Skip to content

The Rational Actor Model

Many social sciences use a model of “rational actors” that completely think through options and make decisions consistent with their preferences.  This model of human decision making is particularly prevalent in the field of economics, where it underlies most of the results and thinking.  However, many other social scientists strongly dislike this model of decision making, arguing that it is has unrealistic expectations on humans and is demonstrably false.   They argue that the rational actor model has humans thinking through infinite streams of “what would he do” arguments, perfect and extensive knowledge about their environment, and otherwise complicated thinking required to be rational.  They also cite numerous studies that are mostly in psychology and even behavioral economics that illustrate cognitive biases and other instances where people act irrationally.
What follows is some of my thoughts about the utility and usefulness of the rational actor model of human decision making.
First, it is valuable to separate two ways that the rational actor model is used: as a normative and as a descriptive model.   As a normative model, the rational actor model illustrates one recommendation for how decisions *should* be made.  Many people generally consider rational decisions to be the “best” or “most correct” decisions because they take into account all of the different factors that influence the decisions and balance them appropriately.  Economists will often point out that irrational decisions open up the opportunity for arbitrage: taking advantage of the decision maker’s irrationality through repeated interactions.
While many people may (at least reluctantly) agree with the rational actor model as a normative model of decision making, many fewer people believe the rational actor model is accurate as a descriptive model.  Indeed, as a description of *how* decisions are made it is almost certainly false.  Most people need to be trained (often through economics courses) to be able to think through what a “rational” actor would do; they do not do so naturally.    It is unreasonable, therefore, to believe that most people actually do think through decisions using this complicated “rational” way of thinking.
However, just because they don’t necessarily *think* rationally doesn’t mean that they don’t *act* rationally.   There are many outside influences on behavior that may have shaped human behavior towards rationality without cognitively being rational.   Indeed, if the rational actor model is a good normative model, it would make sense for humans to end up figuring out ways to act rationally even without thinking rationally.  It is plausible that evolution has shaped a number of human behaviors to be rational.   When you think about it, it makes sense for evolution to shape much human behavior to be as optimal (rational?) as possible without requiring overly-costly thinking; rational behavior without the difficulty of rational thinking could easily be an optimal outcome of evolutionary processes.   Additionally, there are many equilibrium process — arbitrage, competition, etc. — that further shape behavior towards rationality.  Indeed, many behaviors are well-approximated by the rational actor model, including but not limited to many market behaviors such as price/quantity predictions, macro-scale economies, etc..
The fields of psychology and behavioral economics has done an excellent job document a wide variety of *cognitive biases*: patterns in human decision making that consistent deviate from the rational actor model.    It is true that the rational actor model is not a perfect theory of human behavior; however it does apply in many situations.  I think a good analog is Newton’s theory of motion.  Newton theory of motion isn’t a description of how objects decide where to go; it is simply a description of the result behavior.   Planets don’t use Newton’s equations to decide where to go next, but Newton’s equations are still effective at predicting where the planets will be tomorrow nevertheless.  This is similar to the rational actor model: the RAM doesn’t necessarily say how humans come to their decisions, but it does accurately predict what those decisions will be.
Newton’s theory of motion is also a good analog for another reason: it too is imperfect.   There are many things that Newton’s theory gets wrong, particularly at the extremes when objects are moving very fast (relativity) or are very small (quantum mechanics).  But just because it is imperfect doesn’t make it useless; it is still a very useful theory for predicting the behavior of objects in motion in many circumstances, and is still taught in school for that very reason.   For many years (centuries?) there were acknowledged flaws in Netwonian mechanics, but still it was the dominant theory because it was useful and no better theory existed.
I believe the rational actor model is very similar.  Yes, there are numerous acknowledged flaws in the theory; people do not always behavior rationally.   However, right now it is still the best and most comprehensive model of human decision making that we have.  Many scientists are striving to figure out just when and were the model works well, and when and where it should not be used.   I suspect that at some point there will be a Kuhn-ian paradigm shift where a new model of decision making will supplant the RAM as the primary / best model of human decision making, must like relativity and quantum mechanisms supplanted newtonian mechanics.  But until this new theory emerges, the rational actor model is still the best model.
(And now this is when this essay becomes particularly defensive.)  There are many critics of the rational actor model.  Indeed, I have had entire papers rejected simply because I used the rational actor model, which the reviewers believes was “wrong”.  Yes, the rational actor model has flaws, but so does any theory.   What most critics fail to do is to argue that either a) there exists a better / more useful theory of human decision making that should be used instead, or that b) not having a theory, and not making predictions, is better than using the rational actor model.  Obviously, if the predictions are completely wrong, then b) is true.   The “Rick Wash” theory of decision making (where I assume that everyone would make the same decision that I think I would make in that situation) is an incorrect theory of behavior, and probably leads to so many incorrect predictions that not having a theory is better than using the “Rick Wash” theory.  However, the rational actor model has proven to be quite useful in many situations.  It has been immensely useful in understanding the behavior of markets, in understanding political relationships and behavior between and within nation-states, and in understanding behavior within organizations.  Much of the advise used to make major decisions for our economy is based on predictions from the rational actor model.
As a theory of human decision making, the rational actor model is certainly not perfect.  It does not explain *how* decisions are made, and it doesn’t accurately describe all human decisions.  However, it is a very useful model of behavior because it provides strong and simple guidance on how to think through decisions in a way that provides reasonable and relatively accurate predictions about complicated human behaviors.

