What do you think?
Rate this book
Small Wars, Big Data: The Information Revolution in Modern Conflict
How a new understanding of warfare can help the military fight today's conflicts more effectively
The way wars are fought has changed starkly over the past sixty years. International military campaigns used to play out between large armies at central fronts. Today's conflicts find major powers facing rebel insurgencies that deploy elusive methods, from improvised explosives to terrorist attacks. Small Wars, Big Data presents a transformative understanding of these contemporary confrontations and how they should be fought. The authors show that a revolution in the study of conflict--enabled by vast data, rich qualitative evidence, and modern methods--yields new insights into terrorism, civil wars, and foreign interventions. Modern warfare is not about struggles over territory but over people; civilians--and the information they might choose to provide--can turn the tide at critical junctures.
The authors draw practical lessons from the past two decades of conflict in locations ranging from Latin America and the Middle East to Central and Southeast Asia. Building an information-centric understanding of insurgencies, the authors examine the relationships between rebels, the government, and civilians. This approach serves as a springboard for exploring other aspects of modern conflict, including the suppression of rebel activity, the role of mobile communications networks, the links between aid and violence, and why conventional military methods might provide short-term success but undermine lasting peace. Ultimately the authors show how the stronger side can almost always win the villages, but why that does not guarantee winning the war.
Small Wars, Big Data provides groundbreaking perspectives for how small wars can be better strategized and favorably won to the benefit of the local population.
The way wars are fought has changed starkly over the past sixty years. International military campaigns used to play out between large armies at central fronts. Today's conflicts find major powers facing rebel insurgencies that deploy elusive methods, from improvised explosives to terrorist attacks. Small Wars, Big Data presents a transformative understanding of these contemporary confrontations and how they should be fought. The authors show that a revolution in the study of conflict--enabled by vast data, rich qualitative evidence, and modern methods--yields new insights into terrorism, civil wars, and foreign interventions. Modern warfare is not about struggles over territory but over people; civilians--and the information they might choose to provide--can turn the tide at critical junctures.
The authors draw practical lessons from the past two decades of conflict in locations ranging from Latin America and the Middle East to Central and Southeast Asia. Building an information-centric understanding of insurgencies, the authors examine the relationships between rebels, the government, and civilians. This approach serves as a springboard for exploring other aspects of modern conflict, including the suppression of rebel activity, the role of mobile communications networks, the links between aid and violence, and why conventional military methods might provide short-term success but undermine lasting peace. Ultimately the authors show how the stronger side can almost always win the villages, but why that does not guarantee winning the war.
Small Wars, Big Data provides groundbreaking perspectives for how small wars can be better strategized and favorably won to the benefit of the local population.
408 pages, Hardcover
Published June 12, 2018
54 people are currently reading
369 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 13 of 13 reviews
February 10, 2020
I rarely rate a book 4 stars. This one book encapsulated an entire graduate school years’ study on the Political Economy of Civil Wars and Peacebuilding. The references to the First Infantry Division and colleagues only endeared me to their overall narrative.
Read
March 10, 2026Such an exciting title, but it just means turning the regression crank with micro-level data, that is often how these big data things are.
A scholarly monograph, you know the type, careful conclusions that are well defended but seem slippery and contingent.
When they justify the value of this study they sometimes sound defensive, I speculate because of the tons of Minerva etc. cash they got for this.
The beginning is an accessible version of the model [Civilian information can give counterinsurgents the decisive edge. Asymmetric warfare involves insurgents and counterinsurgents influencing civilians to provide information. This is the vital way in which symmetric conflict differ.]
The body of the book is a set of propositions related to this model, supported by studies and data of various conclusiveness, the authors briefly consider alternative explanations as well.
The final chapter, 10, is something entirely different. It is the “why this matters” chapters we expect, asymmetric conflict will increase, and become more important, all the questions the book answers, and the broader project of micro-level data social science, all social science, is a useful endeavor. There is a lot of substance here, both data and speculation, theory and argument, and these are the ideas that I am most interested in, so I wish there was more, though a fraught wish.
