Ship early and listen to real customers - then adapt.
This reminds me of the Customer Development model championed in Blank's "The Four Steps To The Epiphany".
I agree and work toward that philosophy even though it is sometimes difficult. For example, finding a manufacturer that can provide fast turnaround may be difficult and more expensive but it would probably be worth it for the additional agility.
I always thought Henry Ford "won" because he innovated on the manufacturing process with his introduction of the assembly line, which allowed him to produce better and cheaper automobiles more quickly than his competitors.
I had never heard that he embraced a "continuous deployment" model, and this short interview obviously doesn't get into details.
Yes, his idea was assembly line, which by Eric Ries was completely overrated?
And Google - how do you make better search engine by "listening your customers"? No, you don't, you need an idea of pagerank foe example. Without that worthless idea you are toast, no matter how much you listen...
Oh and that thermonuclear reactor that we all dream about and what will resolve all energy problems on earth. No idea how to build it. Listen customers? Good luck with that.
The PageRank paper directly references customer concerns and their approach: "Some argue that on the web, users should specify more accurately what they want and add more words to their query. We disagree vehemently with this position."
* Customers want higher quality, cheaper cars -> Create a production process that satisfies these requests.
* Customers want better search results -> Create an algorithm that finds better results.
* Customers want clean, abundant energy -> Find out about the thermonuclear reactor.
You're looking at it from the wrong side.
If customers didn't want better search results but instead wanted portals, Google would have died, and Yahoo and Altavista would be about as big as Google right now. How do you know they want better search results? By listening to their needs.
Your product-idea might be the best thing since slice bread, but it's worthless if no one wants it.
PS: I'm not saying Google or Ford used this process, of course. But they are extraordinary companies. Also, if enough companies try, at least some should come close to satisfying the customer's needs just by chance and some intuition. You don't hear about the companies that had a great idea, implemented it, and then dropped dead for lack of customers.
If you stop being pedantic, a car really is a faster horse. It just doesn't the physical nature of a horse, that's all.
This is what I don't get about the regurgitation of that "quote". It's obvious that what people are asking for is something faster; "horse" is just a point of reference.
People could have asked for a smaller horse, or a horse that eats less hay, or a horse that doesn't shit so much, and if you're not a blindered literalistic you'll easily read between the lines.
The "users are clueless" meme is perplexing. If they're so clueless, what does that say about the products they choose?
"Your product-idea might be the best thing since slice bread, but it's worthless if no one wants it."
I'm not arguing about that, that I agree actually. But your brilliant execution is equally worthless if you do not have an idea HOW you should execute. Creating an algoritm that finds better results is not something you just sit and do.
But.
- As you may not know. Customers did not what mobile phones. At all. And they absolutely did not want SMS. By listening their needs, we still be using landlines these days.
- Customers did passionately, desperately want videophones, since 70's I remember. :-)
I don't think many people want videophones. That would mean having to wear pants and t-shirts all day and combing your hair. That's to me why it never became reality.
So in effect people didn't know they wanted SMS but when they got it they loved it. People claimed they wanted Videophones but when they understood the drawbacks they balked.
"Listening" to your customers also includes the iterative, quantitative process of testing and measuring your ideas, not just listening to qualitative customer feedback and requests.
Ford didn't invent the assembly line, he mastered it. Ford's naming conventions also hint at the continuous deployments, small batches and quick pivots that Ries is talking about:
When Ford finally found a hit with the Model T, there were 18 prior prototypes (A-S), including many production models - A, C, F, N, R, and S. The revolutionary idea of interchangeable parts (again, congruent with Ries' thesis) had a lot to with with why Ford was able to iterate so much.
