Listen to the podcast here, and read the write up here.
_____________________
Eric Lofgren: I’m here with Anne Marie Knott she’s a professor teaching strategy and entrepreneurship at Washington university at St. Louis. And she’s also a former researcher at Hughes, and we’re here today to talk about her book, how innovation really works. And Marie, thanks for joining me on acquisition talk,
Anne Marie Knott: [00:00:37] thank you so much for having me, Eric.
Eric Lofgren: [00:00:38] I want to start with your experience doing research and development at Hughes, which of course used to be a big defense contractor before they merged several years ago, several decades ago now. So can you just talk a little bit about Hughes and your career there? And then how was R and D managed and how did that change over time at Hughes?
Anne Marie Knott: [00:00:57] So Hughes was fabulous. I was in the missile systems group. We were a big company. So each of the different groups were in different buildings around the Los Angeles area. And, you got to work with really smart people. You got to work on really interesting problems. And one way to capture how the company viewed R and D was that when you wanted to propose a project, it was just a one-page sheet. It wanted you to give the title for the project. What you were hoping to accomplish, how many people you needed and how long you thought it would take. And that was it.
And what would happen is all of these would get submitted up to ultimately what we used to call the college of Cardinals. These were all PhD, scientists and engineers who understood the technical merits of most of these projects or collectively did. And we would all wait for the white smoke to appear.
And, my case did almost always appeared.
Eric Lofgren: [00:01:49] What was interesting, actually, you brought in some of that experience from Hughes into your book, you had a couple vignettes about different technology programs and how they interacted and what you learned from them. So was there a one or two that you might want to just bring up for our audience here?
Anne Marie Knott: [00:02:04] The one that I remember talking about, I never really talked about my own projects, I don’t think. But by way of backdrop there’s a stylized fact that 125 projects of those only that one becomes a commercial success. So the most likely outcome of any project is that it’s going to fail or be abandoned.
And anything about a company like Hughes was that we were so large that if something failed for a particular application, there was often another application somewhere in the company. So one technology was ion beam propulsion, so ion beam technology. And it turns out that military satellites that was its intended use only have a five-year life.
And at a five-year life, the technology doesn’t pay for itself. So it was abandoned for that application. But they discovered that they could use it for implantation of layers on semiconductors in another one of the groups.
Eric Lofgren: [00:02:56] Yeah. I think the, you brought up the success curve, just like one in thousands of projects actually succeeds.
And you talked a little bit about that in your book. And I thought maybe we would pivot to talk about that because it seems like something that occurs across like the markets, we VCs talk about the same thing. You were pointing out that, drugs through FDA approval. It’s one in several thousand.
So they all have this success curve. So what does that mean? If you’re a fund allocator what would you do having that knowledge or what are the strategies available to you?
Anne Marie Knott: [00:03:29] With the, so there’s a few, it’s very similar to BC’s. The first is that you carry a portfolio, so companies recognize only one-in- 125 is successful.
Then you want to carry maybe 125. The second thing is that you want to be very good at pulling a plug. Because your ability to fund projects is a function of how much you spend on each of those projects. Companies that are very good managing a success for curve exit early, most of my research deals with exit.
I think those are the main ones.
Eric Lofgren: [00:04:02] Yeah. I think, when I think about, okay, I need a thousand trials to get one major success. And of course you get the asymmetric upsides from that. But it seems like one of the issues is, and what we’ve been hearing isn’t that — like the scarcity isn’t necessarily dollars that are going after these types of projects or R and D, but it’s actually like a scarcity of people or talent.
So if I do a thousand things are there a thousand people that are, or am I going to be getting the same people coming back through this pipeline does that kind of, does the scarcity, idea affect where you’re going to be re allocating those resources?
Anne Marie Knott: [00:04:41] This is a common view, is that there’s a scarcity of idea and people, it turns out that companies are actually over-investing in R and D.
So they’ve just got the people in the wrong places.
Eric Lofgren: [00:04:52] And that’s a very interesting point of view. And you brought up a lot of instances where, It was actually like a lot of these new startups. They didn’t come out of the blue. Right. They were actually learning from the large companies where they were those ideas and take the discarded ideas and then scale them.
So can you just talk a little bit about that relationship?
Anne Marie Knott: [00:05:12] Sure. Everybody thinks that the solution to growth is to sponsor more entrepreneurship. W from the book and others who’ve read, it will know that my belief is that the growth problem is directly related to the fact that there’s been a 65% decline in company’s R and D productivity.
And I’m not the only one to have documented that was originally documented by Chad Jones, back in 1995 at the economy level. So the thing that people are pointing to when they say we need to sponsor more entrepreneurship, is that there’s been a fairly dramatic decline in the founding rate of firms.
And I did a study where I actually showed that the founding rate of firms follows the decline in R and D productivity with a lag. I can do that across markets, or I could do that just in the U S market. But the reason is, that most ideas from new ventures come from somebody, the founder who’s been working inside a firm.
And he sees a project that the company wants to abandon and remember about 124 of 125 are abandoned. They fall in love with the project. They say this project doesn’t have a big enough market to keep the company happy, it’s got a big enough market to keep me happy and then they leave.
So if companies are doing a poor job in generating projects they’re less productive, then they’re also going to be less productive in producing this byproduct, which is from founders.
Eric Lofgren: [00:06:30] We definitely want to get back into your measure of research and development productivity the RQ, but I want to start with something like the empirics and the big picture questions, because especially in the department of defense, we’ve been hearing a lot of folks talking about how like federal spending on research and development and especially in the military itself has just been, declining over time since the fifties and the sixties, and particularly for the early stages in research what’s your reaction to those claims?
Anne Marie Knott: [00:06:58] There has been a dramatic decline, but everybody thinks, as you said that it’s in research and there’s been absolutely no decline in research. And I don’t understand there’s whole books that are coming out using that as the motivating. The motivating argument for funding, entrepreneurship, et cetera, et cetera, and funding universities, all of the decline.
Going back to the 1950s, all of the decline is in development. And that’s what goes to firms. I don’t understand why this hasn’t been recognized. It’s well-documented in the national science foundation data.
Eric Lofgren: [00:07:33] a lot of people have been asking for the government, for example, to target a higher number, like three, even up to 5% for science and technology.
Do you think that those kinds of quotas are, something that should be done or do you think that it’s overblown and we’re actually at a pretty good place right now in terms of that distribution?
Anne Marie Knott: [00:07:52] Right now, we’re not at a good place because there’s not enough development, but in terms of the percentage of the GDP that the federal government spends on research that stayed absolutely the same.
The problem with having something stay absolutely the same as you’re not running any experiments. You don’t know whether you raise it or lower, it’s going to affect GDP. They only thing that we do know because we’ve changed it is then the level of development. And that’s declined and GDP growth has declined.
