Remote teams have traditionally been bad at breakthrough innovation, but things may have changed
Like the rest of New Things Under the Sun, this article will be updated as the state of the academic literature evolves; you can read the latest version here.
Remote work seems to be well suited for some kinds of knowledge work, but it’s less clear that it’s well suited for the kind of collaborative creativity that results in breakthrough innovations. A series of new papers suggests breakthrough innovation by distributed teams has traditionally been quite difficult, but also that things have changed, possibly dramatically, as remote collaboration technology has improved.
Distant and Colocated Collaboration Are Not Alike
We can begin with Van der Wouden (2020), which looks at the history of collaboration between inventors on US patents, over the period 1836 to 1975. To build a useful dataset, he has to extract the names and locations of inventors from old patent documents, which have been digitized into super messy text files by Google. These digitized patents are rife with misspelling (because optical character recognition scanning is very imperfect for old documents) and lacking in much of any standardization.
It’s a ton of work that involves fuzzy matching text strings to a big list of names which in turn is drawn from the US census, modern patent documents, and an existing database of inventor names. And that’s only the first step - it just tells you the names of people mentioned in a patent, not whether those names are inventors, rather than lawyers or experts. To figure out who is an inventor, Van der Wouden uses a set of classification algorithms that predict the probability a mentioned name is an inventor using a dataset of known inventors linked to patents. It’s not a perfect method, but it is able to find an inventor on about 90% of historical patents. Moreover, the people it identifies as top patent holders, and the number of patents they hold, matches pretty closely other lists of top patentees in US history. He also has to do similar work to pull out the locations mentioned on a patent.
Now that he has an estimate of how many people worked on each patent, and where they lived, Van der Wouden can start to look at how common collaboration and remote collaboration are. We can see that collaboration really began to take off in the 1940s and that the probability a team of inventors didn’t reside in the same city rose from under 5% in 1836 to over 10% by 1975.
Van der Wouden next tries to measure the complexity of a patented invention with an approach originally used in another paper, Fleming and Sorenson (2004).1 Fleming and Sorenson attempted to create a measure of how “fussy” technological classifications were, based on how well they seem to play nice with other technologies (fussy is my term, not their’s, but I think it captures what they’re going for in a colloquial way). If a technological classification is frequently attached to a patent alongside a wide range of other classifications, they’re going to say this isn’t a very “fussy” technology. It can be used in plenty of diverse applications. On the other extreme, if a classification is only ever assigned to a patent with one other classification, then we’re going to assume the technology is very sensitive and very fussy. It only works well in a very specific context.
While this measure is a bit ad-hoc, Fleming and Sorenson also did a survey of inventors and showed their measure is correlated with inventors self-assessments of how sensitive their own inventions are to small changes, and that this measure is not merely picking up how novel or new the technology is; it’s picking up something a bit different. Returning to Van der Wouden (2020), his measure says a patent is more complex if it involves more technologies, and if these technologies are “fussy.”
There are two key results: complex patents are more likely to be the work of teams. And among patents by a team of inventors, the inventors are more likely to reside in the same city if the patent is more complex. It seems that, at least over 1836-1975, it is hard to do complex work at a distance.
Lin, Frey, and Wu (2022) pick up Van der Wouden’s baton and take us into the present day. They look at the character of both patents and academic papers produced by collocated and remote teams over 1960-2020 (actually 1975-2020 for patents), but focusing on how disruptive a paper or patent is. To measure disruption, they use an increasingly popular measure based on citations. To simplify a bit, the idea here is that if a paper or patent is disruptive, you’re not going to cite the stuff it cites, because the paper or patent has rendered those older ideas obsolete. After Einstein, you no longer cite Newton. On the other hand, if a paper is an incremental improvement within a given paradigm, you are likely to cite it as well as its antecedents. This disruption measure quantifies this notion: for some focal document, it’s based on how many citations go to the focal document alone relative to how many citations go to the focal document as well as the documents cited by the focal document.
Across 20mn research articles and 4mn patents, Lin, Frey, and Wu find that, on average, the farther away the members of the team are from one another, the less likely the paper is to be disruptive.
So, over 1836-1975 the patents of inventors who reside in the same cities tended to be more complex, in the sense that they either drew on more technologies, or more technologies that don’t have a long history of successfully being combined with other technologies. And over 1975 to 2020, patents with inventors residing in the same city were more likely to be disruptive, in the sense that they are more likely to receive citations that do not also reference earlier work.