Many social sciences use a model of “rational actors” that completely think through options and make decisions consistent with their preferences.  This model of human decision making is particularly prevalent in the field of economics, where it underlies most of the results and thinking.  However, many other social scientists strongly dislike this model of decision making, arguing that it is has unrealistic expectations on humans and is demonstrably false.   They argue that the rational actor model requires humans to think through infinite streams of “what would he do” arguments, have perfect and extensive knowledge about their environment, and otherwise use complicated thinking patterns.  They also cite numerous studies that are mostly in psychology and even behavioral economics that illustrate cognitive biases and other instances where people act irrationally.

What follows is some of my thoughts about the utility and usefulness of the rational actor model of human decision making.

First, it is valuable to separate two ways that he rational actor model is used: as a normative and as a descriptive model.   As a normative model, the rational actor model illustrates one recommendation for how decisions should be made.  Many people generally consider rational decisions to be the “best” or “most correct” decisions because they take into account all of the different factors that influence the decisions and balance them appropriately.  Economists will often point out that irrational decisions open up the opportunity for arbitrage: taking advantage of the decision maker’s irrationality through repeated interactions.

While many people may (at least reluctantly) agree with the rational actor model as a normative model of decision making, many fewer people believe the rational actor model is accurate as a descriptive model.  Indeed, as a description of how decisions are made it is almost certainly false.  Most people need to be trained (often through economics courses) to be able to think through what a “rational” actor would do; they do not do so naturally.    It is unreasonable, therefore, to believe that most people actually do think through decisions using this complicated “rational” way of thinking.

However, just because they don’t necessarily think rationally doesn’t mean that they don’t act rationally.   There are many outside influences on behavior that may have shaped human behavior towards rationality without cognitively being rational.   Indeed, if the rational actor model is a good normative model, it would make sense for humans to end up figuring out ways to act rationally even without thinking rationally.  It is plausible that evolution has shaped a number of human behaviors to be rational.   When you think about it, it makes sense for evolution to shape much human behavior to be as optimal (rational?) as possible without requiring overly-costly thinking; rational behavior without the difficulty of rational thinking could easily be an optimal outcome of evolutionary processes.   Additionally, there are many equilibrium process — arbitrage, competition, etc. — that further shape behavior towards rationality.  Indeed, many behaviors are well-approximated by the rational actor model, including but not limited to many market behaviors such as price/quantity predictions, macro-scale economies, etc..

The fields of psychology and behavioral economics has done an excellent job document a wide variety of cognitive biases: patterns in human decision making that consistent deviate from the rational actor model.    It is true that the rational actor model is not a perfect theory of human behavior; however it does apply in many situations.  I think a good analog is Newton’s theory of motion.  Newton theory of motion isn’t a description of how objects decide where to go; it is simply a description of the result behavior.   Planets don’t use Newton’s equations to decide where to go next, but Newton’s equations are still effective at predicting where the planets will be tomorrow nevertheless.  This is similar to the rational actor model: the RAM doesn’t necessarily say how humans come to their decisions, but it does accurately predict what those decisions will be.

