Michael Schrage wrote a great op ed piece for the Financial Times on November 8. Under the headline of “For innovation success, do not follow where the money goes”, Michael rips in to those who equate R&D spending with innovation in response to a recent UK Department of Trade and Industry report focusing on global R&D spending.
I urge you to read the whole piece; it is unrelenting in its attack. Let me just quote some of the juicier pieces:
Any policymaker, chief executive or innovation champion who relies on R&D intensity and R&D budgets as a meaningful or usable metric to assess global competitiveness virtually guarantees shoddy analysis and distorted decisions. Few things reveal less about a company’s ability to innovate cost-effectively than its R&D budget. Just ask General Motors. No company in the world has spent more on R&D over the past 25 years. Yet, somehow, GM’s market share has declined.
Michael makes clear that R&D spending is only an input:
The simple fact is that R&D spending – whether in euros, dollars or as a percentage of sales – is an input, not a measure of efficiency, effectiveness or productivity. Ingenuity, invention and innovation are rarely functions of budgetary investment.
He also makes an important point about some of the most innovative companies in the world today:
While Wal-Mart, Texco and Dell have miniscule R&D budgets, their quality, procurement and growth requirements have probably done more to drive productive innovation investment than any five European Union funding initiatives.
Finally, Michael draws some important implications for public policy:
Growing market competition, not growing R&D spending, is what drives innovation. A successful innovation policy is a competition policy where companies see innovation as a cost-effective investment to differentiate themselves profitably.
Right on! In my consulting career, I have participated in many analyses seeking to draw a correlation between R&D spending and business performance in specific industries. The conclusion: there is absolutely no correlation – what you get is a scatter diagram.
I only wish that Michael had gone a bit further and spent more time attacking a related fallacy: equating patents with innovation. At least this approach focuses on outputs, rather than inputs, but it focuses too narrowly on only one kind of output. In effect, it equates innovation with invention. This immediately narrows the focus to product innovation and largely ignores process and business model innovation. The longer I work on innovation, the more convinced I have become that process innovation is far more powerful than product innovation – it has a multiplier effect that product innovation can rarely match.
Bottom line, the only effective measure of innovation activity is the rate of productivity improvement in an enterprise – the growth in value added generated per employee. There are lots of ways to “game” productivity in the short-term – for example, by raising prices or by cutting staff and forcing the remaining people to work harder. But these can’t be sustained – over time, they generate diminishing returns or, in the extreme case, lead to productivity erosion. That’s why static productivity measures can be misleading. What really counts is the ability to sustain and amplify productivity improvements through innovative products, process improvements or new business models.
From a competitive viewpoint, what matters is the relative rate of productivity improvement. R&D spending and patent filings will matter little if they do not translate into faster productivity improvement – in fact, they can be a significant distraction. Those who understand this will have a significant edge as competition intensifies in the global economy.
Solid article/blog post, John!!!
Amen x infinity. :)
Posted by: olivier blanchard | May 28, 2009 at 06:50 PM
Chris - The measure for innovation that I propose is hard to apply and simply can't be done for private companies without access to inside information.
The McKinsey Global Institute has invested significant effort in capturing productivity data by industry across many different countries, but it has done far less in tracking the rate of productivity improvement and, with very limited exceptions, has not focused on the enterprise level. A great synthesis of this work is available in The Power of Productivity by Bill Lewis, the former head of the McKinsey Global Institute.
I have frequently done the analysis on specific public corporations that I proposed in my posting. It is time-consuming but extraordinarily revealing. One can take short-cuts and look at simpler measures like profit per employee or even revenue per employee over time. These are suggestive and informative, but they also have clear limitations.
Obviously, looking at R&D spending or patent filings is much easier but it is fundamentally misleading. The problem is not R&D spending or patent filings - it is achieving market impact from all forms of innovation initiatives.
Posted by: John Hagel III | November 15, 2005 at 05:23 PM
Booz, Allen, Hamilton recently reported on their Global Innovation 1000 research in "Money Isn't Everything." Their study makes a similar point about R&D spending as well as patent output as indicators of innovation. The report noted:
"We found only one strong performance correlation. Higher R&D-to-sales ratios were associated wtih higher gross margins: the percentage of revenue left over after subtracting the costs of materials, labor, manufacturing, and direct shipping, and after paying other expenses incurred in making the products or services sold."
Posted by: Larry Irons | November 13, 2005 at 08:18 PM
As someone who heads a research department mandated to produce product innovations, I appreciate your post on the lack of information provided by R&D spending and # of patents as measures of a company's innovation.
If we limit our interest to product innovation only, I'm curious how those two measures stack up. I would guess that they are still only loosely correlated, and in some cases correlated only through a co-dependent relationship with another independent variable like R&D size.
Posted by: bwedwards | November 13, 2005 at 05:08 PM
Most of the uses of stats on R&D spending and patents granted are in regional competetiveness studies. If they're useless, what would you suggest as an alternative? The productivty growth/employee example you offer is hard to get from outside, especially for private companies.
Posted by: chris anderson | November 13, 2005 at 01:49 PM