Google Analytics speeds up

The GA team have announced three new features that aim to speed up the experience of using Google Analytics, particularly for large accounts. The updated reports should be available to all later this week.

Progress bar for report loading

For anyone who has ever sat staring at the Coremetrics clock of doom, or watched GA or SiteCatalyst grind your browser to a halt while trying to load a heavy report, you will share my pain. However the first new update to GA will aim to address this previously unquantified 'stare idly at the screen' time. They will be introducing a progress bar to visualise how long is left before you can start analysing your report. This may seem like a small change to some but not knowing how long a report will take to load is an analyst's nemesis; so believe me every little helps.

Google Analytics Loading Bar
Google Analytics loading bar


Cached reports

It is entirely fair to say that the first update won't speed up Google Analytics but instead it will simply tell you know how long the reports are taking to load (a welcome addition nonetheless!). The second update however is to introduce caching for the data sitting behind most reports (as long as you have the interface open), which will go a long way towards speeding up the overall experience of using GA. A notification will appear at the bottom of your reports (see below) if they have been cached and if the data does update the GA team state that this will be kept in sync.

Google Analytics cached report message
GA cached report message


Sampling controls

Anyone with access to large accounts will be familiar with "fast-access mode" or sampled reports as you may know them. Sampling occurs in GA when large data sets are queried for data that is not available in aggregate and certain thresholds are met. These thresholds are:

  • 1,000,000 maximum unique dimension combinations for any type of query.
  • 500,000 maximum sessions for special queries where the data is not already stored.

Essentially what this means is that GA takes a statistically random sample from the entire data set (when one of the above thresholds is met) and then estimates the overall report results based on this sample.

The new updates will include a three part change to how the report sampling in GA is handled:

1. You will now be able to see how many visits are used in the sample (very good news indeed!)

GA sample size message
GA sample size message

2. GA will be reducing the minimum visits threshold that determines when sampling begins in reports from 500,000 to 250,000. This change will increase overall speed, particularly for larger accounts where sampling has always occurred frequently. This on its own may seem like a negative change but never fear the third part to the sampling update allows you to set this back to the previous threshold of 500k if desired.

3. A new report calculation control (sampling control) will be introduced, which can be activated by clicking the square grid shown to the top right of the sampling message shown in reports. Once clicked you will be presented with a slider which you can use to control the number of visits used to calculate the sampled reports (up to a maximum of 500k).

Sampling control
GA sampling control

The data updates automatically when you use the slider and your preference is remembered and carried across all reports until you log out. The basic principle is: the lower the sample the faster the reports load; the higher the sample the more precise the data will be. You will be able to experiment with this to find the right balance for you.