Newton’s theory of motion is also a good analog for another reason: it too is imperfect.   There are many things that Newton’s theory gets wrong, particularly at the extremes when objects are moving very fast (relativity) or are very small (quantum mechanics).  But just because it is imperfect doesn’t make it useless; it is still a very useful theory for predicting the behavior of objects in motion in many circumstances, and is still taught in school for that very reason.   For many years (centuries?) there were acknowledged flaws in Netwonian mechanics, but still it was the dominant theory because it was useful and no better theory existed.

I believe the rational actor model is very similar.  Yes, there are numerous acknowledged flaws in the theory; people do not always behavior rationally.   However, right now it is still the best and most comprehensive model of human decision making that we have.  Many scientists are striving to figure out just when and were the model works well, and when and where it should not be used.   I suspect that at some point there will be a Kuhn-ian paradigm shift where a new model of decision making will supplant the RAM as the primary / best model of human decision making, must like relativity and quantum mechanisms supplanted newtonian mechanics.  But until this new theory emerges, the rational actor model is still the best model.

(And now this is when this essay becomes particularly defensive.)  There are many critics of the rational actor model.  Indeed, I have had entire papers rejected simply because I used the rational actor model, which the reviewers believed was “wrong”.  Yes, the rational actor model has flaws, but so does any theory.   What most critics fail to do is to argue that either a) there exists a better / more useful theory of human decision making that should be used instead, or that b) not having a theory, and not making predictions, is better than using the rational actor model.  Obviously, if the predictions are completely wrong, then b) is true.   The “Rick Wash” theory of decision making (where I assume that everyone would make the same decision that I think I would make in that situation) is an incorrect theory of behavior, and probably leads to so many incorrect predictions that not having a theory is better than using the “Rick Wash” theory.  However, the rational actor model has proven to be quite useful in many situations.  It has been immensely useful in understanding the behavior of markets, in understanding political relationships and behavior between and within nation-states, and in understanding behavior within organizations.  Much of the advise used to make major decisions for our economy is based on predictions from the rational actor model.

As a theory of human decision making, the rational actor model is certainly not perfect.  It does not explain how decisions are made, and it doesn’t accurately describe all human decisions.  However, it is a very useful model of behavior because it provides strong and simple guidance on how to think through decisions in a way that provides reasonable and relatively accurate predictions about complicated human behaviors.

2009 12 21 rwash Methodology Notes Comments (0) Permalink

Types of Bootstrapping

I’ve written before about the bootstrapping problem for social media systems: if people are coming to the site because of the content, how do you draw the first users in?  Bootstrapping social media is a classic instance of a positive network effects problem: how do I get enough people to start using my system such that the system provides enough value to everyone to be self-sustaining?

When thinking about bootstrapping, there are really two related but separate issues.   First, there is the problem that a newly created social media site has very little content to draw users in.  This is the site bootstrapping problem: how do you get users to visit and contribute to your social media site when there isn’t enough activity yet from other users to create a community and sustain interest.  A second and closely related problem is related to the difficult that users face when they first visit a website; what kind of contributions are appropriate, and are they feasible?   This is the new user bootstrapping problem: how does a user learn how to contribute to this social media site; what contributions are appropriate?  Note that these problems are strongly related.   A new user that visits a social media site during early site bootstrapping also faces a new user bootstrapping issue.  However, as content is contributed, the site bootstrapping problem begins to go away but the new user bootstrapping problem remains.

I think the new user bootstrapping problem really has at least two pieces of information that any new user must learn before they can contribute.   First, the new user must understand what type of contributions are appropriate.  For example, political rants are rarely considered appropriate when reviewing kitchen appliances on Amazon.com.   There are many different types of contribution, and the new user has to figure out what type should be contributed here.   Appropriateness is a tricky concept because it can decided in multiple ways.   Often the site designer has an idea of what type of content is appropriate: product reviews on Amazon.com for example.  However, there is often a community idea of appropriateness also.  The other users on the site have their own opinions on what is appropriate, and can enforce those by using various voting mechansims, or writing meta-comments.   The fact that appropriateness is often ill-defined and comes from multiple sources makes figuring out what types of contribution are appropriate difficult.

Second, once users understand what information is appropriate, users must also think about feasibility: am I really able to contribute the type of information that is appropriate.  The Onion uses parody to illustrate this concept quite clearly with a story about a site called “Noveller” where people can “macroblog” full-length novels.  Such contributions are clearly infeasible for most, if not all, users.  However, for more real-world sites, feasibility isn’t always clear.  How feasible is it to write a product review for Amazon.com? It depends both on who is writing, and how detailed of a review they want to write. Feasibility is really about cost management; how can I exert the effort to contribute without spending too much time or effort on it?  Users need to figure out for themselves how much effort to put in, and how to make sure that that amount of effort produces an appropriate contribution.