Does Distance Inhibit Strange Combinations?
These measures are not picking up exactly the same thing, but neither are they as different as they might seem at first. As discussed in a bit more detail here, Lin, Evans, and Wu (2022) find that papers that draw on novel combinations of ideas (in this paper, proxied by the kind of journals a paper cites) are also more likely to be disruptive. In other words, it might well be that the reason Lin, Frey, and Wu find papers by distant teams are less likely to be disruptive is because dispersed teams have a harder time connecting different ideas.
We’ve got a few pieces of evidence that support the notion that remote teams have a harder time making novel connections across ideas.
First, both Berkes and Gaetani (2021) and Duede et al. (2022) find some evidence that colocation is an important channel for exposure to intellectually distant concepts. As discussed here, Berkes and Gaetani (2021) show that:
The patents of inventors residing in denser parts of cities comprise a more diverse set of technologies
The set of technologies that comprise the patents of denser parts of cities is more unorthodox: two different technologies might rarely originate from the same geographical location, but when they do that area is more likely to be a dense part of a city
The patents of inventors residing in denser parts of cities are more likely to feature unusual combinations of technologies themselves.
That’s all consistent with the idea that being physically around lots of different kinds of inventive activity increases the chances you draw an unexpected connection between two disparate concepts.
Duede and coauthors provide some fine-grained evidence from academia. They have a big survey where they ask thousands of academics across many fields about citations they made in some of their recent work. Among other things, they asked respondents how well they knew the cited paper, as well as how influential was the citation to the respondent’s work. In the latter case, respondents rated their citations on a scale from “very minor influence”, which meant the respondent’s paper would have been basically unchanged without knowledge of the cited reference, to “very major influence”, which meant the cited reference motivated the entire project.
If we have a way to measure the geographic distance between the authors and the “intellectual distance” between the citation and the author’s normal expertise, we can see how the two are related: does being close in space facilitate learning about ideas you wouldn’t normally know about? Computing distance in space is straightforward: Duede and coauthors just code whether authors are in the same department, same institution, same city, or same country. To measure intellectual distance, they rely on the similarity of the title and abstract of the citing and cited paper, as judged by natural language processing algorithms. This algorithm judges papers to be more similar if they contain words that are themselves more closely related to each other.
Duede and coauthors find if you and the author of a paper you cite are at the same university, then you are indeed more likely to say you know the cited work well and that it was influential on you. But what’s interesting is that the strength of this relationship is stronger if the cited and citing paper are less similar to each other. In other words, if you cite a paper that’s surprising, given the topic you are working on, you are more likely to say you know that paper well and that it influenced you if the author is at the same university. That’s quite consistent with colocation being a useful way to learn about ideas you wouldn’t otherwise encounter in the course of your normal knowledge work.
The second line of evidence is larger, but less direct: physical proximity seems to be quite important for helping people form new relationships, especially relationships that wouldn’t have been formed in the course of ordinary knowledge work. I’ve looked at this line of evidence in two different ways.
In Innovation at the office, I looked at a few experimental and quasiexperimental studies that showed people are more likely to share information or collaborate with people who are physically closer in their workspace. However, this seems to be the case mainly because physical proximity helps connect people who would not otherwise connect. It’s not actually that important for groups that already know each other to be physically close.
Why proximity matters: who you know looked at a broader suite of evidence that communication across distances tends to be pretty robust so long as the people communicating have social ties with each other. But these papers (and lived experience!) also suggest proximity is important for getting to know people in the first place.
So, although remote work might be well suited for some kinds of tasks, perhaps it is poorly suited for facilitating new and unusual combinations of ideas. And there could be a simple reason for that: if remote collaboration does not expose you to new and different people and ideas, then it might lead to less breakthrough innovation, even if collaboration with people at a distance is no longer a meaningful barrier to getting stuff done.
Who you know
Frey and Presidente (2022) provides some further support to the notion that the weak link in remote collaboration might be in forging networks, rather than in facilitating productive work among dispersed people who already know each other. They do this by basically freezing the network of you who know. Like Lin, Frey, and Wu (2022), this paper looks at how disruptive academic papers are when the team of coauthors is colocated versus when they are geographically dispersed. Unlike Lin, Frey, and Wu though, they focus their analysis on what happens to the disruptiveness of a given team’s papers when they are together or apart. To illustrate, suppose a pair of economists reside in the same department and begin a productive collaboration, writing a paper together each year. Later, one of them moves to a distant university, but the pair continues to collaborate. What is the difference in the disruptiveness of their papers when they are apart, compared to when they are together?
