Bell Labs, Skunk Works, and the Crowd Sourcing of Innovation

I’ve noticed that we hear a lot less from corporate research labs than we used to. They still exist, though. Sure, Bell Labs is owned by Nokia and there is still some hot research at IBM even though they quit publication of the fabled IBM Technical Disclosure Bulletin in 1998. But today innovation is more likely to come from a small company attracting venture capital than from an established company investing in research. Why is that? And should it be that way?

The Way We Were

There was a time when every big company had a significant research and development arm. Perhaps the most famous of these was Bell Labs. Although some inventions are inevitably disputed, Bell Labs can claim radio astronomy, the transistor, the laser, Unix, C, and C++ among other innovations. They also scored a total of nine Nobel prizes.

Bell Labs had one big advantage: for many years it was part of a highly profitable monopoly, so perhaps the drive to make money right away was less than at other labs. Also, I think, times were different and businesses often had the ability to look past the next quarter.

Bell Labs can trace its roots back to Alexander Graham Bell himself. Taking the money from the French government’s Volta prize (about $270,000 adjusted for today’s dollars), Bell started a lab in Washington, D.C. called the Volta laboratory. Bell was very interested in sound and the deaf (you can hear an 1885 recording of his voice from the Smithsonian’s Volta collection, below). And there were other research labs at the time with AT&T and Western Electric. By 1925, Bell Labs formed in Manhattan to consolidate the various research efforts. By 1967, the headquarters moved to Murray Hill, New Jersey although there were facilities in many other areas, too.

Nobel History

Bell had a string of firsts, many of which won the researchers Nobel prizes. In 1937, for example, the prize in physics when to a Bell researcher for demonstrating the wave nature of matter. In 1956, the team behind the successful transistor won. In 1977 and 1978, the lab won in disparate areas: understanding the structure of matter and the discovery of the cosmic microwave background radiation. More recently, the lab’s researchers have won for cooling atoms with lasers (1997), explaining the fractional quantum Hall effect (1998), inventing charge coupled devices (2009), super-resolution fluorescence microscopy (2014), and optical tweezers in 2018. Optical tweezers. No kidding. See the video if you want to know more.

The Nobel isn’t the only prize Bell Labs scientists received. Five researchers have won three Turing Awards for topics ranging from error detection and correction, to Unix. Some of the lab’s significant work, too, never won a prize. Karl Jansky set the stage for radio astronomy while investigating why shortwave radio links were experiencing static. Early TV, speech synthesis, computers, cell phones, and more all came from the halls of Bell labs. Why not put it to music? Oh wait, someone from Bell Labs — well, technically Lucent — did back in 2000, and you really should watch it.

Other Notable Labs

Bell wasn’t the only game in town. IBM has an active research organization. Until 1998, they published a technical disclosure that was a joy to read and breathtaking in its scope with topics ranging across many different scientific disciplines. The only remains of it are chopped up on a patent search website.

Lockheed Martin had its Skunk Works (named after a Lil’ Abner cartoon venue that made moonshine). That lab turned out the U-2, the SR-71, the F-117, and many other high tech airplanes — discussed in an interview about the efforts.

Skunk work had evolved into a generic term for these kinds of innovation labs. Typically, they are set apart from the normal company and have a streamlined bureaucracy. Usually they have a reduced drive to make money in the short term. This makes sense. It might take decades of research to come up with, say, a transistor, but look at the payoff from that. IBM’s version of the skunk work turned out the original IBM PC.

One differentiator among labs was if they handled their own marketing. IBM and Lockheed were able to bring their innovations to the market. Bell sometimes didn’t do that directly, but licensed the tech out successfully, in most cases. Then there was Xerox’s infamous PARC which invented modern desktop computing only to watch it get copied away — ironic, really, when you think of Xerox’s core business. Kodak developed — and ignored — the digital camera.

Death of the Lab

You don’t see nearly as much work happening in this kind of lab environment as you used to. Many companies have reduced or shut down this type of operation or have quietly shifted over to more mainstream approaches. Steve Blank wrote about this in Forbes back in 2014. His premise is that these labs are dying — and should die — because companies should know now they have to innovate continuously in today’s market.

I’m going to guess that Steve has a better business background than I do so you might want to take his word for it. But for the record, I disagree with his idea.