2009 12 18 rwash Big Thoughts Comments (0) Permalink

Social Norms as Motivations

I recently attended a talk by the social psychologist Robert Cialdini.  He specializes in “influence” — how to get people to do things.   It shouldn’t surprise anyone to hear he is actually a professor of Marketing at Arizona State University.  He has spent his career studying how to influence people, and has even written multiple books on the subject.  In the talk I attended, he focused on one specific, and in his opinion very powerful, social influence: social norms.

Social norms are societal cues that create behavioral expectations.  When you see a social norm, you feel like you should be doing that action too.  There are two types of social norms that have been well-known for a long time: descriptive norms and injunctive norms.  Descriptive norms are engaged when a person believes many others like them are engaging in a behavior.  While there is no explicit motivation to “join the group”, many people still feel compelled to engage in that behavior also.   Injunctive norms, on the other hand, reflect community standards of approval and disapproval.  Injunctive norms have obvious paths of influence; if your community disapproves of some behavior, that discourages you from engaging in that behavior.  However, Cialdini argues that descriptive norms may actually be a more powerful mechanism for influence because they are underdetected.  (Nolan et al. 2008)

Cialdini did a great job of illustrating how little information is needed to engage a social norm to influence people’s behavior.  In one of this studies, he swapped out the standard “please reuse these towel” messages in hotel rooms.  The standard “do it for the environment” message was replaced with a short piece of text explaining that “the majority of guests reuse their towels.”  This simple change, he argues, engages a descriptive social norm; people read this and believe that many other people reuse towels, and they should too.  This change caused a 25% increase in the number of hotel guests to reuse towels.  (The rate increased from 35.1% to 44.1% of guests.) However, remember that the descriptive social norm is engaged when a person believes that many others like them are engaging in a behavior.  So he had yet another condition; the hang tag said that “the majority of guests in this room” reuse towels.  This increased compliance even further; that condition had 49.3% of hotel guests reusing towels! (Goldstein et al. 2008)

There are a number of recent research papers by Cialdini and colleagues that illustrate how powerful descriptive social norms can be in influencing behavior.  Cialdini espoused an interesting theory as to why descriptive social norms are so powerful: in addition to expressing a social expectation for behavior (“you should do this”), they also express feasibility (“you can do this”).  I suspect that this feasibility component is really powerful, and it is something that most other incentives does not have.  It is easy to provide incentives for a behavior that is actually infeasible to do; a descriptive social norm overcomes this problem by saying “other people like you have done this; you can too.”

References

Goldstein, N. J., Cialdini, R. B., & Griskevicius, V.. A room with a viewpoint: Using normative appeals to motivate environmental conservation in a hotel setting. Journal of Consumer Research (2008)

Nolan, J. M., Schultz, P. W., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V.. Normative social influence is underdetected.. Personality and Social Psychology Bulletin (2008).

2009 11 23 rwash Research Pointers Comments (0) Permalink

Outcomes of Social Media

A lot of my research can be seen as look at the effects of various types of technical interventions on social media systems.   I change something about a social media system, and I look to see how people act differently.  There are many different ways of trying to predict what will happen, but in this post I wanted to think through what are the possible types of outcomes; what changes can happen as a result of these interventions?

Most of the time I (and other researchers like me) look at individual-level outcomes within the site.   Do individual users increase their contributions to the site?   Do individual users participate more by leaving comments or rating items?  Most of the outcomes I identified in a previous post fit in this category.  This is the easiest category of outcomes to study, since within-site outcomes are easy to measure, and there are usually lots of individuals using the site.

A second type of outcome is group-level or system-level outcomes within the site.  These are outcomes that concern group/system level properties.  For example, we could ask the question “is there a sufficient number of users contributing for the site to be self-sustaining?” or “even though some users increase and others decrease their contributions, is the total quantity of information on the site increasing?”   My recent paper on minimum thresholds looks at this second question.   These questions are different because they look at the site as more than simply an aggregate of individuals; they assume that there exist group-level properties of interest.  Often looking at these outcomes involve looking at tradeoffs across individuals, such as the “some people decrease contributions but others increase” tradeoff in the minumum threshold research.   These types of outcomes are hard to study because each social media system is N=1; the system can only have a single outcome.  I have obviously done some work on these types of outcomes, and I intend to do more in the future.  I think economic modeling is actually a powerful tool for this because it can look at group-level properties and tradeoffs.

