are hard to come by.

My co-founder lives on the east coast, so were doing it in parallel, he’s the subject matter expert (an attorney). It’s probably more important for me to get out and interview people at first because I know little about the industry.  That may vary for other tech co-founders

Patrick: How do you feel about surveying?

Rich: I don’t do any surveying. I wouldn’t call that customer development because it doesn’t provide you very general information. It isn’t a good way to explore the problem.  It might be a good way to see how lots of people feel about something very specific.  Like have you noticed that you’ve started converting more customers since you’ve started using our product, but it’s unlikely to provide much insight outside of the question itself and it might even fail at that, since people might interpret the question in a way that you didn’t intend you can only find that out through a conversation.

For instance, I interviewed someone today and his idea of the product was entirely different from what I described in the one line email I sent him asking him for the interview.

Only through a conversation and showing him a demo did he actually understand what I was trying to do

Patrick: Out of curiosity, what did he mix up?

Rich: He thought we were building a product that would be used to improve the efficiency of sending messages to the client, meaning some interface for sending emails.  Our product does send emails but the more important feature is that it provides a website for clients to log into to look at all of their case information. He didn’t get that from my description.  Which also provides evidence that an MVP can be important to have when you’re interviewing. It’s vary hard to abstractly describe the experience of using something.

Patrick: Absolutely.  Did you change your one-line email description?

Rich: That was just an hour ago … so not yet ;-)

Patrick: How many customer calls/interviews are you doing a week (roughly)?

Rich: 10. I’d like to do more but I’m just starting out for this project.

Patrick: That is impressive.

Rich: So this is the initial wave, I’d like to try to do 4x a day, if I could find a way to schedule that many people. That is another hard thing being a developer: The real world moves very slowly compared with software.  It’s easy to measure progress when you’re writing code, at least in terms of the features that you’re trying to build, not necessarily in your goal of building a profitable company.  So you (I) would typically rather write code because I know that I’m going to make progress, even though I now believe that to be a flawed approach.

Patrick: And that feels pretty good, right?  Building something?

Rich: Yeah, it’s easy.

Patrick: But it may be illusory and that is the point of LS/CD.

Rich: Yes to mitigate risk that you’re building the wrong product.    Building something only mitigates it somewhat.

Patrick: I imagine for you it is rather facile to get into “The Flow” when coding – have you been able to get into The Flow when interviewing?

Rich: I think that I’m still working on that.  I’m guessing that I will with time.

Patrick: Are the interviews easier than before?  Same? Less stressful? More?

Rich: Certainly easier than when I first started with Stylous.  It is also getting easier with Clientite as I learn more about the industry and how I approach things.  It’s hard to have a conversation with someone when you don’t know WTF you’re talking about.

Patrick: Good point!

Rich: So the uncertainty around that is reduced as I learn more of the language that people speak.

Patrick: This is something Steve Blank has talked about, I think he has given examples of this on his blog.

Rich: I wouldn’t normally have started a company focusing on an area that I don’t understand but I think that Lean Startup provides a process that you can use to learn about a market.

Patrick: Agreed.

Rich: That combined with having a subject matter expert co-founder gave me the confidence to start the company.  It was also easy to build an MVP and the value proposition seemed fairly clear (at least for my co-founders firm).  So I think that this company is ideally suited to using CD / LS, if I were doing something a bit more “out there” I might spend more time building the experience before having customers use it as truly innovative things can be hard to understand when described abstractly but that just changes the way you follow the philosophy, it doesn’t discredit it.

Patrick: Again, agreed.

Rich: For instance, it would be hard to validate the iPad by interviewing people so you need to start working on building the experience and testing it out with people, but you still validate the idea before you build the factories.

Patrick: Yep.  To end, is there anything you want to add that speaks directly to engineers not wanting to get out of the building?

Rich: I guess I would tell developers to make sure that they really search their feelings and try to understand any resistance that they have and make sure that they aren’t making excuses because they don’t want to experience getting out of the building when deep down they know that it will be valuable.

Judo is a sport which translates roughly into the “supple/gentle/soft way” in Japanese.  The central guiding principle of judo is “Maximum Efficiency with Minimum Effort”.  This distilled means that to defeat your opponent in the context of a judo match you should first unbalance him so:

when your opponent pulls you, you push him.

when your opponent pushes you, you pull him.

In an actual judo match, please believe me, things are significantly more complex, dynamic and chaotic than that.  But the zenith of elegant, beautiful and, I should emphasize, effective judo is always one that preserves this elemental principle:  non-resistance and taking advantage of the opponent’s loss of equilibrium.

Good judo also subordinates physical strength to actual judo technique based on the principle above.  That is not say that physical strength is unimportant — it most certainly is and ceteris paribus, the bigger judo player almost always wins — but that extreme strength should be applied at the right time, once one’s opponent is off-balanced.  In good judo, technique is paramount, while strength is secondary.

Beginning judo players (and often in the heavy-weight category where I fought) tend to “muscle” their opponents in effort to throw them, in other words, minimum efficiency with maximal effort.  And generally speaking, to be a world-class judoka, technique scales much more effectively than strength.

Here is what good judo looks like in in the “real-world” of international judo competition. Watch as the Brazilian judo player (in the blue judogi) times and initiates his throw upon the French player (in the white judogi) as the French player steps forward with his left foot.

It is plain to see that the Frenchman has forward inertia that is being exploited by the Brazilian. The Frenchman has effectively pushed, and in response, the Brazilian has very literally pulled, resulting in a beautiful throw by the Brazilian to win the match…in six seconds BTW.

