How much money do you think Microsoft put into R&D in 2007? $7.1 billion. How about Google? $2.1 billion. And Yahoo is $1.1 billion. That’s over $10 billion in R&D spending in 2007 by just three software firms.
Now, how much capital was put into software startups by venture capitalists in 2007?
Hold it. Hold it.
How does $5.1 billion sound?
While venture investment and income statement expense isn’t an apples to apples comparison, my hunch is that the discrepancy is even greater than 2x. Much of that venture spending went to areas other than R&D (e.g., sales, marketing, G&A), while 100% of the R&D expense from the big firms went into R&D. And while startup firms have revenue that offsets expense (which isn’t captured by venture investment), these firms don’t spend 100% of their venture investment either.
My strong suspicion is that the return earned by investors on that $5.1 billion (in aggregate) will exceed the returns on the $10 billion (in aggregate). If you buy this argument, then there is only one logical conclusion. There isn’t too much money in venture, but rather there are too many good people in large firms.
Why startups are the right place for innovation.
1. The investment model.
Large organizations often lack the will to kill bad projects. It is often more costly to an executive’s career to kill a project he sponsored than to let it continue. With staged financing, venture capitalists don’t need to explicitly kill anything. They just decide not to continue funding it. It’s a subtle difference, but an important one.
It is also important to note that capital often favors experience and pedigree in large organizations while venture capital is relatively agnostic on the issue of experience or pedigree. Especially if your firm is on fire.
The recent emergence of “professional” seed financing should help venture scale (e.g., First Round Capital, HitForge, SoftTechVC). More firms like these in other locations around the world would help tap into the wealth of talent trapped in non-entrepreneurial ventures today.
The even earlier stage Y Combinator is helping to make entrepreneurs of engineers who might have never made that choice. While many people argue that the emergence of these types of firms will hurt venture capital, I believe more early stage professionals will bring more people into entrepreneurship and increase capital productivity — both are very good things for the venture industry and they are clearly great for entrepreneurs.
We need more Y Combinators. It’s a great model, but it can be improved. I would love to see a Silicon Valley based competitor to Y Combinator (they do have a Mountain View “batch”). And then doing the same in Israel, China, India, Canada, and other locations in the US is exactly what we need to scale the venture investment model. Eventually, venture financing will exceed the R&D budgets of the top few firms, but not until the further development of early stage projects emerge to help venture scale.
Centralizing knowledge in order to make good decisions doesn’t scale (see the first paragraph on my post from yesterday). In small organizations, our desire to centralize can work. The larger the organization, the more centralization destroys knowledge.
I really respect Google for recognizing the fundamental limitations of the modern corporation and trying radically different approaches out (letting engineers pick their projects is a bottoms-up way of allocating R&D dollars). I have also heard that Google makes aggressive use of prediction markets to predict future sales.
Why not take this to the next step? Take your entire R&D budget — let’s say your quarterly R&D budget is 100 units and you have 100 employees. Give each employee 1 unit per quarter to invest in projects (instead of stocks). Executives can speak at all-hands or do webcasts to pitch their projects and employees can vote with the firm’s R&D dollars. This will allow for scale on key projects (because popular projects will get more votes) and it will allow for more granular assembly of distributed information. Make compensation partially dependent on employees’ performance and I bet you will get pretty committed ”investors.”
2. Incentives.
With large organizations, it’s very hard to pay 100x more to an all-star engineer or team than a poorly performing one. With startups, that’s exactly what happens. Such outcomes more accurately reflect the distribution of talent and effort.
Good entrepreneurs know that the probability of success equals the number of experiments per invested dollar times the number of dollars. To increase experiments per dollar, you need to be very honest with yourself. As soon as you know a hypothesis is wrong, move on. Instrument your site very well so that you can increase experiment per dollar. To increase dollars, work longer hours. Hire fewer people and make sure they are great. Be cheap. And make sure your CEO can evangelize your company so that you can raise capital when you need it and on good terms.
