SmoothSpan Blog

For Executives, Entrepreneurs, and other Digerati who need to know about SaaS and Web 2.0.

Why Do So Many Big Companies Suck At Innovation?

Posted by Bob Warfield on May 15, 2009

Larry Dignan writes about big companies spending too much on R&D.  What he really means is they’re spending too much on the “R” (Research) in R&D and not getting anything for it.  Larry’s article is largely focused on Microsoft, which has supposedly spent $36B on pure research over the last 5 years.  What have they gotten for it?  The ill-fated Microsoft “Bob” user interface is all I can remember.

But we shouldn’t just pick on Microsoft.  History is littered with big companies investing in fundamental research and then being unable to cash in on the results.  Xerox had PARC.  IBM invented a plethora of things that mostly others made the money on (Relational databases, RISC cpu’s, the floppy disc, the hard disc, and on and on).  There are many others.  I remember being offered  jobs out of Computer Science grad schools with IBM’s Thomas J. Watson Research Center and AT&T’s Bell Labs.  The Bell Labs offer was particularly interesting.  It included a high salary for an engineer and the promise that I would have a lab budget so that I could buy a VAX computer (or the equivalent in cash) and do anything I wanted with it.  Is it any wonder that while Bell Labs produced a lot of interesting stuff very little of it was of commercial value to AT&T?

What then is the problem?

Some companies get into a mode of collecting IQ points.  They think if they can hire all the smart people, they will have a monopoly on innovation, and therefore they will prosper.  You see all the trappings of this at Microsoft or Google in terms of how they go about hiring.  It is focused on programming or thinly veiled IQ tests (why are manhole covers round?).  I’ve interviewed with Microsoft on three separate occassions, and it was very entertaining to watch the dance that took place.  Every person I talked to (and there were at least 5 for each interview) found an excuse to deliver the same short message at some point in the interview:

–  Microsoft is filled with extremely smart people–much smarter people than the interviewer.

–  Bill Gates is the smartest person at Microsoft (and by implication, probably in the world).

Such an emphasis on smart.  But not so much emphasis on Joel Spolsky’s wonderful “Smart but gets things done.” 

Google seems very similar.  My company, Helpstream, is situated in the midst of Google, right next door to Mozilla.  It’s like being in the middle of a college campus.  Bright nerdly kids everywhere so self-absorbed in their thoughts they regularly step off the curb risking life and limb.  Drive very carefully if you’re in the Google Zone!  They have the IQ tests to join, the focus on good schools.  They even have the moral equivalent of AT&T’s “free VAX for every researcher” in the form of the famous 20% time.  Yet with so many projects running around, so much 20% time, so many IQ points gathered together, there often seems to be little innovation.  They are smart, but other than the core business, do they get things done?

In the end, I don’t believe too much in Big Corporate Research.  Microsoft’s $36B and boatload of PhD’s have not produced a Google-killing search algorithm.  Not even close, and you have to guess they wanted to.  It is beneficial to the Human Race to discover new things, but it is not a viable source of commercial innovation that drives shareholder value.   Knowing new pure research is not the same as having a hot product.  Even having a lot of patents, which is how IBM keeps score is not the same, although sometimes it is a little closer.  There is a unique combination of knowing what the customer’s problem or need is, building an insanely great product, and knowing how to package up a message that has legs that have little to do with what goes on in these Egghead Incubators or in University Graduate Schools either, for that matter.  The chip companies like Intel seem to do relatively better in this regard, but I suspect it is because they’re able to focus their pure research in ways that directly lead to smaller faster chip features, and because what may be their bread and butter will often sound like pure research to the rest of us!

What then, should big companies do about innovation?

Whenever you lack a systematic process or algorithm to produce a result, you are searching through some problem space.  Call it a needle in a haystack, but you are searching a very ill-defined problem space.  The best possible algorithm to do that has been present in nature since the beginning of time.  Call it Darwinian Search or a Genetic Algorithm.  You need to take your dollars and invest them according to return.  Return is simply a measure of fitness as in “survival of the fittest.”  In nature fitness is measured at all times and at every point.  Mutations are extremely short lived.  They don’t get to go on and on “in case they might turn out to be valuable” or because their creator is particularly charismatic and smart.  Life is nasty, short, and brutish with Genetic Algorithms.

There is a model for this already.  It’s called investing in startups and venture capital.  It insists on results pretty short term.  It casts the net wide by not investing too much early on.   It doubles down bets that are working and cuts off bets that are not.  $36 billion would have bought Microsoft a lot of investments.   When you consider their money invested alongside others, it goes even further.  If you don’t want to invest in startups, or can’t figure out how to make it work profitably, invest in pre-IPO acquisitions.

Many companies have done this extremely successfully on the product side.  Cisco is my favorite in the Tech World.  There are people who’ve been acquired multiple times by Cisco.  They build an innovation, prove it out in the market, get acquired, and then use the resources of Cisco to figure out what to do next so they can be acquired again.  Clearly Cisco has an enlightened view of how to use an ecosystem to facilitate innovation and a culture that doesn’t just milk the life out of it after acquisition–products are kept vibrant and growing.  Cisco is not an end of life acquisition model like some others.  My old company Borland can trace the roots of every successful product it had to some acquisition or another.  Philippe Kahn had a great way to embody “Smart but gets things done”.  He very much viewed acquisitions of small companies as building the brain trust, but rather than evaluating their raw talent, he evaluated the market’s reception of their product.  Wonderful concept.  It’s how I came to be at Borland myself.  The pharmaceutical world has long worked this way as well.

The right answer in business is so often to create a process that can effectively manage a decentralized and empowering paradigm.  Don’t confuse this with anarchy or writing blank checks to people to work on whatever they want.  Innovation needs a clear cut strategy and business process just like every other important business problem.

Related Articles

Great minds thinking alike.  Bruce Cleveland at Interwest published this the day after my post:  Spin Ins–A Strategic Opportunity for Venture Capital and Large Software Companies

2 Responses to “Why Do So Many Big Companies Suck At Innovation?”

  1. stevetodd said

    Bob, I work at EMC, which has acquired 40 companies this decade. The latest trend within EMC is to use contests and social media as an innovation vehicle, which I describe here:

    Steve Todd

  2. marcospolanco said

    You are pointing in the right direction in identifying the venture capital model as one to follow by large corporations. But the key is not being beguiled by the surface characteristics of the model. The successful innovation projects are utilizing specific strategies such as Customer Development (Video of Steve Blank, the model originator : and Outcomes-Driven Innovation (StrategyN presentation: According to StrategyN, 75% of innovation investments are lost, and I suspect they are being optimistic.

%d bloggers like this: