Professor Andrew McAfee,
Harvard Business School
Professor Andrew McAfee says the power of technology is still far from a universally
accepted idea. As "the crispest possible declaration" of the pessimism he encounters,
he refers to a 2003 Harvard Business Review article titled "IT Doesn't Matter," which
argued that while technology increases productivity, it does not provide competitive
advantage - and senior managers should spend little time looking at it.
From History to Hypothesis
To "beat back that brushfire of skepticism," McAfee and colleagues at Harvard and MIT researched the correlation of company performance and IT investment at U.S. companies from the 1960s to today. In describing his research, McAfee first describes the effects on competition when general-purpose technologies were introduced to our economy at various points. He says the availability of new power sources for manufacturing - from steam power to the electric engine - is analogous to the advent of technology, citing Henry Ford's introduction of the assembly line as the "innovative leap" that helped Ford achieve market dominance. McAfee says that whenever a new general purpose technology comes along, we enter a period of "fairly intense, competitive change-and often a brand new generation of leaders emerges."
To test whether today's competition is having that impact, McAfee and his colleagues first turned to an unlikely proving ground: Ted Williams' .400 batting average. They examined the research of another Harvard professor, Steven J. Gould, on why there hasn't been another .400 hitter since Williams. Gould found that while there has been little variation in batting averages over the last century, the spread, or standard deviation, around that average tells a different story. In a game like baseball, where rules are sustained over a long period of time, performance improves but variation decreases. The game is played at a higher level than it used to be, so everyone converges around the average - there are fewer outliers.
McAfee and his team hypothesized that in the game of business, if technology has changed the rules such that the old rules no longer apply, the opposite phenomenon would occur - we would see increased spread between the best- and worst-performing companies.
Market Test
To test the theory, they collected data points including gross profit margin, epitome margin and market cap from every U.S. publicly traded company from 1960 on. For those companies in industries that spend a lot of money on technology, McAfee expected to see winners increasingly separated from losers - thus proving that IT matters in competitive battles. The findings confirmed his suspicions.
"You've got one flat line where industries that don't spend a lot [on IT] are not seeing much change over time, but in industries that are injecting IT into themselves, the graph starts to diverge and you see increasing performance spread," says McAfee. "The winner, the best and the worst, do not look very similar to each other."
What's more, McAfee believes we might be in an era of "permanently higher turbulence, permanently nastier competition," where innovation and technology can be propagated much more quickly than ever before. If this holds true, he suggests a few investment strategies for success in this environment.
Q&A with Professor Andrew McAfee, Harvard University
Q: What surprised you the most with the results of all of your data?
A: When you do this kind of work, you grab some initial data and you do some quick back of the envelope analyses to see if the finding holds up. When we did that, we honestly -- I thought our computer was broken for a while because what we see on this, came out from our results so strongly, and we looked at it and we said, "OK this can't possibly be right," so we went back and we tried to be more rigorous about our data and to bang on it all of these different ways, and for each of these different outcomes that we're interested in, this performance spread, the turbulence and the concentration, just the raw magnitude of the result continues to be surprising to us. What's even more surprising is that all three of these things seem to be happening at about that same time. So if we use the mid-1990s period as a "before-and-after" differentiator, it's just crystal clear. And because we can so clearly tell the story that the technology world changed as of the mid-90s, that tells us that we are onto the right track with this story.
Q: Do the numbers indicate that if you spend two times as much you get four times the value?
A: We haven't done that work yet for one very simple reason. The problem is that there are no great data sources about IT spending at the level of the company. We got very fortunate because the U.S. government tracks, very consistently, IT stocks and flows at the industry level, year after year. Here is my hypothesis walking in to that analysis. Let's assume a magical data set existed and we could do exactly what you suggest. My intuition tells me that we would find a positive correlation between information technology spending and a business goodness defined as better revenue share, better market share, all these things that we might care about. A weak positive correlation. The reason I say it's weak though is that every time I look at any IT data, the thing that strikes me is never the mean, but it's the variance around that mean. So companies can spend equal amounts of money and get hugely different returns to that kind of spending.
Q: What's your top recommendation or your top couple of recommendations as you speak to folks - just be smarter about where you're spending all the money?
A: My first recommendation is if your senior leadership of the company is not focused on the possibilities that technology brings to them, they had better get there. Because I have yet to see a company where the top leaders didn't kind of, at least appreciate technology if not embrace it. So the first thing I say is, "Look, get on board." The second one is to start thinking about enterprise-level initiatives that you can do with technology. Every time someone goes to shop at Walmart.com, they see user reviews now. It doesn't only happen five percent of the time. I try to make that analogy to the other processes the company executes and say, "Look, if you believe that technology can bring you better processes and business models, one of the things that you need to do is apply those as widely as is appropriate inside the company." When I look at most IT infrastructures and the processes that they support, they're incredibly fragmented -- it's business unit by business unit, the geographies are all different, and there's this complete patchwork of technologies and then business processes on top of that. I say, look, if you have a better mousetrap, it's incumbent on you to spread that around as widely as possible inside the firm. As far as we can tell, technology does not stifle innovation, and does not reduce your scope for doing interesting things. It actually increases your scope for doing interesting things.
