Subscribers - they come, they go

Looking at my subscribers over time they show a fairly steady trend upwards, but that statement glosses over a significant level of day to day movement. During the first six months or so, it wasn't unusual for the low in a month to be less than 40% of the peak value.

What has changed during the last couple of months is the level of volatility. While subscriber numbers continue to change every day, with unpredictable highs and lows, the lows are staying within 65% of the peak value. What it may mean is that as the subscriber base has grown I have acquired a group of subscribers who switch on their RSS reader more regularly or at least more predictably. In the early days, I was looking at a curve which exhibited the statistics of small population and that wasn't really representative of a larger universe.

That's important to remember when we do any form of research, particularly if we are drawing conclusions from the data. It has always been true to say that people access their RSS readers on different timescales which can make subscriber numbers appear volatile. It wouldn't have been true to draw a conclusion that the inherent volatility of subscribers to this blog was 60%. A small population can help illustrate a truth - subscriber figures are volatile - but it cannot be used for accurate forecasting purposes.