Eric Lofgren: [00:08:19] Yeah. So one thing I think we all and get into these problems when we talk about R and D, cause we say okay, it was this many dollars or this many dollars as a percentage of GDP. but we don’t really have a way to connect that to outcomes or the effectiveness of that R and D.
So that then becomes this big kind of social science and maybe like historical kind of project. But you have a measure that is pretty well-defined in terms of data and what it means, and you call it the the RQ. So can you just describe, what is the RQ and how do you estimate it?
So
Anne Marie Knott: [00:08:52] our Q stands for research question it’s intendedly the company equivalent of individual IQ. So I think of it as how smart companies are. So just as smart or high IQ individuals,
Should I, let me, I don’t want to spend too much time going
Eric Lofgren: [00:09:06] back through economics
Let’s presume that the audience already gets a lot of it, so you can go straight in.
Anne Marie Knott: [00:09:10] Good. So the production function is, the most common way to measure firm productivity in the comprehensive way. So let me back up.
Eric Lofgren: [00:09:19] Yeah. Let’s start with if I’m looking at the dataset what am I estimating? And then what are the independent variables that I’m using to get me there?
Okay.
Anne Marie Knott: [00:09:28] Yes. So RQ is estimated from the company production function. A company production function has. I’ll put on the left hand side and it has all of the inputs on the right hand side. So the output that I look at is total revenues because you can’t look at widgets when companies are multi-product and most companies are multi-product.
And then on the right hand side, I have the typical inputs that you would see, which would be capital and labor, but then I also have the intangible assets. Are they intangible in inputs? So R and D advertising and something called spillovers, which is knowledge input that you get for free. It’s the knowledge that’s being generated by your rivals that you can take advantage of and the reason it’s important and to include that is that small companies rely more on that.
Those spillovers than they do on their art, their own R and D. And if you don’t take that into account, you’ll overestimate the productivity of small firms. So what RQ is exactly is it’s the elasticity of R and D in generating revenue. So the output elasticity of R and D
Eric Lofgren: [00:10:32] it’s a very economicsy term, but it really means the, like when you change R and D by let’s just say a percent, then that has an effect on the next year’s revenues.
And you’re trying to measure that relationship. Some firms, a small change in R and D can have a big change in revenue, which means they’re they have a high RQ or they’re very productive and others might be less, right?
Anne Marie Knott: [00:10:57] Elasticity, the percent did change and output for a 1% change in that input.
For a one. So the average RQ is right around 0.1. So for a 1% increase in R and D you get a point, 1% increase in revenue.
Eric Lofgren: [00:11:16] Okay.
And that’s actually a lot more than it sounds because R and D tends to be quite small relative to revenue,
that’s right. Yep. That’s right. Yep.
So what kinds of interpretations, or what kinds of conclusions have you been able to draw from the RQ and then what are the best criticisms or limitations of it?
Anne Marie Knott: [00:11:34] So the main thing I’ve been doing with our queue is. Trying to understand why we’ve had this 65% decline. So I’ve got this hammer and I’m trying to find all the nails I can possibly find to try to sort that out. Unfortunately there’s not that much data on companies, R and D. Well, what they actually do with respect to their R and D.
But I’ve been sorting through the things. The first thing that I was able to do was get access to the national science foundation survey of industrial R and D. So I linked RQ up with everything in there that I can find the big finding. Do you want my big finding from that? the output elasticity of outsourced R and D is zero.
Whereas on average for internal R and D a 1% increase in R and D gets you one point 1% increase in revenue. For outsourced R and D it gets you no increase in revenue.
Eric Lofgren: [00:12:27] What you found there. I thought it was really powerful. The outsource R and D has an RQ of zero. So that means like literally when I increase my budget, that I’m outsourcing for research and development as a company, I can expect zero additional revenues from that outsourcing. So I, to,
on average, let me qualify it.
Of course. So there will be a variation in some will. Some will gains others will not, but it’s much less in terms of the expected additional revenues from actually doing it and building that capability to institutional capability in house.
And of course, as a defense acquisition guy, I look at the government. And the defense primes as well. And it seems almost all of their efforts, especially for the government, almost all of it after the very early stages is pretty much outsourced. do you think this finding has any kind of bearing on, how government contracting is done?
Like should the government actually in-source and do more like arsenal slash Bureau research, like they used to, or do you think that’s a separate situation and these results don’t really may not apply there.
Anne Marie Knott: [00:13:33] Okay. So what I think is going on this gets back to this idea of the 124 out of 125 projects go nowhere.
So one of the values of being any project is that you get the benefit. Those 124 kind of fueling the one that, that that survives. So I think what you, what the implication is for acquisitions or for any kind of outsourcing, is that what you want is you want the body that’s going to benefit most from those spillovers to be conducting that activity.
So if I think about an aircraft does make sense to outsource the engine? And my guess is yes, that there’s so many spillovers in engine design and production that you, as the prime can benefit from that by continuing to subcontract for them. However, if there is a if the design of the engine has a really significant bearing on the rest of the aircraft, then you would want those activities to be
Eric Lofgren: [00:14:33] co located
Co-location that actually reminds me of a Meckling paper where he says production knowledge should be co located with decision rights. So that also, that would also mean if I want to keep the, keep it located or co located that should also mean like the contract structure should be a little bit different as well to allow the decision rights to flow to that.
Anne Marie Knott: [00:14:56] I haven’t thought about it. And it rights from a decision rights perspective. I’m in largely thinking about it from the standpoint of these spillovers, but that’s actually something we’re thinking
Eric Lofgren: [00:15:05] about.
Before actually, we’re jumping ahead a little bit because your book, you don’t just give this measure of RQ and then use it to show that there has been a decline in research productivity, but you actually have a bunch of recommendations for how you can boost the RQ.
And so one of them of course, was the in-house versus outsourcing. But you were knocking down a lot of innovation dogma. as I would call it as you went through. So another one was actually.
Company size. So you found that one of the big misconceptions was that small companies are the innovative ones and they’re much more innovative than big companies. So what does the evidence actually show there?
Anne Marie Knott: [00:15:47] The evidence shows that on average, again that RQ increases with company size. This is not a surprise to economists because for years there’s always been this puzzle that firms were spending more on R and D the bigger that they got, but they were being less productive.
So that makes it sound as though firms are irrational. And what was happening is — I believed is that in these econometrics, people were failing to take into account the spillover effect, right? they were counting how many patents per dollar of R&D, a small firm was getting.
As opposed to something closer to how much value are they creating from their R and D. That gives me an opportunity to say the distinction between RQ and these other measures or count patent counts is that RQ is capturing how much value you create from R and D. Not how many things you, how many things you create.
So large firms are more productive because when they create something. Or when they invent something, they can diffuse it to a broader market. And that’s precisely, I think it was back to Canterbury goes even back for them to that. But the expectation is that larger firms will have greater incentives to invest in R and D precisely because they can because they can get things out to bigger markets.
Eric Lofgren: [00:17:03] By the way have you read some of Ken arrow’s old ran papers on defense acquisition and R and D?