Consider boarding an airplane. Studies show that the way we board airplanes today is actually less efficient than just letting everyone get on whenever they can. Turns out that airlines like to have boarding classes to reward higher-fare customers or frequent fliers. But even then, you can mathematically come up with better boarding strategies. Why don’t airlines implement them? Simple. You can’t get people to actually obey them. If you fly much, you’ve doubtlessly seen people who try to board with the wrong group or — on Southwest — aren’t in numerical order. A few of those will destroy your carefully tailored boarding algorithm.

It is great to say that innovation shouldn’t be the purview of a technical elite in a company. The whole company should innovate all the time. Sure. If that were possible, then that sounds great. The problem is, you have to have a lot of elements in place to make that happen. The least of these elements is engineers to actually innovate.

Let’s face it. Not everyone is going to invent the next transistor or cellphone technology. That’s not bad. We need people of all skill levels. But they can’t all be working on pie in the sky. So our first job as CEO of our fictitious always innovating company is to find those people. Now we have to realize that you can’t predict your innovation. You can’t sit down and say, “Today I will invent something valuable!” So the finance people are going to have to understand that. So are the shareholders. There will be failures and dead ends. Our job as CEO is to explain all of those things to people who are driven to make money this week, this month, and this quarter. If we do our job right, maybe in 10 or 15 years we’ll have a big payday. Too bad the average CEO only holds the job for five years.

So What Works?

One thing Steve Blank said that I do agree with is that the skunk works is like a startup company inside a big company. In fact, we see a lot of modern innovation coming from very small companies with the help of venture capital. Of course, for every one of those that flies, a few dozen crash and burn.

This is almost crowd sourcing of innovation. Let little guys try things. Some will fail. The ones that look good, get money and grow until someone buys them out. Not necessarily the most efficient thing and if you aren’t on the plus side of the equation, not a fun process. It is also difficult for a small organization to do pure research for decades, so I think this system will not generate the kind of advances we’ve seen in the past.

That’s why I’m not in favor of killing the in-house research teams. With the right vision, those teams can work on projects with time and resources you will probably never have in a small startup. Treat them as startups if you must, but you need the stability and resources of a big company without the profit pressure and bureaucracy. That’s getting harder to do, but the companies that can do it stand to win big. That’s not to say that startups don’t also have their place in innovation.

Either way, I think the idea that you are going to get a large 50,000 employee organization to innovate like a startup is difficult to imagine. Jeff Bezos has famously said treat every day like it is day one. He has probably done a better job at that than most. But you are never going to be able to maneuver an aircraft carrier like you do a PT boat. That’s why you have some of each.

For the Future

Of course, a lot of startup companies begin with hackers — the kind of people who read Hackaday. Even more of them hire people we would think of as hackers very early. You have to wonder if that would help some of the corporate research? Of course, many of us don’t want to work for a big corporate overlord.

It might be worth a few tweaks to corporate culture — at least at the skunk works level — to accommodate the hackers. History is full of examples of agile hackers doing what the establishment can’t.

40 thoughts on “Bell Labs, Skunk Works, and the Crowd Sourcing of Innovation

  1. Flame On

    As an ex ‘labber, the reason that the Labs died was because of incompetent, greedy and quite simply stupid management. I can’t claim credit for anything like C or Unix, but I worked on many exciting projects. Back in the day, the Labs hired excellent folks, many from other countries and cultures. It was where I learned to respect the contributions of others, no matter where they came from.
    Yes, there was a lot of “waste” in terms of projects that never saw the light of day, but there was also a lot of successes as well. It’s a shame the business types can’t figure out how to leverage the research jewels that perhaps are not as “efficient’ as they would like.

    Flame Off.

    1. Well, a few decades ago the middle and most of the upper management was populated by engineers, sometimes the best of them. Nowadays, most of the times it’s the opposite. The management is either non technical people or the worst engineers and the real engineers struggle to find their way to do the job properly.

      Another issue is that engineers today are specialized in some specific part of a specific domain. Back then, engineers were skilled in several different domains. They had a wide range of skills. The more you go back in time you realize that engineers were skilled scientists. Now you have an expert in Kubernetes only.

      Anyway, the problem, in my opinion, has roots in the way that modern society has become. I won’t get into the details because it’s a huge and grey area, but I think the modern engineering is the mirror image of the modern society in general.

      1. That’s certainly the way the United Kingdom Atomic Energy Authority has gone. It used to innovate and now seemed to be populated with dead wood.