3. Risk taking.
Why aren’t more people in startups? PERCEIVED risk is a big part of the answer. Does this sound familiar? ”We cannot launch a product before it’s ready. We’re [insert large firm name]. Our customers, the press, investors — they all expect more from us.” So people have meetings to discuss the risks. By the time they are finally ready to launch, they decide that it’s time for a new strategy.
Larry Page recently did an interview with Fortune on innovation. This is what he had to say in response to the question, “How can we increase the number of people doing such [innovative] work?”:
There is a positive selection bias in startups towards an appetite for risk. People have “overcome” their fear — at least enough to be at a startup. When you have an entire group of people who have overcome their fear, you get positive feedback. People egg each other on to break the rules. To “think different.” To be open minded. That sets off a cycle that drives people to throttle risk up and up and…
In most large organizations, the change-antibodies come out to protect the status quo. When you have an adverse selection bias in terms of people’s risk tolerance and a franchise to protect, it’s not hard to innovate. It’s just about impossible.
On top of that, there is often a higher level of ACTUAL risk in a large firm. For example, had Microsoft or Google launched YouTube, the media companies would have sued them right away. Startups can sometimes fly below the radar.
This is a blog about ideas. So the two ideas I am proposing are: (1) build a better Y Combinator in Silicon Valley (or Israel, China, or India), and (2) create a system for distributing investment and other decisions in large firms.
The winner will determine where the R&D dollars go. My money is on the entrepreneurs ;–)
12 Comments
May 23, 2008 at 6:34 am
Nice writing. You are on my RSS reader now so I can read more from you down the road.
Allen Taylor
May 23, 2008 at 6:56 am
Mike,
Good thought on the large enterprise. Innovation and large enterprises do not go together. Managers are wary of risk for the fear of failing or simply for other “untested water” fears. Why try when status quo does the job.
However, these enterprises have the deepest pockets, have access to tonnes of market research data, and have an installed customer base which can be used for cross-sell/upsell opportunities. But, to the chagrin they end up buying the same companies for 10X the price which a small team of their employees could have done sometime ago.
On a different note:
I would love to see incarnations of Y Combinator like coop/incubators/factories in India.
Indus
May 23, 2008 at 7:30 am
working on it.
May 23, 2008 at 2:21 pm
Any idea what the initial $20k went to for something that has been successfully “y combinated” ?? (besides any viral bandwidth costs)
May 23, 2008 at 5:04 pm
Mike,
Funny how some blog posts seem to tap directly into one’s head and just resonate — this is one of them! I have built a startup in Los Angeles and sold it to a big company (which subsequently squashed it), so this makes a ton of sense. As a result I have always wanted to see entities that would encourage the entrepreneurial spirit in SoCal-based technical talent. I founded the SoCal Tech Group to help kick start the effort, but would love to do a Y-combinator style “Startup school” to serve the dozen or so universities that dot the southland. If you are interested please contact me!
May 23, 2008 at 5:21 pm
Large companies should ‘diversify’ their R&D spending into “alternative investments” by adopting an in-house Y combinator like approach. For example, take 1% of your R&D budget (for big companies, even that may be too much money and too bureaucratic, they may need to start with 0.5% of their R&D budget). Set up a in-house combinator! Set up standalone companies where the parent is a minority shareholder. Heck, even if the parent was a majority shareholder, entrepreneurs may go for it if they had significant equity stakes and modest initial capital. Like Y Combinator, avoid interfering too much in the startup, but offer corporate assistance in ways that Y Combinator cannot, e.g. access to the parent’s larger R&D and sales organization and maybe even its group buying power and existing infrastructure where applicable. The most critical thing would be to keep the parent company’s middle management away from mucking with the operations of the startup.
Note that this is a very different from corporate research arms where brilliant PhDs who invent cool stuff end up receiving a patent, a nice plaque, or a few thousand shares of the parent stock. No, the principals of the startups would have real upside.
How many entrepreneurs would jump at the chance to work for a Yahoo/Microsoft/Google for say, minimum wage, $50k in seed money, and 50% equity in their company?!! Sizable companies could do that dozens, hundreds, thousands of times over; the odds of one of those startups becoming a game-changing company is pretty good.