A third type of outcome is the individual-level spillover outcome.  Often people end up changing other parts of their life as a result of their use of a social media system.   The field of communications calls these “media effects.”  For example, people can feel more connected and have more friends because of their Facebook use.  Spillover effects don’t necessarily exist, but they can be an excellent reason to study social media systems when they do exist.  However, it is generally quite difficult to study spillover effects; simply having access to data from the social media system is not sufficient to study spillover effects.  You must also have data from the individual users about life outside of the social media system.

Finallly, the fourth type of outcome is a group-level spillover outcome.  These are outcomes that result to larger social structures as a result of social media use by the individuals in the group.  For example, a number of business have adopted various social media systems (IBM has gone full out, producing their own social networking system Beehive, their own social bookmarking system Dogear, and their own microblogging system).  These businesses hope that these systems have enterprise-level spillover effects.  (They also hope these effects will be generally positive for the business, like making people more productive, or encouraging innovation through collaboration.)  Unfortunately, these are the hardest types of effects to measure because they have all the difficulties of both group-level effets and spillover effects. Often research into these effects are case studies that follow how one particular system has impacted a given organization.

When thinking through the various ways that social media systems can have an impact, this taxonomy of effects might be useful.

2009 11 20 rwash Big Thoughts
Methodology Notes Comments (0) Permalink

Keeping People Coming Back to Farmville

Farmville is a very popular game that can be found on Facebook.  In this game, you are given a plot of land on which you can purchase and plant all kinds of crops; the more crops you plant and harvest the more types of crops (and other goodies) you can purchase for later. In the last month, Farmville has had over 62 million active users.  If Farmville were a country, it would be the 22nd most populous country, falling between France and the United Kingdom on the list. Social games like Farmville are instances of the kinds of social computing / social media systems that I like to study.  When I see something as successful as Farmville, I start wondering what I can learn from it that applies more broadly to other types social media systems.

One of the really interesting things about Farmville is that it has some fairly strong incentives for people to keep coming back and playing Farmville.  Put in more generic terms, Farmville does a great job at providing incentives for user retention.  Here’s the basics of how it works: you buy seeds and plant them in a plot on your form.   Then you have to wait; it takes anywhere from 4 hours (real, wall-clock time) to 3 days or more for the seeds to grow.  (Note: I’m only a level 5 farmer; I’m sure there are much fancier things you can grow as you advance.)  This waiting means that you can’t just harvest or keep playing right now; with the shortest growing time at 4 hours you also cannot just wait it out.   Once the seeds have finished growing, you can come and harvest them for coins and experience points.  However, if you wait too long to harvest your crops, then they whither and die and you lose your money.

It is really interesting to hear stories about how people adjust to this structure.   They plan a lot: people plan to be at their computer at certain times so they can harvest their crops; they plan which crops to plant by thinking about when they will be available to harvest them, and they stick to their plans because there are penalties for not doing so.  Basically, this structure provides a strong incentive for people to plan out how to fit this game, this social media system, into their life on a regular basis.

This is exactly the goal of user retention: we want to encourage users of a social media system to keep coming back on a regular basis, and to make visiting the site part of their daily life.  When you combine this with their viral marketing incentives (to buy certain things you need to convince a number of your friends to become your neighbors in the game), it is not surprising that this game has 62 million users.

I have a guess (a hypothesis, if you will) about  how we can learn from Farmville in designing other types of social media.Farmville schedules times for things to happen.   At 5pm today, my strawberries finish growing and I need to return to Farmville to harvest them.  This works for two reasons: 1) it gives me a reason to come back to the site later.   This is an individualized reason; they are my strawberries that need to be harvested, and only I can harvest them.  And 2) it gives me a reason to wait.  I cannot harvest the strawberries now.  I cannot accomplish everything I want to accomplish on the site right now.  I have to wait, be patient, and come back for more later.  This second reason is counter-intuitive: by making the site not serve all of my needs, at least not right now, I have a reason to return later.  And to keep coming back.  Less functionality leads to more use.  Farmville retains users by not meeting all their immediate needs, but scheduling a time in the future when they can return and get those needs met.

