TCS Daily

Red Sox Technologies

By Arnold Kling - March 9, 2004 12:00 AM

"...skepticism about micropayments is based not on general qualms about the spread of new technology but rather on economics, psychology, and marketing...customers' usage drops dramatically when there is even a tiny payment...consumers are willing to pay more for simple pricing plans, especially flat rates."
-- Andrew Odlyzko, in a letter to the editor of Technology Review, March 2004

This essay is about technologies, like micropayments, that perennially let down their diehard fans. Technologies that always seem so promising, but never live up to their potential. Technologies whose advocates say only need to be a little cheaper, a little easier to use, or to add a few more features, but they always seem to fall just short. In making their pre-season predictions, pundits will say this could be their year, but in the end there is disappointment. Woulda, coulda, shoulda. Red Sox Technologies.

The List

Here is a list of some well-known Red Sox technologies.

  • Micropayments
  • E-books
  • Speech Recognition
  • Video Conferencing
  • Social Networking Software
  • Virtual Classrooms

All of these seem to represent outstanding opportunities. Initial attempts to solve the problems go back years, even decades. Many failed ventures haunt the memories of investors and large corporations. Yet even today, active efforts are under way to try to find solutions that will overcome the resistance that each of these technologies has encountered so far in the cruel reality of the marketplace.

The engineers, tinkerers, and inventors who are still working on Red Sox technologies all labor under the illusion that all that is needed is a better solution. However, I think that you will find that if you examine Red Sox technologies closely enough, you will see that in each case they address the wrong problem. Like a doctor who prescribes the wrong medicine based on a too-hurried diagnosis, Red Sox technologies fail to address the cause of the underlying pain.

Micropayments and E-books

The rationale for micropayments is that information wants to be free, but creative people need to get paid. The idea is that by charging for content in small amounts, producers can achieve wide distribution and earn sufficient income. Ten years ago, to illustrate the concept, I wrote an essay entitled "Would you pay ten cents to read this article?" Many others wrote similar essays.

The problem with micropayments is that the real pain for consumers is in deciding what to read, listen to, or view. The real scarcity is attention. Micropayments give rise to what Clay Shirky calls mental transaction costs, which tax one's attention even further.

The real need in today's information environment is for recommendation and filtering. Content is not what is valuable. People want to know where to devote their precious time and attention. In this context, micropayments are far from a panacea. As Odlyzko points out, "sellers can derive more revenue from selling bundles...than from selling individual pieces."

E-books suffer from the same misplaced conception of the problem. They have the potential to save on the cost of publication and distribution. However, the big challenge for consumers is determining which books are worthy of their attention.

Speech Recognition

Computerized speech recognition has been within one or two years of widespread applicability for a lot longer than one or two years. At some point, we need to take a step back and consider the possibility that speech recognition solves the wrong problem.

Real human speech is an interactive process. When I talk with someone, I instinctively look for verbal and nonverbal cues of understanding. Real people do not listen patiently for entire sentences and paragraphs. They nod their heads, raise their eyebrows, say "Uh-hunh," and finish one another's phrases. Conversations are not sequential processes, in which people alternate between speaking and listening. Instead, when you listen as part of a conversation you are at the same time making verbal and nonverbal gestures. The actions of speaking and listening are simultaneous and interwoven.

By making speech a sequential process, computers try our patience. You can see how impatient human beings are with sequential speech by observing teenagers using messaging software. They carry on multiple conversations at once, because they have so much more mental bandwidth than sequential communication requires. Sequential communication ultimately seems sluggish and time-wasting, which is what makes business presentations so painful. And yet that is what even the best speech-recognition programs aim to achieve.

Video Conferencing

Video conferencing is another technology that solves the wrong problem. The problem that it solves is one of making it easier to hold meetings. But the problem is not that getting to and from meetings takes time. The problem is that most meetings themselves waste time.

Probably over 90 percent of all business meetings fall short of achieving their objectives -- often by a wide margin. It turns out that productive meetings are very difficult to plan and prepare. If corporations want to reduce the productivity lost in meetings, the answer is not to make it easier to arrange and attend meetings. The answer is to make it harder to call a meeting without clear advance communication of the background and expectations, thoughtful planning of the discussion process, and effective facilitation.

Social Networking Software

The idea of creating "friend-of-a-friend" connections has been tried many times on the Internet. I signed up for a site called SixDegrees almost a decade ago. More recently, I was invited to join a social networking site developed by a researcher at Google.

The Orkut network asks users to provide information in three categories: professional; hobbies and interests; and personal. For example, in the "personal" category, it asked, "What would someone find in your bedroom?" I responded, "My wife."

The assumption in social networking software is that everyone is on the make, and they want to use you as a stepping-stone to find career opportunities, people with similar pastimes, or lovers. But most people are fairly satisfied in these areas most of the time. The stress of seeking a new job or a new lover is something that most of us would rather do without.

In real-world social networking, when John asks me for help in finding a mate or a job, I am the one who searches my rolodex, so to speak. What social networking software does is allow John to search through my rolodex himself. I suspect that most of us do not want our friendship to be disintermediated that way.

I think that the real problem is what Howard Rheingold calls reputation systems. There are many situations when reputation is important. You want to go to a reputable doctor, hire a reputable employee, date a reputable person, and so on. The problem is that everyone has an incentive to try to manipulate their own reputation so that it appears to be high.

Having a connection to someone who is a friend of mine is one indicator of reputation. It is by no means the only indicator, and it certainly is not the best indicator. The worst employment decision I ever made was hiring a star employee's roommate on the theory that the roommate would also turn out to be a good employee.

Reputation systems have potential. Perhaps social ties can play a role in tuning reputation systems. But the focus needs to be on building a robust reputation system, not just on exploiting social ties.

Virtual Classrooms

I have not conducted exhaustive research into the technology known as "courseware," but what I have seen has been pathetic. The worst example was the software for my daughter's calculus class at the University of Rochester. The students were given complex problems that required many algebraic manipulations to solve. The software only told them if their final answer was right or wrong. Sometimes, my daughter had made a simple algebraic mistake. At other times, she was completely lost on the basic concept. The software gave her the same feedback either way. The only way that she could complete an assignment without spending hours on each problem was to get help from another student, or -- when she was really desperate -- from her father.

Most web-based education software seems designed to enable a teacher to make course materials, such as lectures, accessible by computer. However, if access to lectures were the issue, then the VCR would have been the ultimate solution in distance learning.

What I always say is that Teaching Equals Feedback. In a classroom, students are giving teachers feedback (much of it nonverbal) as to how well they are absorbing the material. In grading problem sets, papers, and tests, teachers provide feedback to students about where they need to work harder. The feedback loop is essential to the learning process.

Distance learning will be truly effective when tools are available for delivering timely, high-quality feedback over the Internet. For example, that when a student does a math problem, a teacher needs to be able to read the student's solution -- including all of the steps -- in order to point out where the student went wrong.

Break Up the Red Sox

When a technology keeps getting better and cheaper without achieving commercial success, it may be time to stop trying to improve it. Red Sox technologies tend to arise from a fundamental misperception of the cause of pain. Only when entrepreneurs step back and address the real problem can they apply technology to find a solution.


TCS Daily Archives