No,
they’re actually, they’re pretty great.
But I wanna meet him.
Yeah he’s. He died a few years ago, right?
Yeah, that’s right.
Yeah. Yeah. It was like five years ago. He had a long life and so did our mountain and those two were just like, really great together at Rand.
Back to this kind of idea of large firms and actually they have more distribution channels. So like the ideas have, a way to jump off from there. I’m going to give you a couple of pushbacks and I want you to push back on my pushbacks. Large firms may have that kind of market power economies of scale, that kind of optionality, that is engendered in just their size and the different markets that they’re touching.
But startups can actually take advantage of new, potentially lower cost distribution channels and is through those new channels that you get, those kinds of outsized returns as opposed to leveraging what exists. So what do you think about that?
Anne Marie Knott: [00:18:03] So those are what people call disruption, right?
Where we talk about entry barriers is way for for firms to protect them. And Manette monopoly profits. And the disruption is when so one of the things that happens when firms are monopolists is that they tend to do less. They tend to do less innovation because they have already captured the monopoly market.
And there’s no incentive to go beyond that market. So yes there were a number of entry barriers. In a member of industries and one of the really fun things was this is my, the entrepreneurship side of the thing that the work that I do is that they get these, they’ve formed, these end runs around the entry barriers.
And a lot of them has, or point out are distribution channel related. So think about blockbuster getting disrupted by Netflix, right? When they first started doing the mail things in all the things that were being delivered by the internet for example, now,
Eric Lofgren: [00:18:51] Yeah. well, I guess one of the issues is that a large company also provides optionality in terms of where that, I guess R and D effort can go, but a startup can also just pivot itself.
Pivot, it’s go to market strategy as well. Is, does that make any sense or I think maybe the small and the large have the same optionality potentially depending on how nimble the small is.
Anne Marie Knott: [00:19:14] Yeah. I don’t necessarily agree that one has more optionality than the other except to the extent that small firms don’t have any resources.
So they, they are less inertial, they’re pivoting all the time until they figure out what the right way to go is. The big distinction between small firms and large firms is that large firms are pooling all these bets.
Whereas small firms are just single bets, which is why the venture capitalist makes money as opposed to the entrepreneurs. The most likely outcome from entrepreneurship or founding a firm is that you’re going to fail, just like the, for pro at the project level, 124 of them, one 25 projects fail.
Yeah,
Eric Lofgren: [00:19:50] that’s a, that’s another difference there in incentives between the large and the small. So the large can create these portfolios, as you said, and then they actually they take some of the risk reward away from the researchers, who aren’t participating necessarily to the same extent, whereas like a startup, they have much more participation in the upside, but they also have a much greater risk.
Of the downside. So do you see that kind of incentive difference or, is potentially the protection from risk, are researchers a little bit more risk averse? And so large companies actually have a competitive advantage with that, with respect to that,
Anne Marie Knott: [00:20:24] the resource should be in principle and the researcher should be more risk averse inside a company.
Because they. They’re protected by the promo. Okay. That’s funny. I thought you were saying the
opposite, the firms in advantage because researchers want to be risk averse and they want, so maybe they get some of the benefit from a selection bias or something like that.
That’s a neat
idea.
I met. I remember, it’s been great. Not even just need to think about it and thought about it.
Eric Lofgren: [00:20:46] I guess when I think about some of this, the issues with you have a large firm, they’re doing a lot of research. And then, some of the people in the firm will actually leave and they’ll take these ideas and then they’ll build them out themselves.
It seems to me that’s not necessarily a point for the big firms, because they are either discarding them. Or, the people inside the firm actually recognized an error that the large firm was making. And if they didn’t have the option right? you would have been never having these innovations in the first place.
So it’s hard to say, are they like symbiotic? I
Anne Marie Knott: [00:21:19] love the fact that things get spun off, right? Because the company isn’t making a mistake, the company just says, this is not, this project is not big enough. this is not big enough to we move on. Yeah, it would be one word that they might use.
This project is not big enough product to make an investment because we need projects that can take advantage of our whole market, not just some niche in the markets. Things get abandoned for market reasons and they get abandoned for technology reasons. And my guess is that I don’t have any data.
I haven’t seen data on this, but my guess is the bulk of them are getting abandoned for marketing reasons.
Eric Lofgren: [00:21:53] Yeah, I think he did actually a really good job in the book, actually talking about these interactions and how useful they were. And it seems like the disruption in the DOD is even harder because.
There really isn’t that niche. Or as you said, like some, sometimes these things don’t look big enough for the firm or they’re not moving the needle, so they drop them. And in the DOD, it seems like it’s the same. when I think of Clayton Christianson and these, the disruptive innovation starts out actually worse and it’s in these little niche areas, but it can grow.
The DOD doesn’t allow for that because it’s Where are the niches. And usually the next system has to be better than the previous system on every front, or it just doesn’t get funded. So there’s no like real, flowering for these guys. Have you thought about that in terms of, I guess just the government space versus the ability for the private sector to actually harness that.
Anne Marie Knott: [00:22:40] Remember. Okay. So the big problem and the DOD right now is that we’ve doing almost no development. Things are getting cut up at the research stage.
Eric Lofgren: [00:22:48] Can you expand on that? What do you mean? The DOD is not doing enough development.
Anne Marie Knott: [00:22:51] I’m going to want to pick on the DOD, but the DOD is one of the main the main areas of defense of the federal government spending on R and D.
Right? All of the decline in development, all of the decline in federal spending on R and D is in development. So DOD needs to be a big part of that, right? Yeah. Yeah. That was my point. But what came to mind was, and I don’t know whether you would call it this, it isn’t disruptive, there are small and very cool technologies that the DOD has incubated.
And when I say incubator, I don’t mean it in. In a, that institutional sense, think about I robot, that only exists because they were using the robots the, that language inside venture capital for 10 years, they had no customer. And then the government ended up using the robots for defusing IED.
Eric Lofgren: [00:23:37] Yeah, definitely. I hear that. There’s definitely a counter examples. I almost feel like with I robot it’s like a lot of robotics might’ve actually started in like the DARPAs of the world and then they boomerang out to industry. And then when it matures in industry, now it’s ready to come back into the department of defense itself.
So you were saying that the evidence shows using the RQ, that large firms are actually more productive on average with an R and D dollar vendor than a small firm, but then you kind of balance that out. with this observation that I first got from Ben Horwitz, but I think you actually beat him to it, which is the fact that, it’s not about big versus small.
Only it’s also young versus old firms, or like when a firm is actually run by a founder someone that has that experience of taking risks. So what’s that kind of relationship there between size, but then also with age.
Anne Marie Knott: [00:24:30] So the size we already discussed the age. So as you were alluding to, and as you might suspect, the older you get, there’s a small.
Decline in RQ associated with age. And so we think of that as ossifying. . You said besides there was small, there was age, there was. Large, small, there was one other piece that you had. Run by
Eric Lofgren: [00:24:51] founders or not the
Anne Marie Knott: [00:24:53] founders. Yes. Thank you.
