Four Steps to the Epiphany: the Moby Dick of start-up books

Image: Front Cover; Source: Amazon

If your experience of Moby Dick was that you were constantly aware that you were reading one of the best books of all time that was opening your mind to new ideas if only you could keep your eyes open, you understand.  Four Steps to the Epiphany is the great white whale of start-up books for a reason.  Although it is not nearly as easy to read as his disciple Eric Ries’s more famous book, The Lean Start-up, it is much more systematic.  This books has some profound insights about understanding why some start-ups can do it one way and others need to do it completely opposite.

Instead of abstracting and generalizing the insights, Blank focuses on the issues of managing under extreme uncertainty in their native context.  He tackles every aspect of the non-engineering side of the business.  Most of the book is about how to systematically eliminate the market risk for your product, this will be somewhat familiar to you if you’ve read the Lean Start-up.  However, seeing the original idea and seeing it laid out in full detail, in the context it originally sprang from adds a lot of richness and practicality to the idea.  Blank devotes a good deal of time to understanding how to make technology push and market pull work together.  He covers when to go for broke spending money to enter a market and when to hold back and let the customers come to you.  Most importantly, this comes with some practical steps to discover when to do each.  He even covers how to start converting to mature company once you’ve almost made it.

Much like Melville, Steve Blank will say something really profound and insightful, then launch into a description of whaling–er, uh–start-up processes that are needed to implement that idea.  This can make the book a tough slog, because reading a process description around bed time can definitely have soporific effect.  However, this tough slog is absolutely worth it if your a practitioner in the world of technology start-ups.  You can’t hand it to your cousin that works at a big company and expect him to read it.  This is meant for the start-up community.  If you are a start-up practitioner, get this book and make yourself read it.   You will not be disappointed.  I expect my copy to become much more dog-eared than it already is before it gets confiscated for some future company museum.

So how does this relate to robotics…

Reading this book will further persuade you that many if not most management teams of robotics companies don’t have a clue.  You’ll even be able to look at robotics success stories and realize–wow–compared to software our industry’s state of management practice is pretty dismal.  Many successful robotics companies just fell bass-ackwards into their success.  Many were product driven companies to a fault that were able to expensively keep trying until they finally hit a success.  This is not the same thing as systematically eliminating and consciously balancing market versus technical risk to produce the greatest chance of creating successful business that uses robotic technology to make money and make the world a better place.

We’ve got a long way to go as an industry.  Luckily, now that we know that there’s nothing inherently ‘capital intensive’ about the robotics industry we can start addressing why we have so often screwed it up before.

DARPA is about to show the Navy’s shipbuilding plan is bull****

What is a powerful enough word to describe how the Navy’s shipbuilding plan is wasting thousands of man years and hundreds of billions of dollars on prejudices, untested assumptions, and bureaucratic inertia?

Luckily, DARPA is doing exactly what Congress created them to do way back in the Sputnik era: they are creating and protecting against technological surprise.  It would be fantastic if the Navy would jump on board and run phase 7 of this recently awarded DARPA contract.

Source: DARPA

For those of you who do not come from defense, here is my take on the conflict between how the traditional Navy looks at ships and how DARPA and the embattled progressive minority in the Navy look at naval platforms including unmanned naval vessels.

The big, traditional Navy believes–and they have some experiences that gives rise to this belief–that naval ships ought to be flexible, broadly capable, and completely independent assets.  Take a modern Arliegh Burke class destroyer (DDG-51 class), the backbone of the U.S. fleet, as prime example.  It can deploy itself to the theater of operations, maneuver tactically, sense targets, make engagement decisions, engage the target, and retrograde tactically and strategically from the operation.  Moreover these ships can do almost every mission that they might be called on to do.  They are among the most capable ships at anti-air, surface, and anti-submarine warfare.  Additionally, they respond to things like pirates, search and rescue, and humanitarian relief operations.  Sounds pretty cool, right?  And it is.

However, being able to do everything comes with two main drawbacks.  First the ‘jack of all trades, master of none’ phenomenon is far more likely to be true because of design compromises in engineered systems than it is in people.  Second, adding all this capability costs a lot of money.  These destroyers are about $2B a copy and on the order of $1M/day to operate if you add in everything.  This means that we can only have so many and they cannot be everywhere.