"As a result of these developments in method, Ford's cars came off the line in three minute intervals. This was much faster than previous methods, increasing production by eight to one (requiring 12.5 man-hours before, 1 hour 33 minutes after), while using less manpower.[2] It was so successful, paint became a bottleneck. Only japan black would dry fast enough, forcing the company to drop the variety of colors available before 1914, until fast-drying Duco lacquer was developed in 1926.[2] In 1914, an assembly line worker could buy a Model T with four months' pay.[2]
"The assembly line technique was an integral part of the diffusion of the automobile into American society. Decreased costs of production allowed the cost of the Model T to drop within the budget of the American middle class. In 1908, the price of a Model T was around $825, and by 1912 it had dropped to around $575. This price reduction is comparable to a drop from $15,000 to $10,000 in dollar terms from the year 2000."
To me those numbers are quite significant. I am no student of automobile history, but it seems to me that the primary driver of Ford's success was the massive production efficiencies and reduced costs associated with the modern moving assembly line. I have to imagine there were "better" cars than the Model T, or other cars that had more prototypes and iterations.
Anyhow, I don't really mean to dispute the author's thesis here, since there's not really enough evidence to judge it. But I'm having a really hard time being convinced that some factor other than Ford reducing the production time per car by a factor of almost 10 and the price by almost half was primarily responsible for the company's success with the Model T.
Better = more reliable, more comfortable, more capable. There were many cars of the time that were unarguably better than the Model T, but production and cost were the biggest factors of the Model T's success.
Eric - it's great to see you eating your own dogfood. Initial idea = write book on lean startup.
Launch plan = make sure every human being on the planet hears about it.
Adapt = hunt down the few poor bastards who haven't head about it yet.
Adapt again = hunt them down some more.
right now you are a tidal wave engulfing the media space. Gotta hand it to ya.
He says 501 companies had the idea of the IC engine, all but 1 ( ie. Ford ) died.
Ok.
But its not the idea being overrated that killed 500 and kept 1 alive, cause they all had the same idea to begin with. What killed the rest was - competition, traction, capitalism, free markets, better execution, call it whatever. If the idea ie. IC engines, was overrated, you wouldn't get a car from the entire 501. What happened instead was - you got 501 car guys, & over time 500 car guys vanished and 1 car guy survived. That's what you'd expect anyways.
So how does one go from that to dissing ideas themselves ?
Correct me if I'm wrong but for a startup, you only have one chance. So you have to pick the right idea then the idea is all for a startup =>
For me, the right thing for a startup founder is to transform himself to a typical custommer to find the right idea.
Concerning the iteration process, that's a question of beta testing. So you have the right idea, now you have to make a good application with it which is something different.
In conclusion, I would say that the good equation is: Idea+Ship=Startup. (I assume the beta testing phase is as easy as programming or whatever)
I still have a hard time buying into the lean startup method hook, line, and sinker. I think there is quite a bit of evidence in support of it, but a lot that flies in the face of it as well.
Certain features, for which presentation isn't critical, I would certainly want to deploy early and often. But that also seems to be contradictory to the dictum "execution is everything." Of course building something with excellent execution quickly is ideal, but if you release your product before it is what you believe is your best work, then aren't you kind of hurting what little chance you've got?
To draw an example, Palm had a touch screen phone and apps available for download in 2004. But their execution wasn't very good, and their software provided an awful experience. In 2007 Apple released the exact same idea but with far better execution, and the rest is history.
I'm not saying the lean startup method doesn't work, I just think it is highly dependent upon the product strategy.
One, they're an established brand selling to a mature market. Startups are new brands selling to early adopters. Early adopters are very forgiving.
Two, Apple also iterates a ton, just like a lean startup. The difference is that they do it internally. The iPod went through circa 100 prototypes. That compares with 4-7 for the typical consumer electronics company.
If you are as well-capitalized as Apple, by all means iterate in private. Most of us aren't so lucky, though.
Yeah, and I'm not denying that. My point is that if a product is released just to see whether it gains traction without ideal execution, then how can you know whether it was the faulty execution that led to failure? Maybe if the user were more intrigued from the beginning with better execution, better traction could be achieved.
Suppose you release a product with execution you believe perfect and it flops. The problem could be the idea, but it still could be your execution.
Whether you're Apple or a nobody, you still have to keep iterating until it's right. The way you tell what to try next is by studying how the last one went. You just eventually decide to work on a different idea.
