One problem with newspaper RSS stats used to be that they had a huge component of bloglines readers which were excluded when people switched to counting on Google reader. That punished the early adopter sites.

And what about readers who read headlines in sidebar widgets on other sites – do they count? E.g., this third party Times blogs widget has had 2700 downloads and a lot of installs.

http://www.widgetbox.com/widget/times-online-blogs

How many installs will their own widgets have had, and how many readers?

‘Tis a quicksand.

They’re not a fixed cost – they require maintenance. Let’s leave that aside for now.

No one has ever queried the google reader numbers to the extent you have before. It’s clear that they are just one player. But they are a significant one and it’s fairly commonly accepted that they are accurate as far as their own subs go. It’s nothing like download manager, which is trying to measure unknown unknowns.

RSS feeds are a fixed cost to newspapers. They provide short details to readers, and allow them to link to the paper for more details.

I’m surprised at the supposed readership rates the author quotes. I suspect that these numbers are not accurate. The only really useful numbers would come from the server that actually supplies the feed. I get the impression that the numbers the author is quoting are best estimates that the web designer put on that reflect a rough heuristic of use. For example, consider the Microsoft download manager which uses a rough heuristic that causes it to display download times that vary from minutes to hours.

In short, the author should realize that not all numbers are valid simply because they appear on a web site or are automatically generated. Instead, one needs to look at the context in which they appear, and to verify any numbers with those who implemented them. A lot of work? Yes. Tedious? Yes. Coincidently, these are the same ingredients that used to go into good news reporting before it became excessively profit oriented and time obsessed.