Thursday, August 19, 2021

Is Facebook Asking the Right Questions?

I just finished reading a BBC News post entitled “Facebook reveals most-seen posts are inane questions, not politics.” The story itself begins as follows:

After being accused of spreading right-wing content, Facebook has revealed that its most-viewed posts include asking if sugar goes with spaghetti.

Its first-ever report on the contents of people's news feeds in the United States focuses on how many people see a post rather than any other measurement.

The good news is that, at the very least, this counts as an attempt to clear Facebook of any accusations that is has become an instrument for “politics by other means” (or, perhaps, the more activist proposition, “revolution by other means”). The bad news is that they may have engaged inadequate analytic techniques to arrive at questionable conclusions. The fact is that people were bombarded with “things to see” long before Facebook added to the bombardment.

We might do well to think of the mechanisms behind Facebook posts the same way we think about advertising. We are all used to the high volume of advertising that we encounter through television. (There are other sources, such as billboards and newspapers; but my guess is that television advertising draws more attention, simply because it can deploy more attention-grabbing techniques.) Furthermore, the advertising business is not just about drawing eyeballs. It is a matter of drawing the right eyeballs in the interest of getting “the mind behind the eyeballs” to take some desired action (usually buying something, whether you need it or not).

In that context the concept of “most-viewed posts” is a relatively meaningless metric; and rank-ordering the posts themselves according to that metric is even more meaningless. Indeed, even the question of passing a post to others may not be that significant. The more difficult question to deal with is whether acts of reading and sharing lead to actions being taken that have nothing to do directly with reading and sharing.

In other words Facebook needs to be addressed on the same turf as the question of whether the storming of the Capitol building was the result of deliberate incitement. This is a complex question that cannot be answered by simple bean-counting. It is not so much about “spreading the word” as it is about how persuasive those words are and if they are persuasive enough to lead to action.

My guess is that there are no analytical tools for identifying such persuasion, not among Facebook resources or anywhere else. Those in the business of advertising may have a richer qualitative understanding of human nature, but they probably know better than to reduce that understanding to quantitative metrics. The best we can do in the face of a possible crisis is ask, “How did we get into this mess?” Where the National Commission to Investigate the January 6 Attack on the United States Capitol Complex is concerned, the House of Representatives tried to address this question. Whether or not any useful results will arise is still up for grabs.

Meanwhile, we should be as careful with Facebook’s accounts of its own analyses as we are in taking any television commercial to be verifiable truth.

No comments:

Post a Comment