Why I don't trust RSS subscriber figures

I've said before that I'm skeptical about the value of RSS subscribers. Don't misunderstand: I think it is very important to provide RSS feeds; I just don't think counting subscribers means much.

I've given reasons for that:

Just because an RSS reader is fetching a feed doesn't mean that any human being is still actively reading the posts in that feed - they might be scanning the headlines only or might have just forgotten about it entirely - yet that person who is actually not reading anything gets counted as a subscriber. Subscriber yes, reader, who knows?

I realized today that putting Adsense ads on your RSS feeds can help give you a better idea of how many of those supposed subscriptions are really being read. Google's Adsense reports will tell you how many ad impressions are actually seen. Unless someone actually reads the feed, there won't be any Adsense impression.

Unfortunately, what you don't know is whether a reader saw one ad or many. Google puts a variable number of ads into your feeds, and that number is only partially under your control. You also can't control how often people are reading: you might post once per day, but a subscriber might read only once or twice per week. These variables make it impossible to directly correlate Adsense impressions to feed subscribers. However, we can get some clues.

For exemplar, consider that according to Feedburner, we have something around 2,000 RSS subscribers here. As noted, we have no idea how many of those are really reading regularly or at all: all we know is that their RSS reader has been fetching the feed. According to Adsense, there are approximately 2,000 weekly ad impressions from those same feeds .

If all of those 2,000 subscribers were really reading every day, there should be more like 14,000 impressions, right? Therefore, it's obvious that a good number of those supposed readers either do not read, or at least don't read with anything that can display an Adsense ad and thus trigger an impression. I'd guess that the number of actual readers is something closer to 1,000 people or less - possibly quite a bit less!

Again, RSS feeds are an indication of popularity. Obviously a site with 200,000 subscribers is much more valuable than a site with 2,000, no matter what the percentage of actual readers is at either site. At the same time, it is also apparent that in reality, the RSS subscriber figures are probably always overstated.