Thursday, March 4, 2010

The problem of matching social attention and products...


Many people have already stolen the attention-scarcity ideas from Herb Simon and said that the most important problem in our information overloaded society is the efficient distribution of attention. What some have called the "attention economy" is nothing more than a re-packaging of this idea.

In business, of course, getting the consumers' attention is quickly becoming an important aspect of being successful. Traditional ways of getting people's attention is through advertisement, and we have witnessed a dramatic transformation of how advertisements work in the online world in the last decade, from display advertising to search advertising and, more recently, further to action advertising. Increasingly, we can tie advertising dollars to direct consumer action.

For us, it was not a stretch, then, to start thinking about how the consumer actions are starting to quickly feedback to product design. Thus, we now have people talking about crowdsourced product designs. The most agile companies now listen to the consumers via channels such as Facebook, Twitter, and Blog analytics. They do this via services such as brand management consultants and sentiment analysis tools, so much so, they are able to discern tiny changes in consumer awareness of product issues and their desires.

We know also that traditional economic models serves to optimize the distribution of products to people who want them. But these models have also recently been used to optimize the distribution of people's attention to products that might serve their needs. The two usages obviously goes hand-in-hand.

If we can help companies to serve people attention spots just-in-time with the best products, we would have a highly optimized economy that wastes little energy in distributing worthless advertisements (or spam). In fact, the existence of spam points to the inefficiencies in the economic system.

Turns out that versions of this problem exists everywhere in the Web2.0 world:
  • The problem of efficiently distributing the best tweets to the people who want to view them is a version of this attention distribution problem. Any time you see a tweet that was worthless to you is an opportunity for optimization.
  • The problem of pointing experts to the most valuable articles that they can contribute to in Wikipedia is another version.
Solutions to these problems might take the form of recommendation systems or filtering systems, but might also be efficient interactive browsing systems (for products in an online store like Amazon, or articles in Wikipedia). Some thought experiments:
  • What if we can design an expertise finding system that recommends the best articles for you to contribute to in Wikipedia? Would it increase participation rates?
  • What if we analyze your social network everyday and tell you the best tweets that you should spend five minutes on? Would more people retweet more often?
  • What if product designers are better tuned to trending topics and needs, would they enable companies to succeed more often? Are companies like Zazzle and Cafepress the prototype examples of lubricating this path?

Your thoughts?

13 comments:

Anonymous said...

Dan Cosley did some research on suggesting content for people to edit: http://www.cs.cornell.edu/~danco/#suggestbot.

Disclosure: I was a co-author.

Dan

Ed H. Chi said...

Yes, I am quite aware of that research. Mike Krieger at Stanford also has some research in this area called WikiTasks.

Aneesh Karve said...

CMU RADAR is "a cognitive assistant that learns to help a human user in situations of intense information overload" (http://www.radar.cs.cmu.edu/index.jsp?s=2).

To use spam as an indicator of an inefficient market, you need to show that its price and value are disjointed. I don't see a strong dislocation. For the most part, and for the time being, programs like gmail have all but licked spam.

As you suggest, importance engines will be of growing importance. (For the same reasons that folksonomies and reviews have become so important.) It is equally important to be able to authenticate reviews and folksonomies with a machine. Perhaps solving one problem nets us the other?

A fascinating issue will be the inductive bias of the machine algorithms we will use to label our data for importance. These algorithms will not only be lenses into the web, but will simultaneously *shape* it. The same is already true for pagerank. We will see scale-free network effects and...who knows what else?

James Salsman said...

For Wikipedia, I recommend displaying the top-10 or top-20 most popular related articles in the sidebar following the alternate language interwikis on each article. I think this would serve the purpose of showing both editors and readers where attention is being directed.

David Karger said...

The paradox around efficient distribution of attention is that it's circular. In traditional markets, we allocate resources by ensuring that there is good information (through prices) about their value. But we have to pay attention to that value information in order to allocate the resource. In information markets the attention resource we want to allocate is the same one we need to consume in order to allocate it. There are some interesting diminishing returns. More at http://groups.csail.mit.edu/haystack/blog/2010/02/18/cscw-the-attention-economy-and-efficient-information-markets/

Ed H. Chi said...

@ak: I'm familiar with both RADAR as well as CALO. I think it is quite evident that the price and value are disjointed in spam. For email, perhaps some email systems have done a great job in filtering spam, but not all email systems do a good job. In fact, the ACM filtering sometime filters out good msgs that I want. In Twitter or Facebook, the msgs are shorter, and I see more noisy or non-useful msgs more often than I would like. I suspect that this is true for a lot of people. I liked your analogy of "importance engines" as lenses on the web, and I agree.

@James: Displaying the related articles in the sidebar tells people the possible related things to contribute to. It will somewhat make the feature space more featureful, but it does not entirely solve the original question I posed. We want to connect the expertise of the editors to the right articles. This problem is quite a bit bigger.

Jon Awbrey said...

Re: "What if we can design an expertise finding system that recommends the best articles for you to contribute to in Wikipedia? Would it increase participation rates?"

Ed,

This is so clueless that it boggles the imagination.

You have clearly not spent a lot of time actually participating in Wikipedia.

