Have You Read It Yet?

Often times, when interacting with my friends and acquaintances on a day-to-day basis, I find myself asking them "Have you seen that [post I shared/post I wrote]?" Unfortunately, this data must be obtained and is a quite obvious sign of trying to gather reach data from the audience. This is problematic because responses can be skewed and do not give that much insight.

An idea I've been thinking a lot about is social media. Often times, there's an interest in seeing which users have seen which content, when, where, and such things as that. Many people--not just corporations but actual users--are interested in understanding who their posts reaches, what kinds of feedback it gets, and such things. This data can be ascertained by measuring the amount of time that a user spends viewing a certain post. A system can be designed to determine whether or not content is visible and then determine the user's response.

Unfortunately, measuring the user's response is difficult. Ideally, it would be as simple as measuring whether or not they interact with the content, however people are becoming increasingly lazy and such behavior is difficult to measure. For example, Google+, Facebook, and Twitter all aggregate content into a centralized page (called the "news feed" or "timeline") to display to the user, and the user generally does not click anything. Therefore, the data needs to measure not only interactions with content but times that a post has been viewed.

Using statistical analysis algorithms which I have neither formed nor found as of yet, the user's post response times can be measured. There are users who simply scroll through their feed until they find something interesting, and it is important to measure which posts they scroll past as well as what causes them to stop. If a post is visible, a timer would be started for that post, and as soon as it became invisible or was accessed, that timer would be stopped. If the user clicked on the content, then that would count as interest in that content and would count as a reach. If the user read through the content (i.e. if it were a text post), then the reach of a post could be measured through post view time.

I plan to toy around with these algorithms soon. Timing post content seems to be the most difficult thing and differentiating attention also. The main problem will be aggregating and processing it in a timely manner and without considerable lag. We'll see where this idea goes.