martynj.com - things I have learnt

Notes from Google Conversion Uni: Introduction to Google Analytics

I made the following notes by watching the slides on the Google Conversion University Help page: http://www.google.com/support/conversionuniversity/

The Google Analytics Tracking Code (GATC) javscript is executed when each page is loaded. The _trackPageView() method is called and GA first party cookies are written and/or read. Invisible GIF request is then made to transfer data to secure Google servers.

Data is processed throughout the day.

Google Analytics uses only first party cookies (no third party). Many people block 3rd party cookies in their browser preferences, but most allow 1st party.

If user disables cookies, GA won't be able to track them at all since all data is transferred via 1st party cookies. If user deletes cookies, they'll still be tracked, but will be counted as a new visitor, and conversions data will not be attributed to a prior referring campaign. Much less common is disabling JS altogether. The GA code will not run and the user will not be tracked. GA still tracks visits to cached pages so long as the user is still connected to the internet.

Google Analytics does not report on PII (Personally Identifiable Information) or share data with any 3rd parties.

JavaScript errors can cause GA code not to run if they terminate the execution of JS on the page before the GA code is called.

No reporting tool is 100% accurate. Looking at trends is the most important part of interpreting results.

Interface Navigation

Each website with Google Analytics enabled corresponds to one account. Each account can have 1 or more profiles that can be configured differently.

Reports can be added/removed to Dashboard (starting view when entering a profile).

Best practice when comparing data ranges: use same days of the week in before/after ranges -- not just eg. 10 days before/after a given date.

Scorecard below narrative on report pages gives data aggregates and averages for the data.

Report tabs:

  • Site usage - metrics such as PVs, time on site, bounce rate

  • Goal Conversion - conversion rates for each goal

  • ECommerce - shows eg. revenue, num transactions, av revenue

  • Clicks - AdWords campaigns reports show eg. clicks, cost, rev per click, ROI

Some reports, eg. Bounce Rate, allow reporting metrics by hour -- look out for the clock icon in top right corner.

Five views available for each report:

  1. table view

  2. pie chart

  3. bar graph

  4. comparison bar graph view (shows whether each entry in the report is performing above or below site average)

  5. shows summary report with graphs for the traffic you're analysing.

Interpreting Reports

Don't focus on just a single metric. Use them in the context of other metrics.

We can use the "Compare to Site Average" visualisation to see eg. which pages have a bounce rate that is significantly higher/lower than the site average.

A visit or session lasts until the user closes the browser or is inactive or >30 mins. A PV is counted every time a visitor opens a new page, refreshes an open page, or re-visits a page they have already visited. An absolutely unique visitor is identified by a unique visitor ID combined with the timestamp of their first visit.

Time on page is calculated by subtracting the timestamp of the PV of the 'next' PV from the timestamp of the PV in question. If no next page is viewed, ie. the user exits, the time on page for that impression is 0. Bounces are excluded from the calculation of average time on page ie. all zeros are removed.

Average time on site is calculated by dividing total time on site for all vists by number of visits (inc bounces).

Virtual page views and events can be configured for sites that use Flash or AJAX heavily.

What makes a good source of traffic? Bounce rate can be an indicator of this. NB. Blogs often have a high bounce rate because users visit to read a single update, then leave.

By default, GA attributes a conversion or sale to the campaign that most recently preceeded the conversion or sale. Direct visits will not take credit from previous referring campaign.

To prevent a specific referral from taking conversion credit away from a prior campaign, append utm_nooverride=1 to the link to keep conversion credit with the original campaign.

Tagged: google tech analytics

blog comments powered by Disqus