        My point is not to be critical of them but agree with dimtass. Perhaps there needs to be a new round of labs as it seems they enivitably stagnate on the management side.

    2. Yeah, in some places inevitable. Bigger cause might be the changes in taxes on R&D. The big labs dropped like flies. I remember the rich pickens at Boeing Surplus when they closed down Boeing Research, which was huge. There was Hughs Malibu, which survived in a much diminished form and where Robert Forward did most of his “gravitational engineering” work, and lots of others. It was too industry-wide to all be bad management.

  2. In house R&D teams work when the corporate environment supports it and stays out of their way. Lockheed Skunk works keeled over because the environment isn’t the same as it was in 1960. Management morphed it into just another division of Lockheed and staffed it with MBA types.

    Northrop had to buy Burt Rutan’s company because they became brain dead and could no longer innovate.

    The CEO’s and executives see the R&D department as a money sink. If it’s not making money for them right now, it gets cut. This drives out the talented innovators and then management blames the R&D dept and then cuts them some more.

    With big business it’s always the same MO: “The beatings will continue until morale improves”.

    As long as management is focused on the next quarter innovation will not come from big companies. They will just continue to buy or steal the innovation from smaller firms the way Microsoft and others did over the years.

  3. See “Spin-In”, i.e. the opposite of a spin-off. Companies fund “startups” to work on ideas originated by their employees. If the new companies fails, they can write of the investment. If not, the get to pull it back in to the mother company. It’s not pure research but it does allow them to develop new tech in a marketable manner.

  4. “Of course, for every one of those that flies, a few dozen crash and burn.” And many hundreds more get ignominiously acquired. There are so few startups that “fly” it’s ridiculous.

  5. AYFKM? I hear great things coming out of research every week. Ok, Bell Lbs may not be the brain trust it once was. But look at Google and Nvidia. Nvidia is doing just amazingly cool things with AI these days. I don’t think corporate research labs are dead, it is just the old guard is shifting. Who knows in 50 years google may be a stuffy old company like IBM and some upstart may be hogging the limelight.

  6. The issue is that corporate research labs are generally in the business of long-term, high-risk, high-reward research that quarterly-minded bean-counters are challenged to put a value on. Things can be fine as long as the broader financial going is good, but the moment things turn remotely south, corporate labs attract attention as non-(immediate)-revenue-generating operations and get mounted squarely on the chopping block.

  7. I like research and I like technological innovation, and over time I just keep seeing more of it.

    Is this a debate on fewer patents being obtained or less research being done in total, or how fewer companies are responsible for direct innovations?

    Because a quick search shows that more and more patents and research papers get published every year.
    (see this chart https://www.uspto.gov/web/offices/ac/ido/oeip/taf/h_counts.htm .)

    So, is this a debate about how much innovative research gets commercialized?

    In addition, many companies are now doing excellent technology research (Mitsubishi, Glaxo-smith Klein, IBM, Google/Alphabet, Tesla…), the difference now is that instead of one tech supergroup attracting all of the eyeballs (Like Bell Labs used to, because they were the only one who could do it), there are many more groups doing a huge level of awesome research in house, so it is harder to shine because
    a) Talent has more places to go (Not just one lab, not just one company)
    b) More stuff keeps coming out (you have to say look at that to a wider array of stuff).
    c) Some of the stuff coming out may be more esoteric and only appeal to a more niche market (I’m looking at you cryo-electron microscopy 2017 nobel prize).

    In addition, many corporations do a lot of funding at university labs to
    a) Attract talent
    b) Develop people
    c) Develop people who can both innovate and who can train other people but who would be an ill fit in a corporate research lab
    d) Develop new technologies
    e) Commercialize new technologies.

    Maybe a company gets better tax breaks on funding this sort of research at a university than doing so in-house? Also I know a bunch of very useful university professors I could easily see becoming miserable and useless if they had to do their work in a corporate environment. Does this sort of partnership work better for both parties? Can a skunkworker be a good fit at a university and vice-versa?

    I agree with the assertions made by Steve Blank’s that a company must innovate continuously, because disruption is continuous, and that companies must listen to their clients when they recommend new products that they would like to see. I disagree that Skunkworks need to die out completely though, because sometimes you do need to engage in blue-sky style pure research to come out with new products and methodologies that you can package into technologies. At the very least a skunkworks department probably develops a lot of people with above average knowledge of industry affecting cutting edge research.