DARPA and the progressive faction within the Navy believe that there is a fundamental change at hand in naval warfare.  Looking at how the Army/Air Force team conducts operations and the improvements in automation and communication technologies at sea, the progressives believe that the tradition of having big capital ships that do everything is outdated.

In contrast to the completely capable Navy platforms, Army units often only do one or two things.   Almost no Army units have strategic mobility.  Most can only do one or two things.  Intelligence units often only have the ability to sense. Artillery units only have the ability to do tactical maneuver and fire, but cannot sense.  Transport units move other units and equipment but cannot fire or sense; sometime they cannot even maneuver tactically.  The Army has huge staff units that do nothing but process information and make decisions to keep all these specialist pieces working in coordinated fashion on the battlefield.

DARPA and the naval progressives believe that a similar future is in store for the next globally dominant navy.  Which we hope will be, but does not have to be from the Unites States.  They envision swarms of inexpensive specialist vessels such as the one DARPA is building running around coordinated by a few manned ships.  The components of these fleets would be optimized to do a couple things well, be relatively–we’re talking about defense here–cheap, and be deployed in large numbers.

The reason that this is an urgent argument is that there is wide consensus within the U.S. Navy, across both the traditionalists and the progressives, that the Navy will not be able to meet its strategic obligations to our allies and American political leadership in a decade or two.  This is a ways off, but still within the service life of all the ships commissioned in the last decade.  The traditionalist seem to hope for a larger budget and the chance to ditch some missions (the Obama administration just took steps in this direction in their last budget), while the progressives say that if the Navy is receiving half of global naval spending it should be able to keep all its obligations by changing the way the Navy is organized.

The problem is that the traditionalists point out, correctly, that the progressives have not proved their scheme will work.  Then they say that they cannot cut even one ship or submarine which would build about a hundred of these future systems so that this alternate path can be tested.  It sounds to me like someone’s rice bowl is about to be overturned, and deep down they know it.

This is why DARPA’s ACTUV program is so important.  It puts at least one of these vessels out on the water so that people can see with their own eyes that they work.  They will be able to see the SAIC team turning around the vessel in record time and the ship controlled remotely and also sailing autonomously.  They will get to see that anti-submarine warfare works when done with a robot instead of hundreds of men on ships.  DARPA will start smashing the traditionalists reality, or at least put some big cracks in it.

Three cheers to DARPA for their continued work pushing the United States forward whether we all want to go or not!

Robotics capital intensive?! What are you smoking? Don’t believe it.

Robotic manufacturing is not capital intensive, contrary to the popular wisdom.  (Looking at you HBS.)

Unless someone can bring data to the contrary, we should treat this issue as thoroughly decided against the  conventional wisdom.  As we saw previously, robotics companies do not need a lot of fixed assets.  Now, we will see why people who blithely repeat the conventional wisdom that robotics companies are capital intensive are wrong–even if they claim robotics companies are hiding their true use of capital.

First off, robotics companies’ balance sheets look like technology companies’–the internet kind, not the aerospace/industrial kind.  Robotics companies have lots of cash and relatively little else.

Second, robotics companies have gross margins that even companies that don’t make stuff would envy.  The robotics gross margin would probably be even higher if iRobot and Aerovironment were not defense contractors.   There is a lot of pressure to bury as much expense as allowed into the cost of goods due to defense contract rules.   Intuitive and Cognex’s margins are around 75%.  They are even beating Google on gross margin!

Although, it does appear that robotics companies have a bit longer cash conversion cycle than the basket chosen for comparison here, their cash cycle appears to be in line with other complex manufacturers.  Plus, the robotics companies are holding so much cash their management may just not really care to push the conversion cycle down.

Look at the cash required to sell aircraft though!  Manned or unmanned it looks like it takes forever to get paid for making planes.

Although robotics companies have physical products, the value of a robot is in the knowledge and information used to create it and operate it.  The materials are nothing special.  Consequently, these companies look like part of the knowledge economy–few real assets, lots of cash, and huge attention to their workforce.   Next time someone tells you robotics companies are capital intensive, ask them to share what they’re smoking–it’s probably the good stuff–because they aren’t using data.

