Removing Daily Seasonality from Web Analytics Data

While I generally begin to look at web analytics data at a weekly or monthly level, there are times when it is useful to drill down to daily numbers.  This can be when examining the reason for a change in the data or simply to review the previous day’s performance.  But an issue arises which [...]

Continue Reading...

The impact of hot weather on website traffic

I am guessing there would have been a fair few questions asked this morning why websites didn’t perform as well as expected in the UK over the weekend, possibly down around 3% to 5% against last week.  If the usual suspects (online marketing, server going down) have been eliminated, then the reason in many cases [...]

Continue Reading...

Which is the right metric to use?

I get asked sometimes which is the best metric to use when creating a certain report. My rough rule of thumb is to go back and ask what sort of question it is that you are trying to answer with this report. If it is related to: the number of people, use unique visitors traffic [...]

Continue Reading...

What weight division do your visitors belong to?

The engagement merics that I discussed a few weeks previously are a useful method of understanding user behaviour on site. However it must be remembered that they are averages and that there is no such thing as an average user. So while these are useful as a single number representing these metrics, there is an [...]

Continue Reading...

You notice a change – what to do first?

So you have set up some weekly reports, focusing on key metrics – whatever is most important to your business/website in understanding its performance. Monday morning, you crawl out of bed, into work and update your reports so you can check out this performance (hopefully at the touch of a button). But something has happened, [...]

Continue Reading...

Time Periods

So you have your basic metrics and you know the best ways of giving the numbers some meaning. What then is the appropriate time period to use in order to understand performance when looking at this data? Should it be day, week, month, quarter, year or something different altogether?

Continue Reading...

Data gets lonely

Any metric by itself is inherently meaningless. It is a number, a percentage, a ratio but without something to compare it to, there is no way of knowing if it is good, bad or indifferent. The metric needs to be compared against something in order to give it meaning.

Continue Reading...