From History to Hypothesis
To "beat back that brushfire of skepticism," McAfee and colleagues at Harvard and MIT researched the correlation of company performance and IT investment at U.S. companies from the 1960s to today. In describing his research, McAfee first describes the effects on competition when general-purpose technologies were introduced to our economy at various points. He says the availability of new power sources for manufacturing - from steam power to the electric engine - is analogous to the advent of technology, citing Henry Ford's introduction of the assembly line as the "innovative leap" that helped Ford achieve market dominance. McAfee says that whenever a new general purpose technology comes along, we enter a period of "fairly intense, competitive change-and often a brand new generation of leaders emerges."
To test whether today's competition is having that impact, McAfee and his colleagues first turned to an unlikely proving ground: Ted Williams' .400 batting average. They examined the research of another Harvard professor, Steven J. Gould, on why there hasn't been another .400 hitter since Williams. Gould found that while there has been little variation in batting averages over the last century, the spread, or standard deviation, around that average tells a different story. In a game like baseball, where rules are sustained over a long period of time, performance improves but variation decreases. The game is played at a higher level than it used to be, so everyone converges around the average - there are fewer outliers.
McAfee and his team hypothesized that in the game of business, if technology has changed the rules such that the old rules no longer apply, the opposite phenomenon would occur - we would see increased spread between the best- and worst-performing companies.
Market Test
To test the theory, they collected data points including gross profit margin, epitome margin and market cap from every U.S. publicly traded company from 1960 on. For those companies in industries that spend a lot of money on technology, McAfee expected to see winners increasingly separated from losers - thus proving that IT matters in competitive battles. The findings confirmed his suspicions.
"You've got one flat line where industries that don't spend a lot [on IT] are not seeing much change over time, but in industries that are injecting IT into themselves, the graph starts to diverge and you see increasing performance spread," says McAfee. "The winner, the best and the worst, do not look very similar to each other."
What's more, McAfee believes we might be in an era of "permanently higher turbulence, permanently nastier competition," where innovation and technology can be propagated much more quickly than ever before. If this holds true, he suggests a few investment strategies for success in this environment.
Q&A with Professor Andrew McAfee, Harvard University
Q: What surprised you the most with the results of all of your data?
A: When you do this kind of work, you grab some initial data and you do some quick back of the envelope analyses to see if the finding holds up. When we did that, we honestly -- I thought our computer was broken for a while because what we see on this, came out from our results so strongly, and we looked at it and we said, "OK this can't possibly be right," so we went back and we tried to be more rigorous about our data and to bang on it all of these different ways, and for each of these different outcomes that we're interested in, this performance spread, the turbulence and the concentration, just the raw magnitude of the result continues to be surprising to us. What's even more surprising is that all three of these things seem to be happening at about that same time. So if we use the mid-1990s period as a "before-and-after" differentiator, it's just crystal clear. And because we can so clearly tell the story that the technology world changed as of the mid-90s, that tells us that we are onto the right track with this story.
Q: Do the numbers indicate that if you spend two times as much you get four times the value?
A: We haven't done that work yet for one very simple reason. The problem is that there are no great data sources about IT spending at the level of the company. We got very fortunate because the U.S. government tracks, very consistently, IT stocks and flows at the industry level, year after year. Here is my hypothesis walking in to that analysis. Let's assume a magical data set existed and we could do exactly what you suggest. My intuition tells me that we would find a positive correlation between information technology spending and a business goodness defined as better revenue share, better market share, all these things that we might care about. A weak positive correlation. The reason I say it's weak though is that every time I look at any IT data, the thing that strikes me is never the mean, but it's the variance around that mean. So companies can spend equal amounts of money and get hugely different returns to that kind of spending.
Q: What's your top recommendation or your top couple of recommendations as you speak to folks - just be smarter about where you're spending all the money?
A: My first recommendation is if your senior leadership of the company is not focused on the possibilities that technology brings to them, they had better get there. Because I have yet to see a company where the top leaders didn't kind of, at least appreciate technology if not embrace it. So the first thing I say is, "Look, get on board." The second one is to start thinking about enterprise-level initiatives that you can do with technology. Every time someone goes to shop at Walmart.com, they see user reviews now. It doesn't only happen five percent of the time. I try to make that analogy to the other processes the company executes and say, "Look, if you believe that technology can bring you better processes and business models, one of the things that you need to do is apply those as widely as is appropriate inside the company." When I look at most IT infrastructures and the processes that they support, they're incredibly fragmented -- it's business unit by business unit, the geographies are all different, and there's this complete patchwork of technologies and then business processes on top of that. I say, look, if you have a better mousetrap, it's incumbent on you to spread that around as widely as possible inside the firm. As far as we can tell, technology does not stifle innovation, and does not reduce your scope for doing interesting things. It actually increases your scope for doing interesting things.