Now, I have no desire to stretch, exaggerate (and break) judo as a metaphor for life, business or startups — this has already been done — but I do think there are certain parallels between judo and Lean Startups/Customer Development that are worthwhile exploring.

Leaving aside the fact that your customers aren’t your opponents and your aim is not to defeat them, I think much can be said for a judo-like approach of Maximum Efficiency with Minimum Effort, by establishing marketing and positioning (initiating and timing your throws) based upon on what your customers’ perceptions are or aren’t about your product (their pushes or pulls), rather than forcing what you think your positioning should be upon them (only your brute strength).

What does this look like in the real world of startups?

It looks like what Sean Ellis espouses with Survey.io and what Chris O’Donnell of Markitechture describes in a recent post, A/B Testing for Product-Market Fit.

Use your customers’ pushes and pulls to your advantage.  Don’t try to simply muscle them to where you think you want them.

Maximum Efficiency with Minimum Effort.  Can you think of any additional examples?

The Danger of Premature Media Attention

Fan Bi, the founder of Blank Label, has an interesting post up at his blog.  The original tweet announcing the blog post was captioned as “a perspective on the dangers on going lean.”  From the blog post:

We were getting more and more customers, but most of them were what Geoffrey Moore would describe as innovators or early adopters. The people who are okay with working with a less than perfect website, can deal with product delays and willing to give you a bucket-load of feedback, and order again just to see if you’ve taken the feedbac

May 17 was a turning point in the history of Blank Label. It was a great day, and a terrible day all the same. We had a half page feature in the New York Times and our focus on innovators and early adopters, and being lean, were blown completely out of the water. There were some early signs of things going wrong. By 10am that Sunday morning, our site had pretty much stopped functioning. We migrated servers midday, it took about 4 hours. We were getting phone calls asking us how they could order via the phone because they didn’t understand this “web stuff”. Yikes! And by 6pm, we had collected more orders on the day than we had from October 31 last year to the previous day. Chaos!

I think that this post is less about the risks of being a Lean Startup than it is about re-capitulating the many dangers and pains of premature media attention.  Brant Cooper agrees.  As Eric Ries writes:

A marketing launch establishes your positioning. If you don’t know what the right positioning is for your company, do not launch. Figuring this out takes time, and few entrepreneurs have the patience to wait it out, because the business plan does such a good job of explaining what customers are going to think. The problem is that customers don’t read your business plan.

When you launch with the wrong positioning, you have to spend extra effort and money later cleaning it up. For example, we did some early press (in Wired, no less) for IMVU that called us the next generation of IM and compared us positively to AOL. At the time, we thought that was great. Now, I look back and cringe. Being compared to AOL isn’t so great these days, and IM is considered a pretty weak form of socializing. When we finally launched for real, we had to compensate for that early blunder.

I wrote a little bit about the epic launch we had at a previous startup in my post Achieving a failure. We really did it well, with a great PR firm and great coverage. New York Times, Wall Street Journal, CNN, the works. But it turned into a crushing defeat, because we couldn’t capitalize on all that attention. The product didn’t convert well enough, the mainstream customers we were driving weren’t ready for the concept, and the event fed expectations about how successful the product was going to be that turned out to be hyper-inflated.

Nivi of Venture Hacks revisits the same post and adds:

If you ask most entrepreneurs why they want to launch, you’ll get an answer like, “So people find out about my product — are you stupid?” But a thoughtful founder who read Eric’s article recently asked me,

The New York Times wants to write about my company. It will take me no time or money to get this press. What should I do?

You should still consider the downsides of launching:

  1. Launching a product that doesn’t solve a real customer problem establishes the wrong positioning in the minds of customers.
  2. You can only launch once. If you launch the wrong product or you have an un-optimized funnel when you launch, you just wasted a one-time opportunity to harvest and generate demand.

Launching isn’t the only way to harvest demand. You can reach customers through customer development: AdWords, search engine marketing, online ads, contacting prospective customers through LinkedIn, et cetera — get creative. Don’t launch just because everybody else is doing it — be thoughtful.

It is fascinating to actually witness Blank Label struggling to keep up with the demand created by the New York Times article.  Demand by people in some segments that Blank Label really doesn’t want to serve!   (i.e. “We were getting phone calls asking us how they could order via the phone because they didn’t understand this “web stuff”.)

The article, however well-intentioned, may have wrought serious damage to Blank Label as early adopters, the people who, according to Fan,”who are okay with working with a less than perfect website, can deal with product delays and willing to give you a bucket-load of feedback, and order again just to see if you’ve taken the feedback” might feel a bit alienated as they bear the negative externalities of this premature media attention, as evidenced by the message currently visible on the Blank Label site.

Hopefully, this new demand is sustained, the early adopters stick around and Blank Label can pull through and emerge even stronger, but it would behoove all of us to ruminate a bit on this.  BTW reading the NYT article, it is heartening to see that Fan Bi, like entrepreneurs Giff Constable and Hiten Shah, appears to be a what I have come to calling a “Relentless Customer Developer” and does everything he can to talk to his customers on a daily basis.

One goal was to communicate directly with customers. The Web site commands: “Call us. We like to talk.” Depending on the time of day, Mr. Bi answers the calls himself. When he is awake, he also activates a feature that sends instant messages to customers who have been on the site for more than 90 seconds.

Need help? he asks. For several hours a day, he and his partners chat with customers about what they like and don’t like on the site.