More on Netflix

by Henry Farrell on August 2, 2004

“Hunt Stilwell”: has let me know via email that the “Netflix fallacy”: that I talked about last week seems to replicate a very interesting experiment on the psychology of intertemporal decision-making. His email (with permission, and some light editing) is reproduced under the fold.

bq. I just read your post on Crooked Timber on the “Netflix Fallacy.” I had a similar experience with Netflix, which led me to get rid of it. I think that what your experience and mine represents is not a new “economic fallacy,” but an old one. In fact, I think our experience does a pretty good job of replicating an experiment by George Lowenstein and colleagues on intertemporal decision making. In that experiment (a link to the paper is below), participants were allowed to choose films, some of which were “vice” films (e.g., “Armageddon” or “The Mask”), with others being “virtue” films (e.g., “Schindler’s List” or one of your foriegn films with subtitles). Participants made the choices either in a sequence, meaning that the viewings would be in the immediate future, or made them all simultaneously, meaning that some of the viewings would be delayed. Participants who made the choices in a sequence tended to pick mostly vice films, while participants who made the choices simultaneously picked many more virtue films.

bq. It looks to me like our behavior was an instance of a psychological tendency that researchers studying intertemporal decision making have observed over and over again: when making decisions for the immediate future, people tend to pick vices (low immediate reward, high long-term cost), whereas people who are making decisions for the more distant future pick virtues (low immediate cost, high long-term reward). We selected “virtue” movies because we knew our viewing would be delayed, and therefore were more likely to pick alternatives with low immediate cost and (potentially) high long-term rewards. We even picked them simultaneously (3 or three films at a time)! as in the experiment. If we had gone to Blockbuster, we would have known that we would have to watch the films in the relatively immediate future (24-72 hours, depending on whether the film was a new release or not), and thus we would be more likely to pick “vice” films that we might be motivated (by the easy entertainment value) to watch.

bq. I’ve always had a problem with the “virtue/vice” distinction applied to films in the Lowenstein paper, because there really isn’t much of a difference between the cost associated with the two types of films. I suppose the wasted time makes for a cost difference, but I’m not sure that’s enough. However, the pattern is the same. The fact that Netflix demonstrates this so well, and even replicates an actual experiment, lends me to believe that Netflix is really just a big social psychology experiment, and those of us who’ve subscribed are all unwitting participants.

The paper on the movie experiment can be found here:



Keven Lofty 08.02.04 at 4:26 am

My wife and I do this EVERY TIME. We used to be members of Netflix (and quit for the reasons above). Now however when we rent videos or DVDs we do so from Hollywood Video or Blockbuster. We usually get two movies, a vice and virtue movie. Last time it was Bubba-Ho-Tep and 21 Grams. We watched the ‘vice’ movie first (like we always do) and this time we got round to watching the ‘virtue’ movie before it was due back. Many times we don’t. Incidently the time before that we rented Before Sunrise and Lost In Translation. We watched Before Sunrise first because we had already seen it before and it had become the ‘vice’ choice because it was the easy choice.


eudoxis 08.02.04 at 6:22 am

Very interesting. I don’t know how the Netflix phenomenon fits either the simultaneous or the sequential categories in the paper. It appears, at first glance, that the Netflix choice is a simultaneous one with delayed viewing. However, even when movies are shipped simultaneously, movie selection is really a lot more interesting than the experiments in the paper provide for because the decision to view movies is adaptive at the selection end and flexible at the viewing point. The movies in the online queue represent a collection that is formed by several decision making options. The top movie in the queue approaches a sequential decision type when a lowbrow movie is impulsively popped into slot #1 right before shipping. At other times, the top movies in the queue bubble up through a series of simulataneous decisions with changes along the way. Even though there is a shipping delay constant, the real delay happens at the queue where must-watch-depressing movies are pushed to the bottom and may never reach the point of shipping. Once shipped, the number of movies out at one time presents a new selection. This time, another sequential choice is made. Movies with over predicted return value languish on top of the DVD player.

Together with the previous post on this subject, I am given the impression that a lot of people don’t make use of two Netflix features that critically differentiate it from traditional rental stores: reordering the queue and returning movies that are not watched. The Netflix scheme, then, allows more extreme “violation of invariance” and a way to indulge in plenty of lowbrow movies.


bad Jim 08.02.04 at 9:42 am

Another former Netflix subscriber who quit because of the inevitable virtue build-up. Haven’t yet gotten around to viewing “Rabbit-proof fence.” Odd pickings, though. Vivaldi’s “Orlando Furioso” was more strangely memorable than Prokofiev’s “Love of Three Oranges.” I actually bought a copy of “Lady Macbeth” after renting it. “Romper Stomper” was too violent, but “Quills” wasn’t.

