Attention and software

RELease 1.0, March 26, 1992 by Michael Goldhaber

To conventional economists, pcs are something of a paradox. Their production has been an admirable growth sector, but their consumption or use seems to serve little purpose. As discussed in Release 1.0, 9-91, while millions of pcs found their way to corporate offices in the Eighties, they seemingly did nothing for overall productivity. Yet pcs are more than a fad. If people in offices didn't use them for compelling reasons, not just as fun toys, demand for them would have faded well before now. A correct economics should be able to explain the role they play. Over the past few years, I have been trying to put together an economic theory that can do so. This is a report with special relevance to software.

Though we hear the term less now, in the Eighties we were assured that pcs would take us further into the Information Age." The amount of information stored, transmitted, developed and processed continues to rise astronomically, but it is not clear why. Most people have simply taken it for granted that more information is good for business, but standard measures of growth don't reflect-the,explosion in information: As information took off in the past few decades, conventionally defined productivity growth turned flat.

Furthermore, people are just as likely to experience the growth of information as a problem: "I can't keep up," There's an information glut," and so on. If information were valued in itself, like normal commodities, a glut would cause us to cut back sharply on its production. Instead, we keep increasing exponentially the amount of information we put forth.

The satisfaction is in giving, not receiving, information

The question is why we keep producing more information. My answer: We do so because through it we hope to get attention. It is the economics of attention that can explain the value of pcs in today's offices.

Note that in this respect pcs are very different from mainframes; they serve as vehicles for personal communication, as opposed to mainframes or pcs used in mainframe-style tasks such as accounting, which generate data for business processes rather than for people. The following applies to personal computing, however it's accomplished, rather than to pcs per se. Streams of data are not attention-getting in themselves; it is the information in those streams of data that may garner attention, if it is either interesting or attractively presented.

Had we infinite attention to give, we might look at a data stream. But attention is scarce: Each of us has just so much capacity to pay attention and wants to use it to good purpose. One definition of economics is the science of the management of scarcities, so an economics of attention makes sense, in a way that an economics of information, which is open-ended and unlimited, would not. [What is scarce is the unique understanding or knowledge of information that gives you some competitive advantage, but that's another story. - ED]

The micro level: Attention barter

To start at the micro level, attention might seem an evanescent thing: You look at this, you look at that, and then you look at something else. If Mr. Jones gets your attention for ten minutes, what good does it do him in the long run - immediate gratification aside? The answer is that attention can generate further returns for the receiver - give Jones ten minutes of your attention today, and therefore you remember him a week from now, a month from now, maybe even a year from now. Attention is accumulated in the form of other people's memories and in their willingness and desire, on the basis of that, to pay attention again.

Of course, Jones can wheedle ten minutes of your time and then blow it. You may remember him but with aversion; you have no inclination whatsoever to pay attention again. As an economic transaction, from your perspective, Jones took more attention than he gave back, so you certainly don't owe him any more (in fact, you feel taken advantage of). If he talks for ten minutes about something of no interest to you - say, his back pains - you will stop paying attention well before he stops talking.

But if you happen to have a serious backache yourself, you might interpret Jones's remarks as an expression of sympathy, offering attention to you and your needs, an attention you might reciprocate by talking about your own aches. Each of you would feel that the other is giving as well as taking, that the conversation is an equal exchange, and that it leaves you feeling slightly better.

No knowledge of any practical value need be exchanged - no cures or doctors' names, just commiseration between fellow sufferers. This is not an information transaction; it's an attention transaction. The next time you're in contact, you will both remember your talk favorably and willingly exchange a little more attention. Through such more or less equal exchanges friendships are built. Each person comes to occupy a more and more important place in the mind of the other, as each conversation leaves some trace in memory. [You could also look at this as a neural-net phenomenon, sensitizing people or machines to familiar patterns. - ED]