The Lean Startup: A contrarian view | Hacker News Reader

The Lean Startup: A contrarian view

khitchdee.forumatic.com

30 pointssaurabh45 comments

>That assumption is that you have something to learn from the market that you don't already know.

In the Lean Startup book, Eric Ries defines a startup as "a human institution designed to create a new product or service under conditions of extreme uncertainty". "That you have something to learn from the market that you don't already know" isn't an assumption, it's a premise. If you already know about the market, then you aren't facing conditions of extreme uncertainty and you aren't, by Ries' definition, involved in a startup.

Lean startup doesn't claim to be the solution to every business' problems in the world. It's a solution that works really well when you're dealing with a lot of unknowns. One way that entrepreneurs deal with unknowns is by speculating and acting on assumptions, which often turn out to be wrong resulting in waste. Ries argues that this excessive waste is avoidable by taking a more scientific approach: identify your assumptions, validate them through measurement, adjust your product/strategy according to results, and let that process of validated learning pave a clear path for your business through the shadowy unknowns.

If you're planning to startup, what you'll here from most people is how you should follow lean startup principles. The point of this writeup is to give some breathing space to those entrepreneurs who have a top down idea. It's to encourage them to go for it instead of questioning themselves. Typically, the game changing ideas come top down and not bottom up. If you want to change the game everyone's playing, you have to step out of it first.

The central question of the Lean Startup book is: how do you know that your top down idea is correct? Strictly speaking, until you go to market ... you don't.

There are two basic classes of strategy to doing anything:

* Commit all your resources in advance

* Incrementalism

In situations of uncertainty, the latter is a simple risk management strategy. Armies have scouts, marketers work in test markets, engineers build prototypes.

What Eric Ries is selling isn't strictly new. But it's a cohesive repackaging of extant ideas that in my opinion is useful for making a single universal feedback loop the basis of a business.

Going lean and incremental does reduce risk. It also reduces reward.

> It also reduces reward.

How?

If you're going to drive your product based on customer feedback, its growth is necessarily going to be slower and more in line with existing conditions in the market. If instead, you drive your product based on a game changing idea, when you hit the market, your product will be completely different than what's available at the time. The statistics will therefore be completely new and in your favor. That's the theory anyways.

Right, creating new markets. That's fine.

But basically the failure rate on it is atrocious. If you find a different new market that is immediately adjacent, what's the smart move?

When I studied nature-inspired computation, my professor quipped at one point that nature-inspired optimisation algorithms usually show what you got wrong about your problem definition.

Markets are the same. Nobody swoops down like Batman with a flawless plan and mastermind insight that turns out to be perfectly correct. The airport business books might give that impression, but it's total bullshit. What actually happens is that millions of men and women swoop down like Batman and discover that they're not backed by Bruce Wayne and a friendly script editor.

Once in a while somebody stumbles onto something good and then their history is rewritten to fit the monomyth.

There is no such thing as an entrepreneur with a top down idea. You can be an ideas guy, but you can't be an entrepreneur without taking what you have to customers and iterating on it until it's right.

Please give an example of a game changing idea that came about 'top down' ? Where do these ideas come from, in your opinion?

An example of a game-changing idea that came top-down was the use of the touchscreen interface in the iPhone. Someone thought this would be a cool way to interface, worked down to a cool product and people liked it and used it a lot.

See, now we're well into the realm of arguing by warring definitions.

One of the points of the lean startup book is to go where the evidence points to go. Apple et al throw lots of features at the market and a lot of them die with a whimper.

Touch screens turn out to be really popular, so Apple has progressively doubled down on that.

And that's pretty much what the lean startup is arguing they should do. The main difference is that Ries made it an explicit loop that covers all activities in a company.

To me the most useful analogy is simulated annealing. Each startup is dropping into a chosen neighbourhood of the stupendously large space of possible businesses in possible markets in possible societies. Each strikes off in the direction it thinks is where the optimum lies.

What happens is that if you don't periodically check the slope, you'll run out of money. And if you don't periodically jump around the solution landscape, you may well miss out on a better optima and become stuck in a local optimum.

But like all multi-objective optimisation problems, both the solution and the search for the solution are riddled with unpleasant tradeoffs.

The iPhone is just one point in a long line of evolution that started way back with Doug Engelbart's work, if not before.

Apple would like you to believe the iPhone sprang into existence fully formed, but that is complete bullshit. Apple experimented with touch screen interfaces with the Newton in the late 80s / early 90s, and completely failed. Palm showed how touch screens could work in the mid / late 90s. Numerous academic projects worked on touch screens. The iPhone was the right product at the right time, but it was not truly innovative in a technical sense.