Some points that stood out to me: existing data on development is (more or less) useless in conflict zones, but governments and NGOs fail to adapt. Additional micro-level research can explain the interaction between economic growth and political development (especially in conflict areas), which is (they imply) the knot that ties up many failed states. “Sober, reasoned public debate” [Ben Rhodes’ formulation] is likely politically impossible, but social scientists are best able to fill that role. Social science is uniquely good because peer-review and professional reputation create a community with the right incentives for transparent analysis informed by real-world data. Economists have parlayed this into policy through CEA in government and NBER out of it, but social scientists who study security have failed to consolidate their position. They and their allies should seek to emulate the economists (WOW! Destler, Gelb, and Lake?).
Tactical victories do not necessarily translate into strategic victories [“conflict-level political settlement that would end an asymmetric war”]; micro-data informs tactical victories but security studies/ESOC mold should focus more on how tactics build to grander outcomes. Current doctrine has “the aggregation challenge:…it assumes away a range of challenges involved in going from local victories to an overall political settlement” [sounds familiar…?]. “Unfortunately, the current research base provides little consistent guidance on how to meet that challenge”… can it? They speculate yes because local success can 1. Reduce violence to a level society can tolerate and 2. Open up political opportunities to settle underlying issues. They add that principle-agent models (foreign power-local ally) might provide more guidance.
Some say drones are a way around this issue, since they can decapitate insurgencies. Our authors are not convinced: they do not protect civilians because they cannot control territory, they endanger civilians when intelligence fails, and they cannot set up a political solution. However, they can be an effective compliment to counterinsurgency campaigns.
Finally, they offer headline prompting for declassification, I intuit targeted mostly at the military and government, but I am curious for details of declassification politics and anthropology.
In coda, some claims from the substantive chapters, so as not to forget:
1-2: asymmetric wars will be important for military and development reasons going forward. Micro-data is useful for understanding asymmetric wars.
3: model describing interaction between insurgents, counterinsurgents, and civilians. The key mechanism is civilian willingness to provide information to counterinsurgents.
4: Technology that makes it easier for civilians to provide tips is generally good for counterinsurgency (i.e. cellphone service/towers). Sometimes it also benefits insurgents, because it allows them to communicate more easily, but generally, in asymmetric conflicts, it helps the counterinsurgents more.
5: Small, secure, conditional aid helps counterinsurgents win tips, and large aid does not, sometimes it makes conflict worse. Self-interest, gratitude, and grievance based explanations for aid are less important than this: civilians rationally calculate if they should provide more tips to counterinsurgents [and aid can influence this calculus].
6: Service provisions work better when there is a security force suppressing violence. Force can increase or reduce violence depending on how it is applied and how civilians interpret it (the authors say this, almost verbatim, on page 181). There is more here, including about attitudes, lots that the authors say we cannot test well empirically.
7: Combatants avoid harming civilians and mitigate the consequences when they do, civilians have worse attitudes about those that harm them, overall violence goes down when rebels cause civilian casualties and goes up after governments do.
{on re-examining 4-7, if NGOs and others failed to realize these points (as the authors claim), I feel this work is a much much much more severe criticism of those orgs’ institutional sclerosis than prodict of social science. Perhaps the authors thought this but could not write it, for obvious reasons}
8: outside the model in chapter 3, thinking about civilian willingness to provide tips as a key mechanism has other implications, like a market for tips. The supply-demand explanation for insurgents is weak, insurgents require few people and have plenty of supply, so interventions meant to redirect possible insurgents to satisfying jobs fail. Poor rural civilians like insurgents less than their middle-class urban counterparts, contrary to popular belief [based on survey data – as an idea this is truly interesting, there is more to think about here]. Certain types of economic prosperity give insurgents more money and resources making violence worse.