You can never truly know whether a repeated flop is due to your idea or your execution.
The Model T is not a very good car, even considering the time. It's about as simple transmission as you can get, terrible brakes, a gas tank under the front seat (how convenient), would roll over in heartbeat, you'd get soaked when it rained and freeze in the winter, and only came in black.
It was wildly successful because Henry Ford knew one thing - customers wanted cars but they couldn't afford the current offerings. Everything he did was to make the car affordable to the common man.
I will disagree that timing doesn't matter. Ford was definitely at the right-place at the right-time. 5 or 10 years later and he probably wouldn't have been so successful, someone would've beaten him to the punch. The plain fact is that while timing isn't everything, it can't be considered irrelevant.
I heard a different wording of "ship early, ship often" tonight. One that is to me a little more tought provoking even though it says exactly the same thing: "It's better to have unhappy customers than no customers at all."
Because unhappy customers provide feedback that allows you to change your course it's the same thinking as Ship Early or Fail Fast and the rest of the lean mantras. But those have been repeated enough to loose their meaning.
So it was refreshing to hear a rephrasing (that actually predates Eric Ries' lean startup methodology by two full decades) that wasn't as easily dismissed as plain common sense.
It's better to have unhappy customers than no customers at all.
I find it most ironic that the customer development guys are so fond of preaching their philosophy of how to properly build businesses, yet they're failed entrepreneurs and have yet to successfully prove themselves. They're better marketers/bloggers than they are entrepreneurs, buyer beware.
In principle, I can appreciate the need to be 'lean' and nimble when starting out, however, these concepts are not new, and Eric Reis certainly did not spearhead this method.
Eric Reis has done a great job at making noise and he has demonstrated that he is a consummate salesman. While the concepts that he espouses aren't exactly innovative, there is no doubt, a deep appeal, particularly to those who are seduced by the 'gold rush fever' that is driving this current tech boom(bubble?).
Ultimately, it strikes me that this guy is a huckster who is fast becoming the spokesperson for a growing league of entrepreneur-wannabes in search for a piece of 'social' pie.
Exactly! "Ideas are overrated" is neither radical nor is it new. Somewhere we all read these points before, probably many times before, and in some instances we read and posted them on HN. Why this is suddenly celebrated like mana from heaven is beyond me.
You know, I used to agree with this statement: "Ideas are overrated." Until I came up with an idea that was not.
There are some ideas that are so complex - but then they are so simple - that it's hard for me to see other people coming up with that exact idea in that way.
Once it's done, then it becomes obvious - but that's not necessarily true before the fact.
For instance, Google's PageRank is so obviously the best way to rank links (based on the number of incoming links) - but who woulda thunk before they did it?
In fact, when they were just starting out, many of the search players thought it was a dumb idea...which goes to show that they under appreciated the novelty of the idea.
Actually a few people were experimenting with the idea before the Google guys did it. They just did it better. In the plex talks about that in the first couple chapters.
Idea is a single word. You can neither say if it is overrated, nor analyze the market on the basis of a single word.
— But can I write bullshit articles with a flamebait title interviewing some guy no one takes seriously?
Yes, you can.
There are pure ideas that just occur on an impulse. And there are ideas that satisfy a given demand. If you skip one of the steps when implementing an idea that satisfies the demand, you bail on the demand, not on some abstract over-hyped quality that serves no purpose.
In my opinion the really worthwhile ideas go alongside with the demand, and making sure that the demand is met is no overrated practice. Why? Because the demand is just a social definition of a problem that needs a solution.
People that have ideas unrelated to the demand usually write this kind of interviews.
54 comments
[ 2.2 ms ] story [ 107 ms ] threadThis reminds me of the Customer Development model championed in Blank's "The Four Steps To The Epiphany".
I agree and work toward that philosophy even though it is sometimes difficult. For example, finding a manufacturer that can provide fast turnaround may be difficult and more expensive but it would probably be worth it for the additional agility.