However, they add one additional twist.  They add a deadline; if you don’t come back by a certain time, your crops have withered.  You can’t procrastinate your return; you have to return relatively soon if you want to actually get the rest of your needs met.   This is the real key incentive that induces people to plan around Farmville.   Without this, you could just play farmville in your free time, and as people get busy, free time disappears.   But since your crops wither, you can’t just wait till you have free time.  You have to go harvest them now (or within 4 hours).  You have to plan a time in your day to tend your crops.  You have to force time into your busy schedule to visit the site.  You have to integrate Farmville into your life.  And once a person has integrated a social media site into their daily life and their daily routines, then the site really has retained the user.

This strategy of waiting with a deadline has been used by other types of social media.  For example, Facebook uses a very similar strategy to retain users.  You can’t get all your needs met right now; you have to wait for more people to contribute status messages in the future.   By forcing users to wait for the status messages, Facebook gets people to come back later.  Many sites accomplish this by having regularly updating content; this is a well-known important feature of most social media systems.   However, Facebook also uses the deadline approach.  Eventually, new status messages scroll off the screen.  If you wait too long to check Facebook, then you miss status messages from your friends.   And its not very easy to just check the friends you really care about; most people use the “all status messages” live news feed.  Since messages basically expire by falling off the bottom of the page, users feel like they have a deadline for checking Facebook.  This deadline provided the incentive that people needed to integrate Facebook into their daily lives.

Also, this incentive has the interesting property that it is the strongest for people who are the busiest.  Normally, it is the busiest people who can’t afford to fit a social media system into their lives.  However, the busiest people are often the ones with the most “friends” on Facebook.   And that means that status messages fall off the page faster as new messages from friends appear.   The busiest people are the ones with the shortest deadline, and therefore the strongest incentive to integrate Facebook into their lives.

2009 11 13 rwash Real World Observations Comments (0) Permalink

Site Governance

As social computing systems grow, it becomes increasingly important to institute some sort of site management or governance.   Site governance work is important to help keep the site working smoothly as many new people with many new motives and agendas use the site.   Wikipedia is famous for making public its large quantity of governance work: deleting inappropriate articles, reverting unwanted changes, setting site-wide policies, rewarding users for contribution, etc.  All large sites need governance of some kind, though not all sites make it as public as Wikipedia.  Facebook, for example, has a team of employees that work on eliminating unwanted people / pages, developing the site, soliciting partners, etc.

Once a site grows large enough, it needs a non-trivial amount of governance work.  One way to get this done is to form a company and pay employees to do the governance.  Many sites try to automate some of the governance work, such as cleaning up unwanted spam.  Finally, many sites look to their users and try to encourage their users to voluntarily do some of the governance work.

And, to me, this opens up an interesting opportunity.  How can we design social media systems that encourage at least some users to perform the necessary governance work?

Slashdot has an interesting model in this area.   They use part-corporate (the editors are all professional, paid employees) and part-user governance.  Users of Slashdot moderate and filter content in the discussions, and then meta-moderate others to ensure a consistent, valuable moderation system.  Slashdot has a complicated procedure for assigning rights to moderate based on a number of criteria, including how valuable your contributions have been in the past, and how much meta-moderation you have engaged in.  In this way, the ability to moderate, or exercise control, is used as a reward for other types of less interesting governance.  Governance that encourages governance!

Tagged
2009 10 23 rwash Big Thoughts Comments (0) Permalink

The Power of the Ask

I recently got an iPhone, and I don’t know how I lived without it.  For the not-so-brief period of time that I didn’t have Internet at home, it was my lifeline connecting me to my Email and my distant wife.  One of the fun things about the iPhone is the App Store.  You can browse through and download applications from a collection of thousands of applications.  When browsing an application in the App Store, each application displays a distribution of ratings on a scale of one to five stars.  This really helps evaluate applications; if lots of people didn’t like it, then I usually won’t download it.

One of the things that really surprised me when I got my iPhone is the large number of ratings.  Most applications that I looked at had hundreds or thousands of ratings.  That’s as much or more than the number of ratings that products on Amazon.com get.  Why would so many people spend their time rating applications they have used?

It turns out that the way most of the ratings are collected is by prompting users for a rating when they delete an application.  I guess that Apple figures people will have used an application enough to have an opinion about it if they are deleting it.