I'm not sure that is entirely valid. Most smartphones before the iPhone were touch screen - resistive and needed a pen, like the Palm Treo, and the many Windows Mobile Phones before it, - but still touch screen.

Having a capacitive screen only seemed like the next logical step, and of course to do it well, you will need to the change the UX.

In any case I think its disingenuous to contribute the iPhone's success to the top-down touch screen when many of the game changing features were brought to light after the iPhone went to market, such as the Apple-Controlled App Store.

Lastly, there is no way the touch screen could have been considered top-down from Apple in 2007. The touch screen interface had already been iterated numerous times starting in 1987 with the Apple Newton.

Touch screen interfaces on smartphones existed long before the iPhone.

The Lean Startup took off when it did because it represented a new way of building a business. Before the open source stack got deep enough and servers got cheap enough, there was no room for low-capital startups. Then we hit an inflection point where it became increasingly possible to bootstrap highly scalable businesses.

The Lean Startup is simply an old world entrepreneur discovering the possibilities in the new world. It's not that it's better, it's just been more fertile ground over the last several years. However, as more and more smart kids show up on the scene, I think the pendulum will swing back because the stuff that can be done without capital or investment is going to get tapped out by armies of ramen-eaters. There's always room for a great new idea, but if you have access to capital a little barrier to entry is not a bad thing.

This

And capital need not necessarily be direct financial in vestment. For example any Web 2.0 ramen startup will have 15+ years of computer and software experience and education as its capital base. Throw in a hardware need, like making thermostats and the rarity of the education and thus capital Increases Last weeks demo day highlighted this for me - UAV hardware and software plays - that's gonna be a rare combo in any market, and guys trying to offer easy divorce web sites - the capital needed there is quite low

I think if we look at capital as previously invested education for skills needed, the fill in the gaps websites have no moats, the arduino is about to be the biggest barrier to entry for the Web 3.0

I also feel like The Lean Startup really isn't anything new. Its just articulated and repackaged in a different way that's more sexy, or tech-y.

God forbid any of my peers start reading Drucker...

I was actually reminded of Peter Senge's Fifth Discipline.

That's a good one too - luckily they had a second edition or it may have been totally lost in the endless abyss of time. I'd be really surprised to find someone of my generation to have read about it though...

But the thing that caught on about the idea of a "lean startup" in the area I live in (Salt Laje City) wasn't so much the concept of "it takes less money to start a business now" as "your customers know what they want more than you do." Because of this, every investor instructs you to read a book called "Nail It Then Scale It," which is a dumbed down version of The Lean Startup, which is a dumbed down version of Four Steps to the Epiphany.

The process is explained like this. "You have an idea. But is it validated? Go ask a bunch of people if they think it's a good idea." People aren't encouraged to build a minimum viable product to iterate on first.

I see this as a huge step backward, removing the intuition of a founder and killing serious innovation. The thing is this model works great for some SaaS companies that try to figure out which software is needed in industry X then create it for them. But that path of entrepreneurship makes me cringe. That sounds like no fun at all. Will it be successful? Sure, a bunch of $1-50 million companies will come out of that, but just building stuff people tell you to? No thanks.

The truest entrepreneurship, or at least what I want to be involved in, is where you operate according to your own vision and intuition, and give people something they had never imagined. That's incredibly hard to do, but that's what's so exciting about building companies, at least to me.

If there is no market for it, it will die. There's a bit of arrogance in the idea that you know better and can do things from the "top-down". Getting user feedback doesn't mean replacing your ideas with whatever John Doe thinks he likes, but finding out how your ideas fit (or not) the users' needs.

Before someone starts misquoting, even Apple did user testing, although not in the traditional format; there is no point in building someone that suits nobody.

This is correct. Without customer interaction, there are no good ideas, only lucky ideas and failed ideas.

So you contend that Apple must be one helluva lucky company since ALL of their popular products are famously built without customer interaction.

The opening two sentences are this:

"The basic idea is that to get a startup of the ground, you don't need much. In particular you don't need much money and its possible and even desirable to grow your startup organically from the ground up instead of trying to get a cash infusion from a VC."

This is incorrect. The basic idea is that you can apply the scientific method to a startup. It has absolutely nothing to do with avoiding investment.

The top down approach requires you to have deep domain knowledge to build a successful product around your idea. The bottom up (lean) approach starts the same way but is based on measurable feedback. In both approaches, you still start with the idea. Domain knowledge MAY improve your odds of creating a good product but plenty of people have done/tried this and failed miserably. The lean approach improves your odds of success by actually trying to measure success whenever possible and use those measurements to validate or improve the process. The days of "if you build it, they will come" are numbered in the face of more data driven methodologies.