One thing that a venture capitalist may mean when he says that robotics is capital intensive is that it generally takes a long time and lots of money to develop a viable product in robotics.  This may be true, but it is not really the same thing as being capital intensive.   This observation should cause a lot of soul-searching within our industry.  What the venture capitalist is telling us is that we–as an industry–cannot reliably manage our engineering, product development, and business structures to produce financial results.

This is why the conventional wisdom is dangerous.  It suggests that the lack of investors, money, and talent flowing into our industry isn’t our fault and there’s not much we can do about it.  That is what needs to change in robotics.  We need to get better at management.  We need to start building companies quicker and producing returns for our investors.  If we do that the money, talent, and creativity will start pouring into industry.  Then robotics can change the world.

Notes on Data and Method
Data Source: Last 10-k


Accounts Receivable = All balance sheet accounts that seem to be related to a past sale and future cash, so accounts receivable plus things like LinkedIn’s deferred commissions.

Cash + Investments = All balance sheets I could identify as being financial investments not required to operate.   Assume all companies require zero cash to operate.

Did not account for advances in cash conversion cycle.

A response to Singer

I must be an Aristotelian active soul, because I have no patience for those who hold themselves up as philosophers but suggest no way to actually live in the world.  I have a new post up on Hizook taking the leading intellectual on drones to task for sloppy reasoning–then I suggest a realistic path forward.

Where are the Ops Companies?

Really where are they?  Given how many companies are  building some form of robot it seems like there should be some proportionally greater number of companies out there forming to implement, service, and operate these robots.  Where are they?

Frank Tobe isn’t finding a lot of them forming in his start-up list.  Even the RIA seems to have fewer integrators than suppliers.  AUVSI has many more Lockheeds and Insitus than VT Services.  One could make a case that this is characteristic of the peculiar industries that we’re looking at.  The robotic counter example is perhaps the ROV industry which routinely provides the ROV as a packaged service to the off-shore oil and gas industry.  But most consumer robotics are still selling to early adopters.  Our consumer customers are all people who want tech for tech’s sake, not to mainstream customers that are just looking to solve a problem.

Think about other complex goods in our economy.  Computers have a vast cottage industry associated with servicing and maintaining them which is probably as big or bigger than the software industry proper.  All vehicle industries whether air, ground, or sea have vastly more businesses in the business of selling the services than engaged in construction of the vehicles–even if constructors do manage to capture a large share of the total revenues of the industry.

I think our industry has a problem.  I’ve talked to people at the oil and gas majors and heard straight out that robotics companies are producing robots which have a business case to be used several applications, but they will never be used until a credible organization to is there to provide the robot as a service.   It is a bit of chicken and egg, but I think this applies as you go down the chain, not just in large capital projects.

When doing sampling or reconnaissance, customers want actionable data not a fleet of robots or new employees.  I know from experience that infantry brigade commanders love having drone imagery of the battlefield, but don’t want to worry about having to support the drone unit, they just want to see the battle.  This is equally true in forestry, agriculture, infrastructure, and minerals.

Do I really want to own a cleaning robot? No, I would much rather have a business that comes to my house every week and keeps the place clean whether that business uses humans, robots, or both.

Even in medicine, if I were a hospital operator I’d love to be able to push the risk of owning the robot back onto someone else.  If I can pay per procedure and not worry about utilization, maintenance, or obsolescence–I’m much more game to adopt something new.

To date, our industry has done a relatively poor job of making robotics accessible to people and organizations who aren’t willing to organize around robotics and develop organizational competence in robotics.  Providing robotics as a service could greatly expand the number of potential customers.  I think when we see these businesses start cropping up, we will know that our industry is no longer in its infancy.

How many will die before we fix this?

Four airmen died this weekend fighting wildfires.  They died needlessly, compounding the tragedy of their sacrifice.  There is nothing technological stopping drones from taking over the retardant dropping mission.  The hold-up comes down to bureaucratic inertia and a lack of political leadership and attention to this issue.

There have been at least six aircrewmen killed just this fire season!  How many does it take before we collectively figure out how to do this mission with autonomous or remotely piloted aircraft?