So this isn’t formalized because there just aren’t enough founders to be able to do anything statistically significant. But here. I ranked the top 50 firms on this measure and there’s a disproportionate share of founders in that set. Similarly when founders leave a company, I found in the handful of places where I’ve been able to see that they’ve left.
I see that the RQ actually declines. So there does seem to be a founder effect. And people would like to understand that. And I can’t answer the conclusively. So my, but my intuition is that founders have vision, right? They know what they’re doing, the art, they know what they want to accomplish.
They know why they need R and D to accomplish that. And everything is cohesive. Everything makes a lot of sense. And after that, when you lose the founder, everybody else is trying to second guess what they should be doing. And I think they’re too responsive to investors. I think investors are really important, if you don’t have the if you don’t have the.
Kind of the vision. And I hate that word, the vision that the founder does, you’re vulnerable to other people’s suggesting, Oh, maybe you should be doing this. Maybe you should be doing that. And the other thing, the other advantage of founders have is that the investors will leave them alone. The investors are betting on them as much as they’re betting on the company. I mean, Just look at Tesla right now. That’s not a bet on the car. That’s a bet on Elon Musk.
Eric Lofgren: [00:26:17] And I think like when I look at the department of defense, of course, a lot of the big players there, those are very old companies.
None of them have their founder still. Because they’re much older than that. But then you see these entrants, like the space X with Elon Musk, palentier with Peter Thiel and then. Anduril with Palmer lucky, and they’re doing what you said. They have their own vision.
They’re just going for it. And they have, I don’t want to say less regard for investors, th the way that they contract with the department of defense, which is a customer, is also they have more decision rights and they’re doing more with internal funds and shouldn’t the department of defense just be like loving these proven founders that are actually coming in into their sector.
Anne Marie Knott: [00:26:59] I,
I’m not sure that’s an answerable question, Eric.
Eric Lofgren: [00:27:03] What’s your opinion.
Anne Marie Knott: [00:27:05] I’m not even sure I can answer with an opinion, but are these guys exciting? Yes. Should they embrace them just for the sake of embracing them? I’m not so sure. They have to be fulfilling some kind of need, but so new technologies come from one of two places.
There’s one is the technology push. You’ve probably heard this. And the other is demand pull. The scenes since DOD is. Ultimately a customer that they should be playing the role of demand pull. But thinking about what technologies are going to be able to satisfy the demands that they might see.
So it just reminds me of the proposals that we read right. At Hughes. we saw opportunities, loosely defined, in terms of where our technologies might go. But we weren’t doing anything like market analysis and I can get to go back to tell my story.
So when general motors came in and acquired us, I remember the first time we went through the proposal cycle again, I submitted my normal proposal, came back and they said, we want to know the ROI for this project. I said, what are you talking about? this project is technologies, 10 to 20 years away from commercialization.
There’s no meaningful ROI. I could give you for this technology. And they said you need to have one. And so I came up and had my MBA at that point. And so I came up with an estimate, it was, I thought of all the things you could potentially do, how much they would actually be worth. And I said, it’s 204 7% What do you do with a number like that?
Eric Lofgren: [00:28:27] Yeah. Can you talk about, like you said, he was acquired by general motors, but general electric? You had a good story in there about how they used to do a lot of R and D. And then like financialization at the top and it that fixation on, quarterly earnings and like accounting figures.
Yeah. I actually drove them to reduce that and that like really killed their RQ. So can you talk about the GE story there? Yeah.
Anne Marie Knott: [00:28:52] And then what’s nice about the GE story is I think that this is the story of, for all firms generally. It’s just nice and compact, but Jack Welsh was considered to be the hero of general electric.
And his strategy was to exploit all of the assets inside general electric and the stock had a huge. Run up people who are investing with GE at the time did really well. The problem is he dis he picked monopoly markets. He wanted to be one or two in each of those markets. And what you do when you’re in those markets is you insulate yourself from competition, right?
And that tends to suppress that definitely suppresses innovation for the reasons I mentioned earlier. Yes, there was disinvestment in R and D and a decline in RQ, not surprisingly. And by the time that Jeffrey Immelt took over the company, there was nothing left to exploit. And, he just look at what you can see what happened to the stock after, after
Eric Lofgren: [00:29:53] that.
So is the moral of the story you really have to see the quality or does the RQ, like you had been able to see what was going on there with the RQ or does it really tell you that I need to know like the specific qualities of investments and management actually going on in the firm to understand.
Anne Marie Knott: [00:30:13] You just need to read in the case of GE it was, companies are, you’re not there to milk all of the investments that were made previously, right? There’s they’re supposed to be making new investments that that you can continue to exploit in the future. Yeah, gradually over time, improving the set of assets that that you can take advantage of.
So you want to continue, you want to continue doing R and D if you’re a technology-based company, the advantage of RQ is, so this may have been deliberate. It looks to me like it was a deliberate, this investment in R and D in which case, RQ isn’t particularly helpful. But if you’re worried that your R.
You know that your R and D is slipping, which is different than cutting. If the productivity of R and D is slipping RQ, we’ll give you an early warning signal that, you might change the way that you’re conducting your R and
Eric Lofgren: [00:31:00] D. one of my economics, professors actually said this, right?
Like when you go, when you IPO and you get to this big company status, the point of your enterprise is really to milk what you’ve already done. And that’s like your obligation to the stockholders, to an extent. and some of the, I think empirical data.
Seems to show that right? Like the older firms have less RQ. So it almost seems like a self fulfilling prophecy that you should be almost like reducing your R and D.
Anne Marie Knott: [00:31:27] No, no, no higher RQ means that you aren’t very good at exploiting your R and D past R and D. Yeah. Yes. Continual R and D right?
It’s you know, whatever in whatever R and D I’m spending now is going to be enhancing. My, is going to be enhancing my revenues in the future. So th the issue is to continue to invest in R and D so you continue to have more things that you can exploit, make your market’s even bigger.
It’s not, you don’t want to, you never want to stop. The problem is sometimes outside forces cause a firm to have to to do things and ultimately hurt their RQ.
Eric Lofgren: [00:32:01] I think. One of the issues is that a lot of firms are a lot of people look at these firms and they say because of their size of their bureaucracy, they’re actually not going to be doing R and D very well or as well.
So they should, create their own VC funds. But I think you, talked a little bit in your book about the idea of like skunkworks and building out skunkworks as a good way for these large companies to actually harness innovation. So can you talk about, what does it characteristic of a skunk works, factory in an affirm.
And then how do you contrast that with the idea of a central lab?
Anne Marie Knott: [00:32:33] skunkworks historically our project-based. Okay. You want to develop a particular project. You want to insulate it from various problems that you think that the firm ordinarily has in developing something.