    When I read this article, and the amount of complaints regarding how once a major corporation provided so many technological innovations and was such a hub of R&D, that I couldn’t help but think Nortel Syndrome. The belief in a single supercompany that can dominate it all. A Nortel is dangerous, because it represents a single point-of-failure company that can take many people with it when it falls out of favour.

    Nortel syndrome is named after the Canadian telecommunications company Nortel that was Canada’s major tech industry player, until they dramatically deflated and sucked the wind out of Canada’s tech industry when they stopped getting tax breaks for R&D. It collapsed in a rapid and epic manner, laying off thousands of experienced tech workers, and the stock dropped from over 100$/share to 5$/share. This was a classic case of over-concentration, and the need to have more than one player in any industry.

    1. OK, but from your description even if there were several players, they all would be hit and dead from stopping of R&D tax breaks. What happened with the thousands of experienced tech workers, did they dispersed and seeded a new wave of Canada tech giants? Or they just migrated over to the USA?

      1. Biggest effect was that the computer/electrical/physics/compsci graduates of 2001 were competing directly with phd’s and people wiyh 12 years of experience willing to take entry level wages, and many of those graduates found themselves unemployable and many never got into the workforce as engineers.

        1. Biggest effect was that the computer/electrical/physics/compsci graduates of 2001 were competing directly with phd’s and people wiyh 12 years of experience willing to take entry level wages, and many of those graduates found themselves unemployable and many never got into the workforce as engineers.

          My friend was in that sucking hole. To this day, when his University sends alumni donation requests he and a few of his classmates send in pizza coupons.

  8. Something that can happen with a research project is the company sees no future in it then the team that invented the thing quits, forms their own company and becomes very successful.

    It would be interesting to know how the Diablo team was able to separate from Singer without legal issues.

    That’s how Diablo printers was founded. For some reason Singer decided they should get into the burgeoning business of printers for computers. Their R&D team created the daisy wheel “Hy-Type” (they claimed NOT short for Hyper Typer) printer. Singer decided the company should stick with sewing machines. Diablo would go on to produce many variants of the Hy-Type, Hy-Type II, and 630 daisy wheel printers before being knocked out by lower cost and faster 24 pin dot matrix, laser, and inkjet printers. It didn’t matter anymore that a daisy wheel could crack out better than typewritten text far faster than a human, or a teletype. When you wanted doublestrike bold underline, a daisy wheel slowed to a very loud crawl, and it could only do one font at a time. Graphics (other than ASCII style with an ANSI wheel) were impossible, so were different sizes and fonts on the same document – without using tractor feed and making several passes with different wheels.

    Had Diablo moved quickly into dot matrix, then inkjet and/or laser printing the company might at least survive as a brand name in some conglomerate.

    A retro tech article on the rise and fall of Diablo printers would be an interesting read, especially if it includes something about an infamous failure. Some motor driver chips used on a PCB for one of their models were encapsulated in a foreign factory with poor climate control. Some got a bit of condensation embedded. Eventually the moisture could makes its way to the chip die and short it out, causing one or more of the motors, and/or the daisy wheel hammer to turn on at full speed.

    Imagine the chaos of a 15″ wide carriage printer suddenly spewing 4 copy greenbar while the heavy carriage tries to slam through one end.

    1. And the daisy wheels lived on in use for many years, printing onto clay coated paper to produce camera-ready copy. But we don’t use camera-ready copy anymore either…

      Change or fall behind to die.
      I still miss all the different film emusions I could choose from 35mm through 4×5. The good digital is very very good, but there’s something about a VPS trans that digital can’t touch.

      1. I had a Xerox 820-II Information Processor with a Diablo 630. Even inside the massive wood and plexi, eggcrate foam lined cabinet on wheels, that sucker was LOUD.

        It ran CP/M 2.2 and WordStar, which is what it was mostly used for, to write and print letters in the Time Before The Web. I also ported a few text BASIC games to it like Star Trek and Trade War. I even did a couple of small hacks. Built a cable to bring out the internal parallel port to connect a 9 pin dot matrix printer and upgraded from single side to double side floppy drives by moving one of the ground wires at both ends of the cable to the side select location. Yup, Xerox went so far as to leave one wire completely out of the round bundle of twisted pairs so if you wanted to upgrade to double side floppies you’d also have to buy a specific and hellaciously expensive different cable *with just one more wire* in it.