9: interaction of local-knowledge and high training provides for effective counterinsurgency forces. Better trained forces are better at counterinsurgency. Civilians give credit to foreign militaries for disaster relief (good things generally?) when those efforts are visible and don’t when those efforts are opaque. Foreign militaries do not gain and sometimes suffer when their adversaries can steal or obscure credit for what the militaries do. Marginal changes in tip flow can significantly enhance counterinsurgency campaigns, so counterinsurgents can often achieve their goals with small interventions.
10: see above
A scholarly monograph, you know the type, careful conclusions that are well defended but seem slippery and contingent.
When they justify the value of this study they sometimes sound defensive, I speculate because of the tons of Minerva etc. cash they got for this.
The beginning is an accessible version of the model [Civilian information can give counterinsurgents the decisive edge. Asymmetric warfare involves insurgents and counterinsurgents influencing civilians to provide information. This is the vital way in which symmetric conflict differ.]
The body of the book is a set of propositions related to this model, supported by studies and data of various conclusiveness, the authors briefly consider alternative explanations as well.
The final chapter, 10, is something entirely different. It is the “why this matters” chapters we expect, asymmetric conflict will increase, and become more important, all the questions the book answers, and the broader project of micro-level data social science, all social science, is a useful endeavor. There is a lot of substance here, both data and speculation, theory and argument, and these are the ideas that I am most interested in, so I wish there was more, though a fraught wish.
Some points that stood out to me: existing data on development is (more or less) useless in conflict zones, but governments and NGOs fail to adapt. Additional micro-level research can explain the interaction between economic growth and political development (especially in conflict areas), which is (they imply) the knot that ties up many failed states. “Sober, reasoned public debate” [Ben Rhodes’ formulation] is likely politically impossible, but social scientists are best able to fill that role. Social science is uniquely good because peer-review and professional reputation create a community with the right incentives for transparent analysis informed by real-world data. Economists have parlayed this into policy through CEA in government and NBER out of it, but social scientists who study security have failed to consolidate their position. They and their allies should seek to emulate the economists (WOW! Destler, Gelb, and Lake?).
Tactical victories do not necessarily translate into strategic victories [“conflict-level political settlement that would end an asymmetric war”]; micro-data informs tactical victories but security studies/ESOC mold should focus more on how tactics build to grander outcomes. Current doctrine has “the aggregation challenge:…it assumes away a range of challenges involved in going from local victories to an overall political settlement” [sounds familiar…?]. “Unfortunately, the current research base provides little consistent guidance on how to meet that challenge”… can it? They speculate yes because local success can 1. Reduce violence to a level society can tolerate and 2. Open up political opportunities to settle underlying issues. They add that principle-agent models (foreign power-local ally) might provide more guidance.
Some say drones are a way around this issue, since they can decapitate insurgencies. Our authors are not convinced: they do not protect civilians because they cannot control territory, they endanger civilians when intelligence fails, and they cannot set up a political solution. However, they can be an effective compliment to counterinsurgency campaigns.
Finally, they offer headline prompting for declassification, I intuit targeted mostly at the military and government, but I am curious for details of declassification politics and anthropology.
In coda, some claims from the substantive chapters, so as not to forget:
1-2: asymmetric wars will be important for military and development reasons going forward. Micro-data is useful for understanding asymmetric wars.
3: model describing interaction between insurgents, counterinsurgents, and civilians. The key mechanism is civilian willingness to provide information to counterinsurgents.
4: Technology that makes it easier for civilians to provide tips is generally good for counterinsurgency (i.e. cellphone service/towers). Sometimes it also benefits insurgents, because it allows them to communicate more easily, but generally, in asymmetric conflicts, it helps the counterinsurgents more.
5: Small, secure, conditional aid helps counterinsurgents win tips, and large aid does not, sometimes it makes conflict worse. Self-interest, gratitude, and grievance based explanations for aid are less important than this: civilians rationally calculate if they should provide more tips to counterinsurgents [and aid can influence this calculus].