And so you put it in a separate building, put all the, put all the resources necessary inside that building. And there’s lots of neat examples of that being successful. The distinction between that and a central lab is that a central a lab is not project-based. The goal of a central lab is to journal generate is to be generative.
So rather than exploiting one particular idea, It’s doing is it’s generating technologies that ought to be useful throughout all of the company’s business units. Okay. They, they may be in a separate building, but their power, that isn’t necessarily their power now. But I want to get back to the skunkworks idea because my favorite story that’s related to this is interesting because it combines.
The best, the advantages of both small firms and large firms. Okay. And I th I think skunkworks, when people think of skunkworks, what they’re thinking is, this is an opportunity to operate like a small firm inside a big firm. The story that I love is Xerox technology ventures. And this was set up after the company was embarrassed from a book in the eighties called bundling the future and what that book documented among other things was the fact that the market cap of companies that had spun off from Xerox was three times the market cap of Xerox itself.
And that made it sound like the company. The company had made big mistakes and they want to recover from them. As I said earlier, that I don’t view those as mistakes. I view those as a neat by-product that you generate to the world as a, as a byproduct of doing good R and D.
But what they did in XTV, that was very cool was they said, let’s take. They set up a venture fund. David $30 million had an old, the normal structure of a VC, a ten-year life. They actually brought in VCs to run the fund, but whereas normally corporate VCs are investing in.
Side technology is what was happening. XTV is they were in, they were investing in the things that Xerox was going to abandon. they were recycling farm for Xerox technologies. And what was quite cool was that First of all the outcomes. I think the return on the $30 million, I think they ended up making 250 to 300, 300 million.
So almost a 10 X return, which, outperforms any VC, anybody knows. But what was more fun is that Xerox had the right first refusal for our projects once they came inside XTV and there was one where Xerox had originally abandoned it because it was going to cost. I want to say 25 million and take 36 months.
And once it was inside XTV they completed it for. I think it was 4 million in 18 months. So what’s nice about that particular example in all the examples of the XTV, if we saw them in greater detail is that they were in, they had all the advantages of Xerox. So they had advantages to Xerox as manufacturing, to Xerox’s Salesforce, to Xerox’s suppliers, which would never have you would ordinarily not have it.
If you were an independent firm. But you were also, you were also set aside and you had these high powered incentives. And even the employees were compensated the way that founders would be or similarly, that way founders would be compensated. And it worked. It was abandoned.
Eric Lofgren: [00:36:03] It was abandoned.
Anne Marie Knott: [00:36:04] Oh yeah. Yeah. That’s how I think when we teach the case, that’s the big question. Why would the company not do that? The only thing that we can come up with is that those guys, the people on XTV were making more than the people inside Xerox and that just created a, problems.
Eric Lofgren: [00:36:19] I feel like does that model work in the department of defense itself? they’re trying to hire people that have higher salaries than the people that manage them, because there are these technical skills So the workforce issue in the department of defense and that disparity of pay, potentially, it seems to also be a kind of, sticking point. I guess two questions.
One. what would a skunkworks look like in the department of defense and to what kinds of incentives might make that reality? If it’s a good thing, if we want that.
Anne Marie Knott: [00:36:48] so I guess I’m hooked on the idea of, these discards, what is it that the defense department has to be, needs to be recycled more so than the idea of creating this court so that it was a skunkworks with a particular purpose. So yeah, I think it would be great for the DOD to examine what technologies it had abandoned and why it abandoned them. And then see if you could bring in some kind of VC or incubator or some something similar to see if there was anything that they could do with those things.
The Problem in the DOD, if they do it thinking of the labs rather than the companies doing it is, there’s not these other things to exploit, like the sales like the Salesforce, right? Like the marketing, the distribution system, like the manufacturing systems.
That’s what made you know that this complimenter. the best of both worlds, where you have all the large firm advantages and all the small firm advantages is what made the X TVs example
Eric Lofgren: [00:37:42] work. I see. Yeah. So the government could do some technology transfer with other firms that could, that would be willing to take those on.
That might be different. have you looked into the defense firms and thought about what are their incentives to go and do something like this, or are they fixated on just government contract dollars for R and D?
Anne Marie Knott: [00:38:05] So before even thinking about defense firms, I don’t know any other firm who has done that, right?
Eric Lofgren: [00:38:11] You hear about defense firms with their little D VC funds, like BA has a pretty sizable one. Lockheed has a smaller one, but they’re not, I think they’re doing what you’re saying.
Anne Marie Knott: [00:38:19] Necessarily the opposite. They’re looking externally to keep their fingers on the pulse of what’s going on externally.
With the intent of bringing it in-house possibly. So that’s the conventional definition of a VC. So their investments more so than their incubators. Yeah, the Valley of death problem is, that there seems to be all of these things, these ideas that nobody wants, and everybody thinks it’s just so that we’re not selling them on up the problem isn’t that we’re not selling them enough, it’s that, they inherently don’t have enough merit.
Eric Lofgren: [00:38:52] So your view of the DOD is Valley of death. Is that, not that there’s a transition problem per se, in that the, the right programs aren’t getting scaled. You’re saying a lot of those things just don’t have enough merit to actually scale.
Anne Marie Knott: [00:39:05] I think there has to be demonstrated demand.
If companies see that there’s demand for something, they’ll make the necessary investments to carry them forward. I believe short of things that they don’t recognize.
Eric Lofgren: [00:39:17] So when the DOD says. Oh man, we really love AI. Let’s just say that for example, but then it takes them five years to get that fund allocation up to close to a billion dollars.
Like companies look at that and they say that’s not, they’re not really actually interested in what they say they’re interested in.
Anne Marie Knott: [00:39:36] Yeah. Money talks. And again, if we have more development dollars, that’s, that’s the big Valley of death. And the DOD, I think really is that, there’s not development dollars, then the companies have to fund the development on themselves.
And if there’s also no hope that there’s a market out there, there’s not sufficient certainty that there’s a market out there, then they don’t see a return for the investments that they themselves make. So the government, I think they might need to be doing more, both in the development as well as procurement
Eric Lofgren: [00:40:05] in development procurement.
So when you say there’s not enough development dollars, are you saying The government just needs like a higher top line or is it like lot of those dollars are fenced off for an F 35 or something like that?
Anne Marie Knott: [00:40:18] No. And yes, I think you’ve seen that the curve on defense then the defense dollars.
So If you take the, I could show it to you if you want it, but we’re on a podcast so that doesn’t do your audience. But the the funding of research has been at a constant level of GDP for ever. The funding for development went from. I want to say three times that to now being equal to with the research
Eric Lofgren: [00:40:45] funding.
Yeah, I looked at I looked up this a little bit where I just was like, okay, I’ll just go for the DOD budget figures and look at how much dollars were in RDT and E and procurement versus O and M. And it was something like, between 50 and 60 was in the acquisition side and 60 and 40% was O and M and now it’s basically flipped.