I had never heard that he embraced a "continuous deployment" model, and this short interview obviously doesn't get into details.
And Google - how do you make better search engine by "listening your customers"? No, you don't, you need an idea of pagerank foe example. Without that worthless idea you are toast, no matter how much you listen...
Oh and that thermonuclear reactor that we all dream about and what will resolve all energy problems on earth. No idea how to build it. Listen customers? Good luck with that.
Section 3.1 - http://infolab.stanford.edu/~backrub/google.html
* Customers want better search results -> Create an algorithm that finds better results.
* Customers want clean, abundant energy -> Find out about the thermonuclear reactor.
You're looking at it from the wrong side.
If customers didn't want better search results but instead wanted portals, Google would have died, and Yahoo and Altavista would be about as big as Google right now. How do you know they want better search results? By listening to their needs.
Your product-idea might be the best thing since slice bread, but it's worthless if no one wants it.
PS: I'm not saying Google or Ford used this process, of course. But they are extraordinary companies. Also, if enough companies try, at least some should come close to satisfying the customer's needs just by chance and some intuition. You don't hear about the companies that had a great idea, implemented it, and then dropped dead for lack of customers.
This is what I don't get about the regurgitation of that "quote". It's obvious that what people are asking for is something faster; "horse" is just a point of reference.
People could have asked for a smaller horse, or a horse that eats less hay, or a horse that doesn't shit so much, and if you're not a blindered literalistic you'll easily read between the lines.
The "users are clueless" meme is perplexing. If they're so clueless, what does that say about the products they choose?
I'm not arguing about that, that I agree actually. But your brilliant execution is equally worthless if you do not have an idea HOW you should execute. Creating an algoritm that finds better results is not something you just sit and do.
But. - As you may not know. Customers did not what mobile phones. At all. And they absolutely did not want SMS. By listening their needs, we still be using landlines these days. - Customers did passionately, desperately want videophones, since 70's I remember. :-)
So in effect people didn't know they wanted SMS but when they got it they loved it. People claimed they wanted Videophones but when they understood the drawbacks they balked.
http://answers.yahoo.com/question/index?qid=20091102183112AA...
and
http://en.wikipedia.org/wiki/Assembly_line#Olds_Motor_Vehicl...
But he improved on it quite a bit.
When Ford finally found a hit with the Model T, there were 18 prior prototypes (A-S), including many production models - A, C, F, N, R, and S. The revolutionary idea of interchangeable parts (again, congruent with Ries' thesis) had a lot to with with why Ford was able to iterate so much.
http://en.wikipedia.org/wiki/Assembly_line#Ford_Motor_Compan...
"As a result of these developments in method, Ford's cars came off the line in three minute intervals. This was much faster than previous methods, increasing production by eight to one (requiring 12.5 man-hours before, 1 hour 33 minutes after), while using less manpower.[2] It was so successful, paint became a bottleneck. Only japan black would dry fast enough, forcing the company to drop the variety of colors available before 1914, until fast-drying Duco lacquer was developed in 1926.[2] In 1914, an assembly line worker could buy a Model T with four months' pay.[2]
"The assembly line technique was an integral part of the diffusion of the automobile into American society. Decreased costs of production allowed the cost of the Model T to drop within the budget of the American middle class. In 1908, the price of a Model T was around $825, and by 1912 it had dropped to around $575. This price reduction is comparable to a drop from $15,000 to $10,000 in dollar terms from the year 2000."
To me those numbers are quite significant. I am no student of automobile history, but it seems to me that the primary driver of Ford's success was the massive production efficiencies and reduced costs associated with the modern moving assembly line. I have to imagine there were "better" cars than the Model T, or other cars that had more prototypes and iterations.
Anyhow, I don't really mean to dispute the author's thesis here, since there's not really enough evidence to judge it. But I'm having a really hard time being convinced that some factor other than Ford reducing the production time per car by a factor of almost 10 and the price by almost half was primarily responsible for the company's success with the Model T.