6: Service provisions work better when there is a security force suppressing violence. Force can increase or reduce violence depending on how it is applied and how civilians interpret it (the authors say this, almost verbatim, on page 181). There is more here, including about attitudes, lots that the authors say we cannot test well empirically.
7: Combatants avoid harming civilians and mitigate the consequences when they do, civilians have worse attitudes about those that harm them, overall violence goes down when rebels cause civilian casualties and goes up after governments do.
{on re-examining 4-7, if NGOs and others failed to realize these points (as the authors claim), I feel this work is a much much much more severe criticism of those orgs’ institutional sclerosis than prodict of social science. Perhaps the authors thought this but could not write it, for obvious reasons}
8: outside the model in chapter 3, thinking about civilian willingness to provide tips as a key mechanism has other implications, like a market for tips. The supply-demand explanation for insurgents is weak, insurgents require few people and have plenty of supply, so interventions meant to redirect possible insurgents to satisfying jobs fail. Poor rural civilians like insurgents less than their middle-class urban counterparts, contrary to popular belief [based on survey data – as an idea this is truly interesting, there is more to think about here]. Certain types of economic prosperity give insurgents more money and resources making violence worse.
9: interaction of local-knowledge and high training provides for effective counterinsurgency forces. Better trained forces are better at counterinsurgency. Civilians give credit to foreign militaries for disaster relief (good things generally?) when those efforts are visible and don’t when those efforts are opaque. Foreign militaries do not gain and sometimes suffer when their adversaries can steal or obscure credit for what the militaries do. Marginal changes in tip flow can significantly enhance counterinsurgency campaigns, so counterinsurgents can often achieve their goals with small interventions.
10: see above
July 8, 2018
Excellent overview of the academic literature on small wars and insurgencies developed by the authors over the previous decade. More than most political science books, Small Wars, Big Data has a strong and clear sense of purpose along with a lot of evidence that can help academics, think tank researchers, and government analysts--particularly by proposing their "information-centric" theoretical framework, which has empirical implications that can be examined in a range of contexts, as illustrated by the examples of applied research that illuminate each chapter.
December 24, 2021
Despite this book's title which leverages the popular technology term "big data," this book is actually more of a portfolio of work done by the think-tank where the authors work. This is not a bad thing since it provides a critical statistical analysis of recent conflicts in Iraq, Afghanistan, and other places. Throughout this book, the authors take time to not only unpack their methodological practices, but also the practical factors which limit data collection for conflict studies work like obtaining restricted (classified) information and dealing with very dangerous and, often, passionate situations in which to collect objective data in or from.
While I highly recommend this book to anyone doing conflict studies research, it has nothing to do with big data despite its title. A great definition of what big data is follows from the Oracle website (www.oracle.com):
"The definition of big data is data that contains greater variety, arriving in increasing volumes and with more velocity. ... Put simply, big data is larger, more complex data sets, especially from new data sources. These data sets are so voluminous that traditional data processing software just can't manage them."
While this book's authors do find and successfully mine new data sources for analytical insights, this data is far from being so voluminous that traditional software cannot handle it. I applaud the authors for looking outside of their field (conflict studies) for an interesting title, but do not forgive them for misusing this term so blatantly. Despite being recommended by a host of senior leaders and generals whose names are publicly well known, I do not recommend this book to someone unless they are looking specifically for conflict studies analysis techniques, methodological discussions, or to see how such research can be applied to low-intensity, counter insurgency conflicts.
While I highly recommend this book to anyone doing conflict studies research, it has nothing to do with big data despite its title. A great definition of what big data is follows from the Oracle website (www.oracle.com):
"The definition of big data is data that contains greater variety, arriving in increasing volumes and with more velocity. ... Put simply, big data is larger, more complex data sets, especially from new data sources. These data sets are so voluminous that traditional data processing software just can't manage them."