So it looks like a lot of the sustainment kind of work is crowding out the research and development. And I guess one of the problems that we’re facing is. For the DOD, you have these existing missions and these existing high sustainment costs that you have to pay. But if you want to get to where, you can flip that where you’re taking advantage of information technology, you’re doing new things that actually bring down production and sustainment costs using new technologies.
You need upfront investment to get there, but if the investment’s already crowded out, Yeah, you’re stuck between a rock and a hard place. So is the idea almost like you just have to bite the bullet in the near term in terms of capabilities, like for sustainment and just reallocate those funds to development, well,
Anne Marie Knott: [00:41:51] yeah, or they should get more funds, right?
Eric Lofgren: [00:41:53] Or just overall,
Anne Marie Knott: [00:41:55] Yeah. Then the big money is in development.
If you think about, if you look at any given project, the amount of money you spend on the research phase, relative to what you spend on the development stages. Trivial, we don’t need, we don’t need more ideas. We don’t need more technology. We need people to actually either want to buy that technology at the end, which gives firms the incentives to make their own development expenditures, or we need to fund more development so that the amount of, the market size that we need isn’t as big as it’s currently to justify investment.
Eric Lofgren: [00:42:28] Here’s one of the things might result from that is I often see like the DOD always wanting to like leapfrog. So it’s if you have all this, like great research, but not enough money for development, it’s it’s
like, okay my development of that idea is going to take another 10 years.
Let’s just supplant that and do the next, leapfrog. And you’re always trying to leap frog, but you’re never quite, fully developing the system that you need.
Anne Marie Knott: [00:42:49] Yeah. Leap frogging is a nice idea, but usually what happens is it. One entity leapfrogging, another entity. I’m not sure you can actually leap frog yourself, but
Eric Lofgren: [00:42:58] that’s a very good point. I never thought about it that way, so, going back to this idea of how companies do the research and development, another thing. You brought up, was this idea that over the past decades it has been, it was pretty fashionable to think about, Oh let’s decentralize our research and development into the operating divisions.
And the people who are closer to the operations will be able to direct and pull along the research and development much better verse this idea that you were describing of a central lab. So can you talk about. What’s the pros and cons of each and what your thinking is there.
Anne Marie Knott: [00:43:33] the idea of decentralizing is to be closer to the customer. And the thought was that the central labs were too ivory tower. And, therefore we ought to make that and more responsive to the visions. So first of all, one problem with that is that Christianson says that’s exactly what will lead to disruption, right?
So too much attention to your current customers makes you vulnerable to technologies that are suitable to a different set of customers that are ultimately, it will be on this trajectory where they’ll overcome, the technological progress you’re making in your own, in your. In your current technology with that set of customers.
Chris would recognize that de-centralization might be a bad idea. My favorite thing is a quote from, I think, from Steve jobs, which is that sometimes customers don’t know what. They want until you show that to them. And related to that as the fact that if the customer is telling you what they want, and maybe that they’ve seen it from somebody else, in which case, when you develop it, it’s not going to be worth very much because it’s going to be competed in a way.
So the, the research on centralization suggests that companies that have centralized R and D in it. And when I say centralized R and D, it doesn’t necessarily mean that it’s in a lab. Like we were talking about with regard to stunk works. It just means that the corporation is deciding where the, how to place the bets.
They’re doing all the allocation. But the research suggests that firms that are centralized have 40 to 65% higher or twos than those that are decentral.
Eric Lofgren: [00:45:01] Can you had a nice whole story about Proctor and gamble with respect to this. Did you, can you just expand on that to give an example?
Anne Marie Knott: [00:45:08] Oh, sure. So they alternate decentralized, but one of the examples of the merits of the centralized R and D is that white strips is probably their last big product. And white strips required technologies from all the different operating divisions.
So it so bleaching would have been, the household cleaning thing. The adhesives was from another group. I can’t remember the other the other group, but definitely an effort that benefited multiple groups and came, took technologies from multiple routes.
Eric Lofgren: [00:45:42] let’s focus on this idea of, when you decentralize and you’re closer to the operating units, you’re closer to the customer. especially in department of defence, it seems like a lot of the labs they tend to talk about there’s this constant debate that’s been going on for a long time.
Should we be taking more requirements from the customer and kind of defining, our programs around that? Or should we be like operating with more discretionary funding? And, it is, we that kind of are able to determine. Our own path because we are like the science and technology people, we should be presenting options to the operator folks rather than taking their preconceptions and just trying to make it work.
Anne Marie Knott: [00:46:21] Isn’t that what the role of IR&D
Eric Lofgren: [00:46:25] it’s supposed to be? The role of IR&D? Yeah. That’s all that’s for the companies. But it’s also with, within like the labs themselves, they went for their project and how they request funding. So I guess it’s at both levels. Should there be more discretionary funding or should there be more, requirements pull from the customer?
With respect to these types of labs,
Anne Marie Knott: [00:46:46] the obvious answer is going to be that you need to balance both, but what’s the record on this being too divorced from the customer. When I don’t have the record is the wrong word, but heard historically that was happening inside companies. Where, they were developing those technologies and they found that they couldn’t even get, they couldn’t get their divisions.
They were in. So one of the concerns is always that the government labs is generating a bunch of things and then, nobody’s picking them up. That’s that was the Valley of death that I was referring to earlier. I think you have a different one in mind. And all we need to do is we need to sell these things to the customer to manufacturers I was doing some research on central research labs.
And the firm itself told me that they, the central lab can’t get their own divisions to buy their technologies. So it’s a bigger problem. I think I’ve lost track of your question. So just remind me what you
Eric Lofgren: [00:47:40] add. Yeah, sure. I want to just bring out a quote here from your book talking about it was from a Hughes research lab guy and he says, quote, HRL.
Hughes research lab should be advising sectors or customers on what was important rather than the other way around that would make a world-class lab. HRL needed to be looking further ahead than the sectors of the customers. And maybe we should be doing exactly those things that no one in the sectors knows are needed.
So I guess that’s where I was coming from, because it feels like yes, in the department of defence, a lot of. The projects are directed from these requirements, from the customers, someone coming up with that idea and then directing it and then even for, Independent research and development in the companies.
A lot of times that stuff is just directed right back at the requirements. So it’s like I’m only going to do an IR&D if I see some kind of program or government sponsorship down there. it seems like there’s less of those kinds of big bets on their own thing with the IR&D. So I guess on the margin, in my view, I think we need more of.
What that HRL guy was talking about in terms of letting, letting those folks lead through discretionary funding? No.
Anne Marie Knott: [00:48:52] Yeah, no, it’s terrific. I love that quote. And the HRL was great. But the research that I’m doing now is interesting. We’re the preliminary results seem to indicate that these, that companies went to central labs.
Have, and this is different than centralization. Okay. Companies with central labs have lower RQ.
Eric Lofgren: [00:49:11] Oh, so it’s actually, you’re finding kind of the opposite now, what you were expecting
Anne Marie Knott: [00:49:17] eight on it before, right then. So I’m continually in search of data and we, there were manuals of the American industrial.