In contrast, Lin, Frey, and Wu, discussed earlier, compared all papers by geographically dispersed teams to all papers of geographically proximate ones; most of the time these papers would be by completely different teams. If our story about network formation being important is right, this difference matters: geographically colocated teams might be more likely to be comprised of people who wouldn’t normally encounter each other, but for the accident of having offices near to each other. That, in turn, might enable them to come up with more disruptive ideas. With the new paper by Frey, and Presidente, we take the potential differences in team composition out of the equation, because we are always comparing the output of the exact same teams when they are all together versus when they are apart.
When you do this, you find that prior to 2000 your team produced less disruptive work when it was spread out versus when it was all together. But that trend has been weakening over time, and from 2000 onwards, dispersed teams have not been less disruptive than collocated ones. In fact, over 2015-2020, your team produces more disruptive work when it is separated than when it is together (though the effect is pretty small)!
There is a pretty obvious explanation for this finding, at least for 1980-2015. Better digital communication (and possibly cheaper travel) has steadily eroded the difficulties of collaborating at distance. In fact, Frey and Presidente show the historically negative effects of going remote are smaller for researchers who reside in countries with more internet penetration. The finding that colocation matters less and less for innovation is quite a consistent finding across a host of different studies (see here, here, and here for more on this). In fact, the post An example of successful innovation by distributed teams: academia looked at this issue specifically for academia and argued physical proximity has mattered less and less for specifically academic collaboration, at least for the exchange of ideas and the productivity of researchers. We can now see the same thing for the nature of research performed.
But the post-2015, calls for a different explanation (though at the same time, I don’t want to lean on this too hard, because the effect sizes are small). Conditional on a team of collaborators already knowing each other, after 2015 they are more disruptive remotely than together. Why should that be? We’re used to thinking of better communication technology reducing the frictions of communicating at a distance, not that we might communicate/collaborate better at a distance. What’s going on?
I’ll turn to Frey and Presidente’s theory in a second, but as a quick aside, one possibility is that remote collaboration short-circuits some of the processes that normally hamper group collaboration. For example, sometimes the free exchange of ideas and perspectives in a team can be inhibited by status hierarchies or dominant personalities: a shy member of a team, eager to get along, just goes along with the group rather than sharing a concern. But a zoom call, slack conversation, or email back-and-forth can have very different social dynamics that might make it easier for some types of people to share their ideas. Alternatively, a colocated team might tend to brainstorm together (for example, over lunch), which might lead to fewer ideas getting expressed, as groupthink and social dynamics take over the process. If it’s unpleasant to have these group brainstorms over a zoom session, the alternative might be that each team member independently generates ideas, and then shares their thoughts. You might end up with more total ideas expressed.
Pulling from a non-local pools of ideas
The above is one reason why it could be that remote teams have recently begun to produce slightly more disruptive ideas than colocated teams. Frey and Presidente’s theory, however, is different. Suppose we believe, as we’ve argued, that colocation is useful for exposing us to ideas we have not normally been exposed to but also that in the modern world it is quite easy to collaborate at a distance. We can then ask, which team will collectively be exposed to more intellectually distant ideas: the one that’s all in the same place, or the one that is spread over many different geographies?
It would seem to be the latter.
To test this, Frey and Presidente look at what happens to the disruption of a remote team’s papers as they move to more highly ranked univerisity (as judged by the Times Higher Education supplement). They find that, after 2010 (but not before), when remote team members move to more highly ranked universities, their papers become more disruptive. There is no analogous bonus for a colocated team. One possible explanation is that a pair of collaborating researchers who move together to a single highly ranked university enjoy the benefits of being exposed to one university-worth of ideas, while a pair of researcher who split up and move to two different universities enjoy the benefits of exposure to two universities worth of ideas.
Possibly, but it’s worth noting that it’s normally hard to identify much benefit to individual researchers of relocating to higher ranked departments (though note departments are not the same as universities). However, I think Esposito (2021) provides some nice complementary evidence to this exercise by Frey and Presidente.