May 23, 2008 at 8:53 pm
Love this post, couldn’t agree more. I can’t resist one quick shot:
I have been arguing heavily for the past few months that Google should adopt a style that is a cross between a keiretsu and an SRI type model. Right now they allow people to start projects in a very good way. They can keep this bottom-up model, which as I understand it is to develop something in your 20% extra time, pitch it, build a small team to work on it, and convince the company to put more resources behind you. The change would be in what happens when projects are identified as having merit and backed.
Currently, small teams are built in house and work towards proving their concept - at which point they become big teams (if deemed the best place to put resources). Internally I estimate there are about 20 good projects a year, and let’s say 5 of which really get backed.
Instead, Google could take these 20 good ideas and invest in each (set 3-5 automated funding milestones, take 49% ownership of the company, give the starting resources to the company, and set it free under a CEO). They would have to keep loss-leader programs internal (analytics works better for them as a free resource that works with adwords then as a separate company), but they wouldn’t have to limit their attention to focusing on strategy for a few products, and could instead focus on helping many possible new ideas and products and looking for opportunities for any two of them to work together. All the benefit of startup innovation, and if any project fails they can roll the IP, team, and anything built back into the GOOG mothership.
May 24, 2008 at 3:13 am
Indus, I agree that big companies have many advantages. Sales & marketing is where many large firms shine. I was really focused on the innovation part of R&D specifically… And I would *love* to see an India-based Y Combinator.
McClure, can’t wait!
Scott, I don’t know how people are using the capital. Presumably for the basics - basic salary, bandwidth, space…
Dan, let’s chat. Can you send me a note at mike [at] shv [dot] com?
Jolly, love your thinking. I think it makes much more sense outside a big company (no issues with consolidating all revenue and costs, no right of first refusal issue, no issues with overall parent situation impacting you). However, I agree that the reality is that many people feel safe under the umbrella of a large organization. And it would be great to see so many of the people who are “check-out” because they lack control over their fate to get back in the game. The world would be a better place.
Tyler, very similar thoughts to Jolly’s. I agree. It would be great to either see large firms cut back on innovation-focused R&D and outsource it by M&R (many are basically doing this already) OR start experimenting with radical models to get better results. You have got to give it to Google - at least they are trying something bold.
May 31, 2008 at 6:32 am
[...] Venture capital, entrepreneurship, big companies, and innovation [via Laserlike] A friend of mine and I were just talking about this… (tags: VC R&D Innovation Corporate America Entrpreneurship Ventures) [...]
May 31, 2008 at 4:48 pm
Another take on corporate R&D: connect + develop from Procter and Gamble (just what von Hippel would advice).
June 4, 2008 at 2:52 am
I would love to see a sustainable model / system for innovation in large companies, driven by an distributed investment / decision-making platform powered by employee choice.
It faces a couple pretty large problems, however, in how decisions get made in companies and in how people work in groups. Asking people who are “adversely biased” against risk (self-selection of people into larger companies) makes it pretty tough to expect people to then think like entrepreneurs.
Not only that, but people in large companies are hired, promoted on their ability to perform, judged by the people above them on the food chain, who have worked and been promoted based on the same self-propagating system. Asking people to then not think about the values and “behavior models” of execs is a difficult request. Not exactly the best breeding ground for a diversity of opinion.
Information asymmetry is also a big concern. Execs lack knowledge of what is going on at the front lines, and front-line managers lack the aggregated big picture. Forcing either to make decisions without the other’s information is tough. Yet this is how large companies operate.
And what do markets require to work? A mass of people, information, diversity of opinion.
I know about the 20% time, the open idea platforms, internal prediction markets et. al. to spur internal innovation, but all of them have success and failure stories, and all work through a single filter at some point.
Still, I’m still a big believer that if people learn to ask the right questions they can deliver the right answers. People in large companies have the ability to be innovative and creative. The difficulty is that almost all large companies don’t ask, tell, or make it valuable for people to ask the right questions.
June 11, 2008 at 3:49 pm
[...] capitalist, Mike Speiser over at laserlike.com has some great insights on this. He looks at the 2007 R&D investments at 3 major companies (Microsoft: $7.1 billion, Google: [...]
Leave a Reply