web analytics reporting - I'm stuck

[alfa daley]

Hi folks

I'm a bit confused with web analytics and hoping a pher could shed some light.

Does anyone know when presenting web analytics figures for a site if it's better to compare absolutely identical periods eg. 30 days vs 30 days and ignore a traditional month on month comparison, because some months are longer than others and have more saturdays etc. Or would I be better comparing 4 weekly mon-Sundays period concurrently, so 28 days?

Slightly confused...

Edited by alfa daley on Thursday 18th February 19:24

[Simpo Two]

I'm not a web expert but the general rule of science is only to change one variable at a time. If two sets of data differ by more than one variable, you can't accurately attribute any difference.

Was that any help?

[JohnnyPanic]

We normally just do calendar months as it's easier for people to understand at a glance.

If there's anything statistically interesting pull it out and highlight it separately but I wouldn't overcomplicate right from the top.

[nsa]

Are you using Google Analytics? It might not be able to do everything you want, but I'm sure there is an export facility where you can dump the results into Excel and cut it your own way.

http://www.google.com/analytics

[jardinec]

if you have the longer term data then i would normally compare current vs. same time period 1 year previous.

plus if traffic is up the difference might be greater and look better

might not matter so much if not a seasonal business - although i cannot think of anyone that isnt seasonal in some way...

[purplepolarbear]

Does your business have any concept of financial reporting weeks / months - it may be better to have reports that run on the same days as them, and help tie web stats together with other financial reports.

If not, how much does your traffic differ over weekends? If it's broadly the same, then it will be less confusing to report on calendar months. If there's a significant difference, then report on 4 week periods.

[alfa daley]

Update on this: I attended TFM&A and spoke with a digital marketing consultant after his seminar and he was of the opinion that it's best to stick to monthly reporting rather than 28 days or exactly the same time periods as you are looking for 'trends' which should still be obvious. It also aligns with finance and other aspects of the business as the poster above has mentioned.

Quite a lot of interesting discussion about adding google tracking urls for campaigns to emails/banners so that you can identify sources in google analytics easily and see which one's are converting.

[evenflow]

I would say it depends on what type of trends you are trying to report on - for example, traffic volumes on Sundays? Month on month? Who are you producing the reports for? Maybe ask them what they would like to glean from the data?