R and D manuals, going back to the twenties that you know, are now being digitized. We can create data sets from them and that’s what the preliminary results show, but it’s just preliminary. I could be wrong.
Eric Lofgren: [00:49:37] I’m glad. You’re letting the evidence, you’re letting your opinions follow the evidence, I believe in central labs and I will find it,
Anne Marie Knott: [00:49:43] yeah no, of course. No, we’re very disappointed by the results. It makes you dig a little deeper to try to see why are you getting the results, that’s where, your own passion or passion gets you to see. Stick with something longer than you would, if you didn’t have a bias the opposite
Eric Lofgren: [00:49:57] direction.
Yeah. And that was the big, criticism of Xerox park, that they were like on a Hill and then they just had that Valley of death problem and they couldn’t transition their own stuff. And it took someone outside to go do that. But so
Anne Marie Knott: [00:50:10] to the extent that it actually was feeling the product divisions, that would be great.
What I don’t, what we don’t know from Xerox is whether it was right.
Eric Lofgren: [00:50:17] Maybe one of the aspects of that is I had Steve blank on the podcast, but the lean methodology of just okay, it’s one thing to formulate this hypothesis and then just go build it in a vacuum. And then once you’re done, go sell it.
But it’s another thing to start out on that and then constantly be talking to. The customer or different types of customers, potential customers and iterating on feedback. So you’re you’re not doing this monolithic one effort and then try to sell it off but you’re like bringing in the customer into the process and then pivoting where needed.
I wonder if there’s, those
Anne Marie Knott: [00:50:49] are two different stages, right? If you’re a venture, you have to have a customer in mind, because the whole point of forming a venture is. So that, to make money, so you’re already outside of a lab. If you’re looking for a customer.
So these central labs do something different, they, what they’re trying to do in principle is just move the technological frontier with no. With no intended product or customer insight. Other than that, that we’re going to be investing in these technologies because we know we can see that, the world is going to evolve in such a way that we’re going to be able to use these technologies at some point.
Eric Lofgren: [00:51:25] So I want to touch on we’re running out of time, but I want to touch on one of the big factors that affects the RQ and it seems to effect it overall. Want you to talk about like, how does con competitive versus kind of monopoly conditions in a given market sector affect the RQ for that sector and for the firms across that sector?
Anne Marie Knott: [00:51:45] If it’s monopoly, there’s not many firms. Yeah, no. So the So what happens and I’ve alluded to this earlier in the conversation, when, what happens is the monopolists clearly have an economics background, but there’s a monopoly there’s monopoly output, right?
And what happens is if you’re a technologist and you’ve achieved the monopoly output, you have no incentive to continue to innovate. And we can see that in lots of instances. The neat thing is I believe if the market’s structured appropriately, that you can solve that problem with as little as two competitors, because what will happen? And there’s a formal model, not mine, but colleagues Dan Levinthal and rod Abner. Wrote this really nice model where, what they show is that what will happen. And as long as you’ve got to, as long as you’ve got two competitors, is that they will continually compete with each other on either product innovation or process innovation to capture customers.
And you continue to get innovation. You continue to get innovation. Cause then they compare that to a model where they haven’t monopolist and show where the monopolist stops innovating.
Eric Lofgren: [00:52:51] You have this nice quote from Richard Branson. Of course it made me think about the department of defense, but he said the customer has been ripped off or underserved where there’s confusion about whether the competition is complacent.
So does the DOD compliance barriers negate this as like a, place that companies want to enter, or should the DOD really be like something that like people would love to enter? Cause It seems like they’re being underserved. And they would want to switch.
Anne Marie Knott: [00:53:19] Okay. So monopsony is different than monopoly, right? So the problem with the problem with the DOD is that they’re the sole customer, right? So what you’ve got is the opposite. You’ve got a bunch of contractors that are all trying to satisfy. That are all trying to satisfy one customer. And then I don’t remember my economics seminar monopsony but typically you won’t get somebody to enter that market.
The basic theory is nobody should want to serve it, but it must just be so big that people are willing to do that. And of course, once they serve at once, then they can continue to serve that market because they’re so specialized. Yeah.
Eric Lofgren: [00:53:55] So I guess, you brought it back to, I think the answer that I like to.
Poke on, because it seems like monopsony from the government side. So you’re saying the monopsony structure of the government and its decisions over programs, because it used to be, you might have multiple customers in the DOD, right? The services could have competed and done stuff like competed against each other, had redundant programs, but that’s gone away.
So because of the, monopsony where the government is, the single buyer, it actually makes that an unattractive place for people to enter because. I guess the whole point of monopsony is that you have the market power to drive down their prices so that they’re always at zero economic profit or something like that.
Anne Marie Knott: [00:54:37] Yup. Yup. And then they make it worse by things like second sourcing.
Eric Lofgren: [00:54:41] So if we’re talking about defense industry competition as a problem, and we always hear about government folks, we need more competition and the whole defense industry would be. A giant monopoly, right? Like in we’re trending that way.
No monopoly. Yeah. Actually, because right, because Lockheed Martin has basically a monopoly on fifth generation fighters, Northrop Grumman has a monopoly on stealth bombers. Boeing is going to be the only tanker maker. So the government’s monopsony is creating monopolies or poor competition incentives on the other side.
So if, when government folks talk about. competitive problems in the industry, they should be looking at themselves. Is that kind of what you’re? I think
Anne Marie Knott: [00:55:26] so. So to look at a commercial equivalent, the closest commercial equivalent that I can think of is Toyota. So Toyota. Is a huge customer.
So they could be the equivalent monopsony for many of their suppliers, taking like 80% of their output. And yet they still continue to maintain multiple suppliers for most of their goods. So it’s interesting that they’re able to do that. But that might be a good model for the DOD.
Eric Lofgren: [00:55:55] So I want to be right before we wrap up here. I want to get to this big question that has been concerning a lot of folks and myself included, and it’s that debate on whether science is slowing down. And one of the big examples that a lot of people bring up is this idea that, okay we have Moore’s laws, so that’s increasing the.
The number of transistors that we can put, or is increasing, double every 18 months. But I actually have to put more and more resource inputs into this. So the number of researchers has gone up 24 times just to maintain this constant growth rate in terms of Moores law. So one of the big explanations everyone’s pointing to is the low hanging fruit.
So we’ve already captured all the low hanging fruit, and now it’s more difficult to make those next advances.
Anne Marie Knott: [00:56:40] I see a very differently as you probably know. So I actually have a, there’s a paper out it’s called our ideas getting harder to find and it makes this argument. And they specifically go into the example that you’re talking about, which is the Moore’s law. That paper is based on a theory from Chad Jones that that he advanced in his paper from 1995, which also was the first one too, which is the first place I know of that documented this.