“Carrington” led me to read Eminent Victorians and The Guns of August, both of which were excellent, of course. This is where Angels and Insects goes, I think, and perhaps also “The Affair of the Necklace”, which I probably would never have seen otherwise (though I now think I’d watch anything with Hilary Swank, including The Core).


Dave 08.02.04 at 6:35 pm

E. Haldeman-Julius managed to do well while doing good by taking advantage of this effect; he offered his little blue books at 5 cents each, with a minimum order of $1.

Not only did the minimum order of 20 encourage getting a few virtue books to pad out a vice order, but he was not above changing the title on virtue books to sell them in vice clothing.


netflixfan 08.02.04 at 7:48 pm

The fact that Netflix demonstrates this so well, and even replicates an actual experiment, lends me to believe that Netflix is really just a big social psychology experiment, and those of us who’ve subscribed are all unwitting participants.

Of course if you replace “Netflix” with “Blockbuster” and “who’ve subscribed” with “who are members” this statement would be just as valid.


David Edelstein 08.02.04 at 7:58 pm

The outcome of this “experiment” is depressing but not surprising and has confirmed the feeling that many of us have that our living rooms are not necessarily the places to have great motion-picture experiences. The language of TV involves a lot of close-ups and a very punchy kind of pacing, whereas most of the art films you will rent have longer and lengthier shots. Throw in subtitles and the prospect seems wearying. I have often rejected b&w subtitled classics for bloody shlock at the end of a long day, to my enduring regret.


Ted 08.02.04 at 11:50 pm

Suggested solution: turn over control of the Netflix queue to the household member who cares the most about it. Our 8-year-old daughter is in charge, and we churn through the disks quite fast. Every third time or so she lets us get a grownup movie in.
I like Netflix as a screening tool. We’ve bought a few movies after renting them from Netflix, notably Yellow Submarine (awesome for all of us, each for our own reasons) and James and the Giant Peach (ditto).


Hanah Metchis 08.03.04 at 12:51 am

I’ve been a Netflix subscriber for over a year, and I’m totally devoted to it. I have made several interesting observations about the way my moviewatching habits have been changed by Netflix, including:

1. I watch way more movies than I used to. They’re right there in my living room, so why not?

2. I watch way, way more “virtue” movies than I used to. For me, these tend to be classics and foreign films.

3. The “virtue” movies do sit in my apartment for longer than the “fun” movies, but 4 weeks would be an extreme outlier. I manipulate my queue to make sure I have no more than 1 “virtue” movie at home at a time.

4. I go out to movies in the theater less often.

When I moved in with my boyfriend, we thought about cancelling one of our Netflix accounts and switching the other account to a service plan that would give us more movies for less money than our two accounts combined. But we decided we weren’t ready for that level of commitment.


bad Jim 08.03.04 at 11:07 am

Thanks to Netflix, I’ve seen the first two Harry Potter movies. It’s not clear whether to assign this to virtue or vice.

I also, the other day, saw a guy with a “Republicans for Voldermort” T-shirt, and I’ve yet to find anyone else who gets it. Perhaps that’s what happens at fifty, but I’m talking about well-educated family members.


bad Jim 08.03.04 at 11:13 am

My brother claims to have seen a bumper sticker reading: “Frodo failed! Bush has the ring.”

At least he knows what I was talking about.


Scott S 08.03.04 at 11:26 pm

It seems to me that the people who left “virtue” movies sitting around were always choosing some other pursuit that was more interesting to them at the time (including watching and then returning a “vice” film).

If the more interesting thing was usually not a different movie, then the obvious conclusion is that, for these people, watching movies was not a real joy for them, but just a distraction. So they were right to drop NetFlix–it’s a service for people who like to watch films.

If the more interesting things were always other Netflix movies, then they were getting the benefit they wanted. If Netflix only let you have one movie at a time, but you ALWAYS watched it right away, and sent it right back, you’d see 10-11 movies a month, which makes the price an awesome deal. So if one of your three slots is unused as a virtue movie sitting around, as long as you’re watching movies pretty consistently, it’s still a good deal.

Soooo, what I perceive here is that people who dropped NetFlix because of not watching virtue movies either weren’t avid enough film-watchers at all to profit from the service, or they allowed themselves to feel guilty at not watching the movies that were “good for them.”

Wow, that was longwinded!


Another Damned Medievalist 08.06.04 at 9:09 pm

Although We’ve got three virtue films hanging around (for a couple of months) I still think NetFlix pays off in the end. Sometimes we go through several movies as quickly as we can — other times, we’re too busy to take two hours and watch — especially if there are good mysteries on TV. Since next week’s mysteries on BBC America are really old Prime Suspects, I’m thinking the queue will move soon

Comments on this entry are closed.