Esposito (2021) takes us back to patents, and like Van der Wouden (2021), is interested in how remote teams of collaborators produce different kinds of innovations. Esposito looks at patents in two ways. First, he classifies patents as novel or normal, based on whether they are assigned unconventional combinations of technology codes by the patent office. Second, he measures the impact of a patent based on the number of subsequent patents that are assigned the same combination of technology codes, with a weighting scheme so that patents get less credit the more patents there are out there making the same combination at the same time or earlier. Note these schema are distinct: you can be novel and high impact, or novel and low impact. Or you can be normal and high impact or normal and low impact.
Esposito builds on Van der Wouden (2020)’s dataset to determine whether patents are created by colocated or distributed teams. But he goes one step further, and divides locations into two types. First, knowledge-diverse regions are regions where a diverse set of different kinds of inventive activity are taking place. He identifies these by finding regions patenting in at least 10% of the currently active technology classifications. Second, knowledge-homogenous regions are those where patenting is concentrated in a small number of fields, or where there is patenting in less 10% of technologies.
So now Esposito is in a position to see what kinds of factors encourage the production of high-impact patents. The following figure shows how the average impact of a patent changed over time, as a function of what kind of patent it was, the type of region where it was invented, and whether the team was colocated or distributed. What we’re actually seeing is the difference in impact from what would be predicted by knowing just the year the patent was granted and the number of technological components assigned to it.2 We want to strip that stuff out, and focus just on the impact of novelty, type of region, and whether a team is remote or not. The figure is a bit complex, because it cuts the data up in several different ways, so let’s take it bit by bit.
There are 8 different lines on here, because Esposito divides up the data into novel and normal patents, and then he subdivides those patents into ones by remote or colocated teams, and then he subdivides each of those categories into the type of region the teammates reside - a knowledge diverse city or not. We can see, prior to 1975, all those divisions didn’t make much of a difference in terms of the extra impact they imparted on a patent. But then, after 1975, things begin to drift apart. Most notably, the green line shoots up.
What is that green line? That is the average impact of a novel patent, where the team is remotely collaborating from different knowledge diverse cities. In other words, this figure shows the highest impact patents where those where a team of remote inventors, exposing itself to multiple cities brimming with diverse technologies, made a strange combination. This is exactly the story we were telling above - in a world where remote collaboration is easy, teams with knowledge “scouts” embedded in multiple knowledge-rich communities may have an advantage over teams colocated in a single city, even if it’s a knowledge-rich city.
In a distant second, Esposito finds a tie between novel patents created by colocated teams residing in knowledge rich cities (orange line) and normal patents by remote teams residing in multiple knowledge rich cities (green line). And further underscoring the importance of being in knowledge diverse cities for doing novel work, the purple line and pink lines that are fighting for last place on impact since 1975 are novel patents born from teams residing entirely in knowledge-poor regions (whether multiple cities or just one).
The bottom line is for most of the twentieth century, the conventional wisdom about the importance of being together in the middle of a vibrant scientific/technological environment in order to do cutting edge work seems to have been correct. Teams working with each other over distances had a harder time doing complex work or disruptive work. Maybe that was because remote communication was so hard. But maybe it was also because it was harder to make connections across distance, in two senses of the word: it was harder to meet new people, especially people you wouldn’t otherwise meet in the course of your interests, and harder to learn about intellectually distant ideas, which might be usefully connected to stuff you already know.
I think it’s plausible that these remain serious barriers to remote innovation, even today, though I look forward to the inevitable studies about whether digital platforms can take on some of the role of physical proximity.
However, what seems to have changed is that once you have found your collaborators, it may now be possible to do very creative breakthrough work even across distances. And in fact, if a team can work together nearly as well apart as they can together, then it may now be optimal to scatter the team across multiple places, the better to draw in ideas from far corners of the world.
Thanks for reading! This post mentioned the following articles I’ve written for New Things Under the Sun:
As always, if you want to chat about this post or innovation in generally, let’s grab a virtual coffee. Send me an email at mattclancy at hey dot com and we’ll put something in the calendar.
New Things Under the Sun is produced in partnership with the Institute for Progress, a Washington, DC-based think tank. You can learn more about their work by visiting their website.
The rest of this paragraph and most of the following one shamelessly plagiarize my description of the Fleming and Sorenson measure in Science as a Map of Unfamiliar Terrain.
That is, we’re seeing residuals in a regression. Given the way Esposito measures impact, patents will have higher impact over time, as the number of patents increases, and if they are assigned more technology codes (whether novel or not).