On the flip side, the lean approach typically results in a new product that is only incrementally different than the one it replaces whereas in a more top-down approach, there is the possibility of creating something game changing.

Some context here:

http://khitchdee.forumatic.com/viewtopic.php?f=21&t=297

(The site is a forum being used as a blog / CMS.)

I have my own reservations about the Lean Startup book -- but most of it comes down to an allergy to it that I developed from reading HN. I got around to it last week on a personal recommendation and it wasn't half bad.

I'd say the biggest problem with how people approach it is that they fixate on the MVP rather than the cycle. The MVP is just a particular cycle.

As benjaminsull pointed out, the Lean Startup disclaims coverage of domains with well understood product models and mature feature-development ecosystems.

I'd say the biggest problem with how people approach it is that they fixate on the MVP rather than the cycle. The MVP is just a particular cycle.

It's even worse than that. The problem I see is folks who focus on particular forms of the MVP. Not every damn product starts best with a launch page of some Google Adwords.

Lean Startup is neat. I'm finding a bunch of ideas from it stupidly useful. But it's going throught the Hype Cycle (http://en.wikipedia.org/wiki/Hype_cycle) amazingly quickly.

Yeah, it's not half bad. Like I said, I tend to instinctively avoid whatever book is being waved like a bible in HN circles.

So as I say, it was a surprise to see how different the expressed vision of MVP seems to be from what I got from the book. I got the impression that it was the first thing you could actually give to a user and have them use. Even if it was buggy and had a single feature.

The launch-page "MVP" isn't actually a product. It's market research.

I know this is somewhat heretical - but I don't actually like The Lean Startup book that much.

I was pointed to Lean Startup back in 2010 as a solution to some of the problems I was looking at. The folk I talked to, places like the LSC mailing list, and reading the online/offline writings of Steve Blank, Eric Ries and many others helped a lot.

I was really looking forward to Eric's book - but it didn't really hit the spot I was hoping it would hit. It's very much aimed (I think) at explaining the ideas to the world in general - rather than being a guide to folk who want to apply the ideas. There's nothing wrong with the former, but I was hoping for the latter.

Well if I had a complaint about Eric's book it's that there's a ton of undergraduate-essay padding at the front end. It could have been about 2/3rds as long and said all the same stuff.

If you do a copycat with the sole goal to copy a business model that works for someone else or might work for someone else in the future (exit driven) it might be a "top-down" situation.

Let me make an example: There is Stripe.com and some German clone called paymill. Of course the clone didn't need to start lean because Stripe already validated the market for them. On the other hand beeing non-lean and non-agile will be very painful, when the "original" pivots or expands into new markets/offer products that are not validated yet)…

Top down is also covered in Peter Thiel's course under the name of 'determinate optimism'.

http://blakemasters.com/post/23435743973/peter-thiels-cs183-...

Big multi-year projects like Tesla Model S or Tony Hsieh's Downtown Project are much more top down than bottom up.

This is merely a reflection of the amount of resources required to build a valid MVP of a product like an electric car, revitalising an urban space or building a space launch vehicle.

So, no true Scotsman?

In what sense? No true scot would be excluding these examples from consideration based on unrelated information. If anything, I'm skirting whether or not the lean startup model is actually falsifiable (it is).

This is actually a pretty traditional model. It even has a name: waterfall. :)

If you have a great team you can probably pull it off with a waterfall method. Team almost always beats the methodology used. A bad team will still fail following the latest leanest meanest methodology.

No methodology is going to be perfect in every situation. Take what can (if anything) from every new methodology you encounter and mold it into something that works for you and your team. The key concept I took from the lean movement was "continuous validation" for your ideas, like continuous integration for your code. Also known as fail faster, ship often, etc. If you're going to do something risky, measure the results and react as quickly as possible. The sooner you discover you're heading down a dead end, the sooner you can perform a course correction.

There's a place for every approach and the one you mention works well in many situations. Look at Apple!

However, Lean handles one of the great problems with this one (which is quite traditional as aytekin has already pointed out): it can take more resources than your typical startup might have. Sure, they can raise capital somehow, but that too could require more resources than are available.

Often, an entrepreneur will want to start from scratch, and may have no experience building a product or a business. That's where the Lean Startup approach does well.

Apple is smart in that they let everyone else do the iterating for them, then extract the parts that work best with a laser beam focus. Apple have no original, 'top down' ideas.