This reminded me a lot of an “incentive mechanism” that is well known in the field of Philanthropy: the “Power of the Ask.” (Anderoni, 2006)  Basically, when you want someone to give money to your charitable organization, one of the most effective methods is to approach them and ask them to donate.  Outright asking works for a couple of reasons: 1) it changes the question from “which charity should I donate to” to “should I donate to this charity.”; this helps the charity doing the asking at the potential expense of other charities.  2) It solves the problem of when to make the donation; people cannot procrastinate or put off donation when being asked to do so now.

Thinking a little deeper, though, the mechanism used by Apple’s App Store is subtly different. The main difference is that Apple is asking for a contribution of information rather than money.  All information isn’t the same; there is good information and bad information.  Or, more accurately, there is useful information and biased / wrong information.  When asking for money, a charity doesn’t have to worry about getting bad money; if they get any money at all it is good.   However, Apple needs to worry that people will provide bad information.   Asking isn’t enough; you need to somehow ensure that the information is high quality and useful.

I think Apple might have done this wrong.  Specifically, they only ask for a rating when a user is deleting an application.  So all the unsatisfied users who delete an application rate it low.  All the satisfied users who don’t delete it and keep using it never get asked, and never contribute a rating.   This means that the overall ratings are biased to be much lower than the community truly thinks they are.  This illustrates one reason why “the power of the ask” doesn’t always for in information settings in the way that it does in charity / money settings.

Unfortunately, solving this problem isn’t easy.  You could randomly prompt people for ratings, or ask them N days/weeks/months after they install an application, but that would be very annoying for the users.  If you don’t ask, then your ratings will be more unbiased, but very few people will contribute.

J. Andreoni. Philanthropy. In S.-C. Kolm and J. M. Ythier, editors, Handbook of Giving, Reciprocity and Altruism, pages 1201–1269. North Holland, Amsterdam, 2006.

2009 10 19 rwash Real World Observations Comments (0) Permalink

Designing Incentives

Designing social computing systems / social media systems is difficult.  This is at least partially because our existing design strategies don’t work very well for the extremely-social systems like Facebook or delicious.com.  The basic problem comes from trying to straightforwardly apply user-centered design.  You certainly can put a user in a lab and watch them use these social systems.  But that won’t actually help you understand their design much.  At best, it can suggest a few minor tweaks.   But the problem is that users don’t use these systems by themselves; they use these systems are part of a (sometimes rather large) group.  And many (most) of the actions of a user are reactions to things other users have said or done.  Basically, what this means is that when designing these social systems, you can’t look at users in isolation.  Treating users as if they were independent doesn’t work.

But it is actually more complicated than that.  These systems frequently don’t work unless users behave in specific ways.  Delicious doesn’t work as a social system unless people bookmark pages of interest.  Digg depends on users voting on stories in a consistent and coherent way.  Wikipedia depends on logical contributions and collaboration across editors.  Designing technologies like this is more than just making it easy to do this required behavior; it is important for the rest of the system that users behave correctly.

Incentive-centered design is a different way of thinking about designing these systems that is being developed at the University of Michigan.  First, the designer figures out what behaviors are necessary and important for the system to function properly.   And then second, the designer designs a technology that induces (or encourages) the appropriate behaviors from its users.  It is different from user-centered design because it focuses primarily on behavior, and how the technology influences how users choose to behave.  It also focuses on positively influencing behavior: it tries to encourage good behaviors rather than trying to eliminate difficulties and problems.

It is important to remember that both parts of the design philosophy are important.  It is not always easy to know what behaviors are appropriate.  Tools like game theory are vital to understanding how users react to each other, and how users’ actions aggregate to form system-level properties.   And once the designer understand how he or she wants users to behave, getting users to behave in that way is difficult.   Technology can only do so much in influencing how users treat the system, and we don’t have much research into appropriate designs for encouraging specific behaviors.

Separating the design process into these two steps also provides guidance for my research agenda.  Both steps in the process need research support.

Tagged
2009 08 28 rwash Big Thoughts Comments (0) Permalink

Boom Blox

I haven’t been updating this blog much in the last couple of weeks because I’ve been doing the final writing push for my dissertation.   After spending 12+ hours a day writing the dissertation, it is hard to sit down and write more for a blog.  But now that is done, and Emilee and I took the whole weekend off of work.  And we spent most of the weekend playing the new Boom Blox Bash Party game for the Nintendo Wii.