Do you have any idea how many "expertise enhancing efforts" have gone already gone down in flames on Wikipedian shores?

You need to put aside your wiki-pipe-dreams, spend a little more time talking to people outside the Wikimediocracy, and try to grasp the real dynamics of what goes on in Wikipedia.

Ed H. Chi said...

@Jon: I know your past gripes about Wikipedia, but let's put that aside for a second. My understanding of past expert enhancing efforts is that none of it has been automatic (i.e. done by a recommender engine). Further, the following post to this one suggests making the system a lot more transparent and social (verified real names available as a feature, wikidashboards to show contribution patterns). I don't believe I'm putting forth a proposal that, in aggregate, resembles past expert enhancing efforts.

Having said that, you're quite right that there are many aspects of wikipedia dynamics that worries me and bothers me. I have made some hundreds of edits, and have encountered edit wars, but I'm not an elite editor and don't claim to be one. I understand that, unlike the Utopian image it has cultivated, there are a lot of undesirable (uncivil?) behaviors in the system. It's not all that it is cracked up to be, but that doesn't mean I'm going to throw the good parts of it out with the bad. It still informs a lot of people with lots of important knowledge.

James Salsman said...

@Ed, I missed that you were suggesting a recommendation engine. I am sure that can work if you use categorical k-nearest neighbors instead of manually trying to convert the huge folksonomies of categories and wikilinks into quantative data. I hope you will also implement the much easier top-20 related popular articles list in the interwiki margin area so you can do a side-by-side comparison of effectiveness.

@Jon, this is a very reasonable suggestion. You are right that there have been similar efforts -- e.g. SuggestBot -- with respectable opt-in rates among editors in the 1000s which I think would be more accurately characterized as chugging along than having gone down in flames. That's not to say that most of SuggestBot's suggestions are any good :) but editors seem to tolerate the handfuls of chaff to get at the few grains of wheat. Which efforts were you referring to?

Jon Awbrey said...

Let me rephrase your suggestion:

"What if we can design an expertise finding system that recommends the best articles for you to contribute to in [Ideal System X]?"

That, of course, is brilliant idea, and well worth trying somewhere or other, almost anywhere but Wikipedia.

The catch is that there is already a culture in place at Wikipedia, a culture that is antithetical to every condition that it would take to make such a utility work.

This is true of every attempt to bring Wikipedia up to par with the intellectual virtues usually associated with responsible scholarship and journalism, and automation would not make the slightest difference, since you cannot implement disciplined routines without the consent of the mob.

You could try to follow the 5-year debacle over "flagged revisions" if you wanted an ever-ongoing example.

Ed H. Chi said...

@James: The problem with SuggestBot is that it is too passive. What is needed is a system that suggests pages for you to edit as soon as you have edited one. Imagine the scenario that you're editing a page about Entropy. As soon as you're done with your edit, it should say, "BTW, there are these pages that are related to Entropy that has been marked as needing attention. Would you like to take a look at what the issues are?"

Think Amazon's recommendation engines for product purchases for each specific user. We need the same for Wikipedia. A personalized interface, just for you, of your information interests and recommendations.

Dan Cosley, the main PI behind SuggestBot, is a good friend of mine, and we're alumni of the same research group, so it's no surprise that we're thinking of similar issues.

James Salsman said...

@Ed, great! Let me give you an example why I think simple top-20 popular related articles will be a strong contender against even very refined subject matter recommendations: If you just finished editing [[Dead zone (ecology)]], would it help you as an editor interested in that subject more to know that [[Carbonic acid]], which gets 1300 views per day, is a high-importance article still at B-class, or merely to know that [[Water]] gets 10000 views per day?

@Jon, the lack of flagged revisions for biographies of living people is an excellent example of a serious problem. I recently learned that the Wikimedia Foundation has never once been sued for copyright, but does occasionally get sued for defamation, even though there are safe harbor laws for both making such suits much less cost-effective than a simple OTRS email. If only we could find a way to re-focus out-of-proportion copyright concerns on implementing flagged revisions for BLPs! But that in addition to all of Wikipedia's other serious problems don't really support your universal quantifiers, "every condition" and "every attempt."

Ed H. Chi said...

@Jon:
To continue my thought in the other thread, if wikipedia is a patient, you seem to be saying that it is so sick and so beyond help, that any treatment is going to be futile. You would not even attempt to resuscitate it. I suppose you're entitled to your opinion, but that's not how I see the situation.

To draw another analogy, it may very well be the case that global warming means that Earth is doomed, and that Mother Earth is so sick that any treatment we apply will be futile. But we don't have another Earth, so shall we just sit and wait to die as humans?

Currently, I don't yet see any viable alternative to Wikipedia. You maybe right that it is doomed, and perhaps even someone will eventually start a different project that will gather similar amounts of attention and effort, but until that happens (until we find another Earth), I'm going to keep working on this one.

@James: It very well could be that your simple recommendation engine would work. I see it as an empirical question that is worth trying. Assuming that there is still an opportunity for us to take action before Wikipedia is beyond help, we should try a number of these different techniques to change participation architecture of Wikipedia. It may be that Tool-Ware (system) will start to affect People-Ware (culture), and we can avoid the death of the community.