Ah yeah, good point - that's mostly true yeah :)

The biggest problem with the lean startup stuff is it results in this ridiculous false dichotomy. There is no "top down" or "bottom up" thing. If you assume you are in one of these two buckets you already lose because you either are abandoning your creative thinking (bottom up) or abandoning listening to what customers are telling you (top down.) I realize the common retort to this is that all startups need to have top down thinking ("vision") but the bottom line is the lean startup philosophy is at the very least poorly communicated because the large takeaway from all the writing on it is that startups should simply perform customer development and let the market dictate most of the actions they take and the products they build.

It's always a matter of deciding when to ship and how much time and resources to commit between data collection points (ie, releases.) This is a higher order function and changes as well in reaction to how you explore the solution space. Different domains have different constraints as to how fast you can iterate. Tesla, for example, is forced to iterate much more slowly than your average social gaming app can. Shipping too early, or shipping too late can result in wildly different outcomes. Drawing the line correctly is a matter of experience, taste, common sense, and luck.

Everyone knows you should try to minimize waste. Everyone knows you should prefer iteration over big releases. The question is always what constitutes "viable". A car that has three tires is not viable, and would be a market failure. The fourth tire takes a little more effort but makes the difference between success and failure. The problem is it takes some intuition to know people need all four tires to be willing to pay for a car. It also takes intuition to know that the three-tired car being a failure does not mean you should throw in the towel. I don't care how many metrics you measure, you can't use metrics to determine where an asymptote is going to appear because it's not going to be a part of any logical predictive model. Disruption is just that, disruption, and usually comes about due to a synergy of several things that were previously unknown to have magical properties when combined. You generally can't hill climb to find to these things, unless you set your step size (alpha) to a "Goldilocks" value: not too large and not too small.

I understand the appeal of creating a mental framework and slapping a label on it to help people think about and communicate ideas clearly. I think there are definitely some important ideas contained in the lean startup philosophy. But in the interest of creating a conceptual framework of absolutes like any other dogma the lean startup philosophy has large gaps in it's methodology that fail to explain for the large number of startups that launched products that many people would consider well beyond "viable", but were successful because their definition of "viable" happened to be the right one.

>If you assume you are in one of these two buckets you already lose because you either are abandoning your creative thinking (bottom up) or abandoning listening to what customers are telling you (top down.)

I find I write better code when I deliberately abandon my creative thinking. It's counterintuitive but it works.

Yes, but do you build better products? In my own experience success or failure is essentially non-correlated with applying deep creative thinking or raw hypothesis hill climbing. Both seem to work in different circumstances. The lean startup method claims that it's anti-correlated since you will fall into the trap of acting on poor assumptions. But guess what, sometimes your assumptions are right and you have a unique perspective to create something almost nobody else would think to.

Also, in my experience it's the subset of successful products that involved more creative thinking that I am most proud of and look back on my time working on more positively. Hell, this is even true for some of the failures. The raw hypothesis driven successes were high-five worthy, but not pride-worthy (other than perhaps a pat on the back for demonstrating my faith in the Church of Science), and the failures tended to be fairly painful since the work itself was unrewarding intrinsically and the outcome was negative. Nothing is really worse than hacking out code in misery because, counter to your intuition and your gut, signals point to it potentially driving a metric up, and seeing it fail anyway. You have to resign yourself to the process and tell yourself that, despite being right, your intuition is meaningless -- not a fun exercise as a human.

It turns out, in my experience, that the most successful creators are talented at both: trying lots of stuff and learning from their failures, but also having that sense of empathy, intuition, and taste to simply "not be wrong" more times than your average person. You can't A/B test your way to having an opinion. And you can't let an A/B test let you change your opinion on the things that matter.

I agree with both. I tend to have Red Days and Blue Days. Red Days are coding days, working from my plan, no creativity allowed - that's the enemy of lean code.

Blue days are creative days, very little coding - maybe a bug fix if creative me says that bug is really impacting experience. These days I plan big picture, I talk to potential customers and testers. I draw big picture plans on the whiteboard that Red me will implement.

Can break up a day too, but you can never ever be creative and engineer well at the same time. At least I can't, nor can most people I've observed.

Did you know that the entire lean movement grew out of car manufacturing industry? Lean applies different to assembly line and engineering, but it still applies. The key premise of lean is waste reduction, on car manufacturing it means ironing out bugs from repetitive processes (among other things), in software engineering it means reducing the cumulative cost of all iterations required to ship a useful product.

Reply on news.ycombinator.com