One of the interesting things about this new game is that it supports user-contributed content: anyone can use the built-in level creator to create new levels and upload them to a central content server.  You can also download and play levels that others have created.  As expected, I found a few really fun levels and a bunch of not-so-good levels.  Fortunately, the Boom Blox online system allows users to rate levels with 1-5 stars, and you can sort the levels by rating when looking for new levels.

This is a great example of how this user-contributed content is becoming pervasive.  Even games like Boom Blox have a user-contributed content portion of the game.  And I think that this feature could have benefited from hiring someone who studies incentive-centered design.  There is a basic contribution question here: why should users contribute levels?  And more importantly, what types of levels should users contribute?  So far, the creators of the game have been rewarding complex Rube Goldberg type levels by highlighting them on YouTube.  It would be interesting to think about different ways that the game can reward other types of levels; for example, it could pick the highest rated levels of each type currently in the game and make them available as a downloadable “level pack.”  This might encourage users to create interesting new levels to extend the game’s current gameplay.

Another ICD issue is the rating system.  Right now, after playing an online level, the user is asked to rate it with 1-5 stars.   But, why should the user provide a rating?   And, more importantly, what metric should he or she employ?  A level might get 5 stars on “cool” but 1 on “playability”.  By thinking about what behavior you want people to do, you can design the system to elicit that behavior.  For example, you might want people to rate levels based on how much fun they are to play over and over. Do elicit that type of rating, the designers might allow people to save levels from online (which is currently supported) and automatically sort them by the rating I gave it by default.   Levels I give 5 stars are shown first, then 4 stars and on down.  This way, levels I have an incentive to rate highly the levels I want to play again (to make them easy to find), and rate poorly the levels I don’t.

This is actually an example of the side effect mechanism I’ve talked about before.  Users have a private reason to rate levels — to make it easier to find the ones they want to play again.   And the incentive to rate levels is aligned with the goals of the consumers, who want to use the ratings to know what levels are best to play.

Tagged
2009 08 17 rwash Real World Observations Comments (0) Permalink

Facebook uses a Minimum Threshold

I recently blogged about minimum threshold mechanisms: set a minimum threshold for contribution and exclude anyone from using the system who doesn’t meet that threshold.  I recently encountered a great example of a popular social media system using a minimum threshold to encourage contribution: Facebook.

In order to use Facebook, a new user must do three things: 1) sign up for an account, 2) join a network, and 3) contribute a list of her friends (her social network).  Users who don’t do all of these don’t receive access to any of the information on Facebook.  Notice #3 (and to some extent #2); it includes contributing information to the system that benefits others.   Facebook doesn’t have lurkers in the classic sense of someone who doesn’t contribute anything.  They decided that it was worthwhile to lose the true lurkers in exchange for the increased contributions.   By knowing everyone’s social network, Facebook can offer additional services (like strong privacy controls and better friend suggestions) that they wouldn’t be able to offer if the social network information was less complete.   This is a great example of the minimum threshold in action; you gain a little bit of information from everyone at the expense of losing the lurkers.   For Facebook, they figured out a “little bit” of information that allows them to really make their system valuable.

If you are skeptical, compare Facebook with Twitter.   Twitter does not have a minimum threshold of contribution.   There are true lurkers on Twitter; many people read twitter without accounts.  These people have never contributed to twitter but still use it.  Twitter values these lurkers and as such does not require a minimum contribution like Facebook does.

And sure enough, a number of the predictions of my minimum threshold paper have come true on Facebook.  A number of people have chosen to join Facebook, but only contribute the minimum.  I’ve seen a number of users who are my “friends” who never contribute pictures or status messages; all they do is contribute their social network information so they can get access to the information of others.   Also, as Facebook has gotten larger, more people who would have preferred to lurk have decided to contribute the minimum threshold.   Larger systems provide more value, and hence more people are willing to make that minimum contribution in order to join.

The Facebook example also illustrates a useful point for using the minimum threshold mechanism: everyone who remains in the system will have contributed at least the minimum.  If the threshold is chosen carefully, the system can take advantage of the knowledge that everyone has contributed this information.   On Facebook, all users have to contribute social network information.  Facebook is then designed to take advantage of the fact that all Facebook users have provided their social network information to the system; many features such as the privacy controls assume the existence of contributed social network information.  This well-chosen threshold enables many useful features.

Tagged
2009 07 20 rwash Real World Observations Comments (0) Permalink