And wikipedia agrees with your conclusion:
> Soon, companies had to have assembly lines, or risk going broke by not being able to compete; by 1930, 250 companies which did not had disappeared.
It also notes iteration in the perfection of the assembly line itself, though going by its older name of trial and error.
What definition of "better" are we using? (For example, higher top speed isn't relevant if you have bad roads.)
> or other cars that had more prototypes and iterations.
Prototypes and iterations cost money. Ford had more money, so why don't you assume more prototypes and iterations?
In conclusion, I would say that the good equation is: Idea+Ship=Startup. (I assume the beta testing phase is as easy as programming or whatever)
Certain features, for which presentation isn't critical, I would certainly want to deploy early and often. But that also seems to be contradictory to the dictum "execution is everything." Of course building something with excellent execution quickly is ideal, but if you release your product before it is what you believe is your best work, then aren't you kind of hurting what little chance you've got?
To draw an example, Palm had a touch screen phone and apps available for download in 2004. But their execution wasn't very good, and their software provided an awful experience. In 2007 Apple released the exact same idea but with far better execution, and the rest is history.
I'm not saying the lean startup method doesn't work, I just think it is highly dependent upon the product strategy.
One, they're an established brand selling to a mature market. Startups are new brands selling to early adopters. Early adopters are very forgiving.
Two, Apple also iterates a ton, just like a lean startup. The difference is that they do it internally. The iPod went through circa 100 prototypes. That compares with 4-7 for the typical consumer electronics company.
If you are as well-capitalized as Apple, by all means iterate in private. Most of us aren't so lucky, though.
Whether you're Apple or a nobody, you still have to keep iterating until it's right. The way you tell what to try next is by studying how the last one went. You just eventually decide to work on a different idea.
You can never truly know whether a repeated flop is due to your idea or your execution.
It was wildly successful because Henry Ford knew one thing - customers wanted cars but they couldn't afford the current offerings. Everything he did was to make the car affordable to the common man.
I will disagree that timing doesn't matter. Ford was definitely at the right-place at the right-time. 5 or 10 years later and he probably wouldn't have been so successful, someone would've beaten him to the punch. The plain fact is that while timing isn't everything, it can't be considered irrelevant.
Because unhappy customers provide feedback that allows you to change your course it's the same thinking as Ship Early or Fail Fast and the rest of the lean mantras. But those have been repeated enough to loose their meaning.
So it was refreshing to hear a rephrasing (that actually predates Eric Ries' lean startup methodology by two full decades) that wasn't as easily dismissed as plain common sense.
It's better to have unhappy customers than no customers at all.
Eric Reis has done a great job at making noise and he has demonstrated that he is a consummate salesman. While the concepts that he espouses aren't exactly innovative, there is no doubt, a deep appeal, particularly to those who are seduced by the 'gold rush fever' that is driving this current tech boom(bubble?).
Ultimately, it strikes me that this guy is a huckster who is fast becoming the spokesperson for a growing league of entrepreneur-wannabes in search for a piece of 'social' pie.
There are some ideas that are so complex - but then they are so simple - that it's hard for me to see other people coming up with that exact idea in that way.
Once it's done, then it becomes obvious - but that's not necessarily true before the fact.
For instance, Google's PageRank is so obviously the best way to rank links (based on the number of incoming links) - but who woulda thunk before they did it?
In fact, when they were just starting out, many of the search players thought it was a dumb idea...which goes to show that they under appreciated the novelty of the idea.
— But can I write bullshit articles with a flamebait title interviewing some guy no one takes seriously? Yes, you can.
There are pure ideas that just occur on an impulse. And there are ideas that satisfy a given demand. If you skip one of the steps when implementing an idea that satisfies the demand, you bail on the demand, not on some abstract over-hyped quality that serves no purpose.
In my opinion the really worthwhile ideas go alongside with the demand, and making sure that the demand is met is no overrated practice. Why? Because the demand is just a social definition of a problem that needs a solution.
People that have ideas unrelated to the demand usually write this kind of interviews.