While this book's authors do find and successfully mine new data sources for analytical insights, this data is far from being so voluminous that traditional software cannot handle it. I applaud the authors for looking outside of their field (conflict studies) for an interesting title, but do not forgive them for misusing this term so blatantly. Despite being recommended by a host of senior leaders and generals whose names are publicly well known, I do not recommend this book to someone unless they are looking specifically for conflict studies analysis techniques, methodological discussions, or to see how such research can be applied to low-intensity, counter insurgency conflicts.
June 22, 2024
Well, I finished the book. Overall it did bring up some good points and was a sort of boring read. Not much that was written was particularly new, but overall brought up some good points. The book as mentioned is a bit boring, but I thought it had some good information and helped me understand what to write for future counterinsurgency books, and what to account for, and show what's done right, and what also is and could be done wrong. It was a recommended book from the Irregular Warfare Initiative best reads for the year, and can say it's informative, but kinda dragged out.
August 24, 2022
dnf---how to use data analysis to strategize in symmetrical and asymmetrical armed conflicts. some interesting info, like how in one extended conflict both sides disrupted cell service. so technically, for one side, that decision was wrong.
data analysis might be helpful in many aspects of life... i just didnt care about it's use for battle tactics and strategy
data analysis might be helpful in many aspects of life... i just didnt care about it's use for battle tactics and strategy
August 15, 2018
A solid piece of work on how data analytics can enhance the effectiveness of COIN techniques in the "small wars" of modern conflicts. Helped by use of data from beyond just the Afghan and Iraq/ISIS theaters
December 2, 2018
An another wise fine and insightful book marred by the authors inability to present a clear consistent argument.
August 14, 2019
Remember when we thought technology might be like, anything other than totalitarian hellworld weapons?
September 10, 2021
Basically a large research paper
December 30, 2021
simple hypothesis that easy to understand - should have been 50 pages.
November 30, 2022
Did not like the format or content of this book.
November 6, 2018
A 2012 Forbes article How Target Figured Out A Teen Girl Was Pregnant Before Her Father Did showed the power of data collection. In that article, author Kashmir Hill wrote that every time you go shopping, you share intimate details about your consumption patterns with retailers. And many of those retailers are studying those details to figure out what you like, what you need, and which coupons are most likely to make you happy. Target figured out how to data-mine its way into the womb, to figure out whether a person has a baby on the way long before the parents need to start buying diapers.
The notion of big data has not only revolutionized retail, it has changed the way nations deal with war and conflicts. In Small Wars, Big Data: The Information Revolution in Modern Conflict, authors Eli Berman, Joseph Felter and Jacob Shapiro have written a fascinating book that shows how big data and data analytics can be applied to modern warfare.
In the book, the authors show that by using gathered empirical data from the battlefields and locales, countries can create a new paradigm to deal with localized conflicts and save countless lives. While wars in the past had two large armies fighting against each other, todays conflicts are with small rebel insurgencies. By using data and analytics, the information gleaned can make the difference between a battle won and a battle lost.
It's an old saying that generals always fight the previous war. In Small Wars, Big Data, it shows that many generals have learned that lesson well, and are using the right data to fight and win these new conflicts.
The notion of big data has not only revolutionized retail, it has changed the way nations deal with war and conflicts. In Small Wars, Big Data: The Information Revolution in Modern Conflict, authors Eli Berman, Joseph Felter and Jacob Shapiro have written a fascinating book that shows how big data and data analytics can be applied to modern warfare.
In the book, the authors show that by using gathered empirical data from the battlefields and locales, countries can create a new paradigm to deal with localized conflicts and save countless lives. While wars in the past had two large armies fighting against each other, todays conflicts are with small rebel insurgencies. By using data and analytics, the information gleaned can make the difference between a battle won and a battle lost.
It's an old saying that generals always fight the previous war. In Small Wars, Big Data, it shows that many generals have learned that lesson well, and are using the right data to fight and win these new conflicts.
Displaying 1 - 13 of 13 reviews