(My criticism of a lack of political leadership is not a partisan criticism.  The technology to remotely pilot fire-retardant tankers has been around for at least two administrations and neither party in congress–which is the body that will really have to act–has shown much leadership on this issue.  But seriously, our guys are getting killed.  The robotics industry knows how to fix this problem, let’s get everyone at the same table and get the barriers cleared so we quit making widows and orphans every fire season.)

Compared to what?

The Pakistani affiliate of the International Herald Tribune has a great summary run down of recent drone strikes under the Obama administration. It details strike after strike against militants that are operating freely in Waziristan and the tribal areas.  It also produces a summary statistic of civilian casualties which, even if taken at face value, seems to indicate that these strikes exhibit a very high degree of proportionality and discrimination.

I’m out here in San Francisco and I sometimes get some troubled reactions when I tell people that I led a drone unit when I was in the Army.  Eventually telling  these people that my drones were not armed quiets them down, but there is a fundamental misunderstanding about how drones are different and similar from other types of war.  Drones unfortunately have some misperceptions about how they compare to other types of military engagements.

I would like us each to sit in the chair of senior commander or official who has just received intelligence from multiple sources (as it seems these strikes are probably based on) that is really good, but not 100% certain.  There is a militant–who the Pakistanis allow to travel, train, and organize unmolested–who is intent on killing not only U.S. soldiers, but also our civilian allies and females who dare to leave the home to get an education.  He believes that it is his god-given moral imperative to kill all these people.  What would you do?

A)  Nothing

B)  Ask the Pakistanis who have been arming this guy to arrest him

C)  A drone strike

D)  A manned aircraft strike

E)  A commando or sniper raid

Okay, even Marines get this one: when in doubt, Charlie out.  Seriously though, C is the only ethical choice.

Doing nothing means that you have let an evil, violent man go about his plans to kill someone’s son or daughter because he or she believed in a world where everyone has the freedom to make something of themselves.  By the same token, B is even worse, because now you’ve not only let him go free, you’re endangering your sources and methods of finding him in the first place by telling the Pakistanis what you know.

On the other side of the spectrum, there is every probability that any method besides a drone attack would be worse when measured by the standard of proportionality and discrimination.  Drones usually strike using the smallest guided air to ground missile in the U.S. arsenal, meaning they tend not to cause more damage than is absolutely necessary to destroy valid military target.  Would a commando team be as precise?  What if the team itself was attacked by our Pakistani ‘allies?’  Could it or a manned plane follow the target and wait until this small missile was the appropriate munition to use to ensure the destruction of the target?  The answer is no, they could not.

Of most concern to all of us who treasure innocent life is the principle of discrimination which means that you are only attacking legitimate targets.  Although this is where drone are most criticized, this is where they beat all other methods of apply force.  Can you imagine any other weapon so precisely attacking only legitimate targets?  There is no other weapon system where every action of the user is recorded, second guessed, and subject to real time absolute supervision by higher authorities.  Once a pilot is over a target his judgment of the scene and need to protect himself while completing the mission takes precedence over a second guessing boss back at base–and rightly so.  Even more so with a commando team, the commander of that mission has complete autonomy and discretion once his forces are committed.

But let’s be clear, commandos and pilots make more mistakes, not less, by being on scene.  They are affected by all kinds of pressures, are in mortal danger, and are being asked to make snap judgments on their own.  They screw-up.  They drop bombs on the wrong house–or even the wrong army.  Snipers in over-watch shoot civilians all the time–they don’t mean to, they just do.  None of these things are war crimes, they are just mistakes–unacceptable tragedies–but still just mistakes not crimes.  I think that most of us, civilian or military accept this.

Drone operators on the other hand work in secure locations, can loiter over their target, and stalk him for days or weeks until they get a clean shot.  They even have lawyers looking over their video feed as they work!  These are carefully supervised, deliberate operations.  That said, this is still war, where even professionals make mortal mistakes with the best information available.  But even or especially in war, we must do our best to protect those principles we hold dear and every analysis says that drones are the most moral option.  To those who would say that drones are immoral, I would ask in comparison to what?


It is well that war is so terrible–otherwise we would grow too fond of it.  -Robert E. Lee