  9. What about universities? At least here in Australia, universities are often funding their research from industry, so a lot of the innovation and IP generation is being outsourced by companies, basically. Unfortunately the researchers involved rarely get any proper recognition for their work and of course the uni retains any IP rights.

  10. The current crop of large corporations kill innovation. They see a smaller company which is driving forward innovation and take it over to add it to their portfolio then completely change the dynamic of that company.

    They force the company to be measured by a short term year-on-year profit margin to maintain the corporations overall profit margin for it’s share price. They impose huge management process burdens tracking progress and accountability to those margins and deadlines. However, worst of all, they generally impose a risk-averse culture in which all R&D must be goal-driven with some sort of end customer outlined in order to make back the investment.

    The smaller company never again develops a new, innovative, disruptive technology because generally the market is not clear.

    1. “Current crop”??? Are you aware that what you describe is the status quo, everywhere and at all times? Your comment could be lifted intact from a newspaper editorial of the 1930s, describing the auto industry, or from a magazine of the 1950s describing the commercial aircraft industry. Humans will always behave like greedy humans, it is as inevitable as gravity and aging.

  11. As for the death of the innovation lab, that is greatly exaggerated. Yes, there is a move to try to bring the spirit of open innovation into the enterprise to drive digital transformation but I am producing a conference on Open Innovation and, as part of my due diligence, I have a “Innovation Lab” Google alert and I would argue there is now a flurry of activity around the opening of innovation labs, etc..You could maybe argue that a lot of the Fortume 500s have Labs (and a select few have closed or scaled back their labs) but, globally, it’s incredible how many organizations, institutions, governments and companies of all sizes that have new lab initiatives. Just set your Google alert and you’ll see.

  12. I used https://www.flvto.biz/ to DL the 1885 recording MP3 then played it back with Windows Media Player. I don’t know how to upload a screen shot of the graphics eq window but setting the 31.5, 62 & 250 Hz. bars to 0 and the 500, 1000, and 2khx bars to ~ 75% gets rid of the thumps and bumps. I set the 4khz to 25% , the 8 and 16 khz also to 0, and I can hear a lot more almost plainly. Fun to fiddle with,
    Russ

  13. I have worked as an engineer and went for but dropped out of an MBA program in order to have the ammunition to fight these battles with the same words and ideas.
    The current environment where while not law still the only factor considered by execs and board members paid in stock and options is stock price gains, every manager is looking for his or her chance to take home an efficiency bonus for cutting something not needed immediatly.
    When the governance only cares about stock price long term profit and innovation is not a consideration, the invisible hand wants to grab all it can, there is almost no way for consumer pressure to change that because they have little effect on stock price.
    Look at Toys-R-Us or sears, crash the company sell the real estate and any valuable properties including the name of the destroyed company. If you can’t strip the copper from the walls and sell the employees kidneys to make the next quarter you are thinking long term and directly hurting your own compensation package.
    FWIW this is also why on the job training is almost unheard of, outsource the expense to the employee and never show that employee any loyalty.

    1. It doesn’t have to be that way. Privately held companies don’t suffer from these issues. They do not get the press of public companies so most people are ignorant of their benefits. Seek out these companies, work for them, do business with them.

  14. “Then there was Xerox’s infamous PARC which invented modern desktop computing”

    “Uh, excuse me, Mr. President, but that’s not entirely accurate.”

    Might want to google “Englebart”.

  15. “you need the stability and resources of a big company without the profit pressure and bureaucracy. That’s getting harder to do”

    Or, startups can bootstrap, instead of taking on investors. I’m running a tiny boostrapping project, and we’re innovating all over the place.

    True, it’s taking a lot longer. Would be much faster with funding, but it’s investment money, then will suffer from the pressures you describe.

    Well, why shouldn’t the gov’t support innovation in bootstrapping startups, WITHOUT profit pressure? Government already pays off giant companies, in tax write-offs and other corporate-welfare.

  16. Usually these days government and other big contracts are very inflexible and specify exactly what they want delivered.

    There’s no room to sell alternative or new ways of doing things, so big companies that carry out the work, whether it’s a new jet fighter or nuclear site, have no incentive to research things, although they will still be receptive to new ways to do the same things more efficiently.

    Although there is a lot of innovation at big firms such as Alphabet, these are primarily IT-focused, and it’s a shame few companies take the same approach to civil, mechanical, and electrical problems these days, as there’s still a ton of good innovative ideas waiting to be discovered.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.