Eyeballs or data -- May 13, 2005

Carlton Vogt's

Enterprise Ethics

Volume 3 Number 13 May 13, 2005

Eyeballs or data

Which is the proper currency for reading online information -- or is it both?

Recently, one of my favorite news websites decided that it wanted people to register before visiting the site. The registration process was actually pretty invasive. They wanted to know such things as how much money I make, whether I planned to buy a car in the next three months, six months, or year. And they wanted my email address.

This is becoming more common and the information asked for in the registration process is more or less invasive, depending on the site. So, the inclination to fudge takes over. Do I tell the site managers who I really am? Or, am I a 95-year-old woman who makes $150K a year and is interested in needlepoint and motocross?

But it all raises the question of how much personal information -- if any -- we are ethically required to trade for the privilege of reading online information.

We could take a hard-line position and say that it's a question of fulfilling a "contract" on the honor system. These people put the information out there and request only that we make a "small payment" of some information in exchange for reading it. By reading the material, we are accepting their conditions and are therefore duty-bound to honor their request.

It's a lot like when I was a kid and the woman down the street would put sweet corn from the family farm on a table in front of the house. Other people would come by, take a dozen ears of corn, and leave the money in a box that was on the table -- even making change from the bills and coins already in there. (OK, I know I'm dating myself with this one.) People were expected to leave money for the corn, and they did.

Although I don't think you see these arrangements so much any more, I think a lot of people would still say that someone who took the corn without leaving the money was doing something wrong. I'm not sure the data-exchange request enjoys the same level of support.

When I lamented about the supermarkets matching our purchases against our identities through their "special saver" cards, quite a few readers wrote in to tell me that I was a boob if I actually gave them my real name and address when signing up for the card. That leads to two possible conclusions: the world has gone to hell or that the data-exchange scheme lacks a certain symmetry that existed in the corn-for-cash arrangement. A third possibility is that it's a little of both.

Without trying to figure out which of those three possibilities is at work, let's just assume that there is some kind of essential difference between the corn scheme and the "give us your personal data in exchange for reading our stuff" scheme. And I think the answer can be summed up in one word: eyeballs.

"Eyeballs," for those who aren't in "the business," refers to the number of people who visit a Web site, and site managers have become very adept at being able to tell how many people visit their site, whether they are return visitors, how long they spend there, where they go, which parts of the site they read, where they come from, and so on.

And this is the real currency of online information. If you sell ads on your site, then you need to tell advertisers that you had one million visitors a month for the last three years, that 800,000 of these were return visitors, and that they spent an average of 20 minutes looking over the content, including, everyone hopes, the ads. Lured with that information, the advertisers will buy an ad on your site.

At the end of the day, however, the bean counters at the company doing the advertising will count up the "click-throughs" -- how many of those visitors actually clicked on the ad to get more information on the company or perhaps to make a purchase. If there is a high percentage of click-throughs, and these are measured in extremely small percentages, the "ad buy" was a good deal and might be repeated.

Now, Web sites would dearly love to be able to "qualify" their visitors, or at least develop a demographic that would allow them to sell their ad space more readily. It would be a tremendous help, for example, if they could tell potential advertisers that 45 percent of their readers were between 18 and 35, made an average of $50,000 a year, and were planning on buying a car in the near future.

The problem with trying to collect this information, however, is that the process is horribly unreliable. If someone is telling the Web site that they are a 95-year-old woman who makes $150K a year and is interested in needlepoint and motocross, then it's safe to assume that other people are lying too, and that raises the question of whether the "unspoken contract" idea is so broken down that it's relatively meaningless.

Even at controlled-circulation trade magazines, where the magazines are only sent to those who "qualify," the circulation people know that a substantial portion of their readership is bogus. Not all of their readers supervise more than 19 people and are the decision-makers for over $500,000 in IT expenditures every year. But the trade-off among magazines is that they assume the proportion of bogus subscribers is constant from one magazine to the next, giving the relative figures between books some kind of meaning.

I could also make a pretty good argument that what is of real value to the Web site are your "eyeballs," the page views and other tracking data that you leave behind when you visit the site. This is the real currency of online information. The demographic information they're seeking goes beyond that, and the more invasive it is the more it constitutes trying to get you to "pay" more for their information than it's worth.

I solve the issue in several ways. Sometimes, depending on how much I value the information, I may just stop visiting a site when it tries to collect information from me that I consider inappropriate. There are some sites for example that want your gender, age, and zip code. I don't have a problem with giving those. When they want my name, email address, and income level, I think they've crossed the line.

For sites that I visit infrequently -- or sometimes once -- I have no qualms about using "Bug Me Not" which will give you a login and password for non-subscription sites. My argument is that because I'm such an infrequent visitor or will most likely never visit this site again, any demographic information I give is meaningless and quite possibly misleading.

The non-subscription angle is important. I don’t advocate or condone giving out logins and passwords to subscription-only sites. On the other hand, once I've paid my money to that site, I consider my end of the bargain upheld and still view any attempt to get demographic information as their trying to get something for nothing from me.

Sites that I visit frequently, I will give general demographic information, as I see that as a small price to pay for their efforts in putting the information online. I won't, however, divulge anything of a personal nature -- name, address, income, phone number, or email address -- that could be used against me by a spammer or unscrupulous marketer.

For email addresses, I use an email anonymizer -- Emailias -- which allows me to create a unique email alias for every such request. This send email to my main account, but under an alias. It also tracks who I gave the address to, so that if I'm spammed, I know who did it. And I can cancel the alias at any time without having it affect my main email account.

But what happens when the site "requires" you to give more information than you're willing to give during the sign-up process? Should you just say goodbye to the site and let them suffer, if indeed they do, from a drop-off in visitors because people don’t want to give our personal data? My best guess is that the site won't necessarily suffer, because, if my powers of observation are well-tuned, most people will just spoof away and help the company build a totally unreliable demographic database. Given that reality, could you just say that the process is so flawed from the outset, and everyone realizes that the data is flawed, that your spoofing your information has no additional deleterious effect?