A What’s Changed Report for Google Analytics

With all the online reports provided by Google Analytics, the one that seems to be missing, years after it was pointed out, is the What’s Changed Report. Way back in 2007, Avinash Kaushik wrote in his blog (Occam’s Razor) and his book (Web Analytics, An Hour a Day) about the wonderful value of this report (from ClickTracks, now Lyris). The conversation continued through several of Kaushik’s other writings since then, with other vendors picking up the baton (Juice Analytics, ClickEquations).

Such a simple concept – show me what, of significance, has changed from one time period to the next. This has huge value to web analysts because it makes you aware of things that simply wouldn’t appear on typical top-10/25 or consolidated trend reports. This is the daily report you should look at first thing in the morning. Has something unexpected happened that I should rearrange my day to investigate?

With the Next Analytics implementation, we wanted to provide immediate value, yet permit easy customization (we all share some common needs, but there are a few differences). By blending simple Excel cell references with the flexible Next Analytics scripting language, we have provided a workbook that will deliver value with the first refresh. Using cell references in the script commands make it easy to change the time periods (either duration or fixed dates), the dimensions (countries, sources, campaigns) and metrics (visits, page views), as well as the sensitivity (how much of a change is ‘significant’ to you) of the report.

Speaking of sensitivity, just what is a practical formula for ‘significance’? Well, after looking at all the debate around this issue, we decided we would avoid the statistical definition since it is way beyond most people’s ability to understand the implications. What good is a report if you can’t understand it? We chose a much easier-to-understand definition – the change in the percent of total. If visits from Pakistan went from 0.1% of your total visits to 1.1% of total visits, that would be a delta of 1.0%. Are you going to do anything about an increase in visits from Pakistan? That is for humans to decide. Is a 1% change ‘significant’ enough for your site? Change the sensitivity to a larger or smaller number. You are in control and there is no mysterious algorithm operating behind the scenes.

We’ve included this report in our most recent dashboard package, and will be issuing variants of it over time. If there is something specific you’d like to see, just let us know at support @ nextanalytics.com.

, , , , , , , , , , , ,

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>