Decline at the aggregate level in R and D productivity. And his argument is that there’s a problem with Romer’s endogenous growth theory in that, or not that if there’s a problem with it, but that he is a special form of a more general functional forum in which the so I don’t know if romer, but basically growth in the economy is tied to growth in ideas or growth in knowledge.
And the growth in knowledge in turn is a function of how much knowledge you have to begin with. How many researchers you have doing R and D and how productive the researchers are in doing R and D. And so Chad Jones, his revision of that is to say that the Romer has argued that the exponent on knowledge is one.
So constant returns so what Jones was saying is no, there’s not constant returns to knowledge, there’s decreasing returns. to knowledge. he calls that fishing out. The idea that all the good ideas have used up. And so it had been used up. And so we’re continually fishing out core important ideas.
But he never, so they have a paper, the paper that came out cheer had or last year now it was April of 2020. Which includes Chad Jones as one of the coauthors is nominally testing has idea, but they actually never test the exponent on knowledge. And so we have a paper where we do that and show that’s, that there is no evidence of decline in the elasticity of the stock of knowledge.
Okay. So empirically there isn’t support for it now, how do they, how, what are they doing in a paper in the paper? They are taking specific examples of technologies. Or fields and they’re showing decline. Now we know for all time, three different domains have characterized these S-curves and technology, right?
So we had them in the sociology of science. We have them in evolutionary economics and we have them in the management. Yeah, we actually have one in the management field. And S the norm is that when a new technology comes out, it’s slow to it’s slow to produce anything. Then we’ve got this dramatic rise.
And then we get. A point of diminishing returns in that technology, which is where we are with respect to with respect to Moore’s law and with respect to all the things that they look at in this paper. Now, the really hopeful thing is, and, we know this from past sociology of science, et cetera, et cetera, is that as a new technology becomes exhausted.
It becomes replaced by another technology, right? So there’s this renewal. And we find in this, in the work that we’re doing is that there is in fact declining art that RQ declines within within domains over time. So if you take the maximum RQ, so their argument is motivated by mean, meaning.
Average RQ, but if you actually, if it’s true that ideas are declining, the whole distribution of RQ should be declining. And what we find is that it’s at the economy level, it’s actually increasing over time. RQ. If you take the maximum, you’ve taken all the firms in each year and look at the maximum RQ in that year and go year a year, that’s increased that level of RQ is increasing over time.
It’s only decreasing if you look within industries which I think is really cool. It’s, as technologies decay, and then what’s happening is we create new firms or new technologies that replace them, from organic growth. So I think it’s
Eric Lofgren: [01:00:35] optimistic. Yeah, you definitely have the optimistic view.
And I think it’s also an intuitive view that kind of conforms with, when I look at the world, it seems to make sense there. And it seems like a lot of those guys are like we we don’t have the same progress in, physical stuff, but then it’s okay you’re throwing out all of the progress we have in the digital space and saying that’s not, part of this S-curve.
One of the things that. I’d like to get your view on this. It seems like the researchers, what they’re trying to do was be agnostic as to like industry. So they’re just saying I want a general measure of something like how many computations can I do per dollar per second or something like that.
And so I should be jumping between S curves in order to keep that same growth along that attribute path. So why If I’m using that kind of measure why does that not work out in the same way as what you’re saying? If let’s okay. Moore’s
Anne Marie Knott: [01:01:29] law’s density on a chip, isn’t it
Eric Lofgren: [01:01:32] correct?
Yeah. So that is a technical measure, but I think, I don’t know if they were using moore’s law, they were using some kind of. Surrogate for it, which was getting back to these more general performance attributes. So I guess, another one that they used was like speed or something of transportation, which would be like, okay, I can move from a cart, a horse-drawn cart to railroad to truck, transport, to aircraft transport or something like that, and so you should be jumping. How do you think about those measures? Are, if they’re all going down. What does that say? Or is that just we’re just not measuring the right thing. We should be shifting how we measure things. So let me give you a
Anne Marie Knott: [01:02:10] cue. So first of all, firm, when I talk to companies, they laugh about, at this idea that we’re running out of ideas, right?
Think about self-driving cars. You think about the internet, you think about smart phones. So my so what I would throw out to you is, okay, let me let me give you a measure. And so let me give you a measure, which is how many keystrokes can I do in an hour. You can measure that, but then, what would be, that’s not meaningful, right?
Because pretty soon we’re not typing keystrokes anymore. We’re doing other things.
Eric Lofgren: [01:02:40] Yeah. I think that, that’s gotta be the answer. I think that we have all these numbers of attributes we could be measuring and those are ever changing and they’re all incommensurable against each other. So once I choose one attribute, I have to like, A cell phone what am I going to measure like that it’s picture or it’s like ability to communicate or it’s access to the internet and then they’re all incommensurable.
So I don’t really know how to collapse all that. So I think that’s the key is maybe just like we’re focused, we’re trying to focus on the wrong thing and potentially that’s where the RQ and your S your studies are helping. Cause it is a more general measure. Right then, like I have to select the one thing
Anne Marie Knott: [01:03:21] each of the measures that they look at or are no they try to do an equivalent of our RQ too.
They look at firms. But yeah, that’s the, you were asking before about the disadvantage of RQ. So RQ is not a project level measure. So we can help you with that. And it can’t help you with. It can help you with things like labs, right? Because it relies on a production function and production functions are in the labs.
Don’t generate revenues from their own. But getting back to the, where you were talking about the issue of measuring it, as soon as you identify a technology and you get a measure that’s tied to that technology, you run the risk that, you’re going to miss the, your measurements are going to miss the the things that you care about.
Eric Lofgren: [01:04:01] So what’s next for you or any big questions that you’re trying to tackle?
Anne Marie Knott: [01:04:07] It’s still looking for the nails, right? So one of the new nails is the, the corporate research labs. Mostly what I’m trying to do is I’m trying to diffuse the RQ. Thought firms, as soon as they’ve had this measure would just jump at it.
And they actually aren’t, which has been really frustrating for me because I thought as soon as they would have it, they would want to improve their RQ. And then we would solve that whole, we would reverse this 65% decline and we would revive economic growth. That’s not working. And what I’m, what I’ve been doing is realizing that the two levers that are going to be pretty more powerful for me are the investor level, because firms are responsive to their investors, but I also like the DOD lever.
So I’ve been working, not just DOD, but the federal government in general. And I think that trying to come up with ways that policy makers can. Provide the right incentives for companies to improve their RQ. So one would be, changing tax credits so that they reflect RQ or improvements in RQ.
Having iRead reimbursement, be tied to RTU and having R and D context contracts. Utilize RQ is one of the criteria of merit,
Eric Lofgren: [01:05:10] some interesting stuff. And I’m going to be interested to follow up and see how that goes. Anne Marie not thanks for joining me on an acquisition talk. Her book is how innovation really works.
So audience, please check.
Anne Marie Knott: [01:05:22] Thanks very much, Eric. It’s a pleasure speaking with you. I appreciate the opportunity
Leave a Reply