Google recently announced that the Google Analytics 360 export into BigQuery now has the ability to run up to every ten minutes, giving you super fresh, almost real-time, granular data!
The export of GA data into BigQuery previously only ran every eight hours. For example, exporting data for 00:00 to 08:00, 08:00 to 16:00, and 16:00 to 00:00.
Although it was useful to have some intraday (current-day) data in BigQuery, this x48 increase in data freshness is a game changer in terms of how fast you can react to user interactions on your site using BigQuery.
How can this benefit me?
Think about merchandising and personalisation. The ability to collect granular user behaviour as it happens opens up the possibility of automatically analysing / modelling on those behaviours and changing what users see on the site in that same session by feeding insights back to your content management system via BigQuery.
You could also suppress media spend quickly where it’s likely to be wasted using propensity scoring. Let’s say a user shows some kind of intent on your site, not necessarily linked to a high likelihood of a conversion. Rather than targeting them immediately with everything you’ve got, use the super fresh BigQuery data to work out the likelihood of a conversion from that user and exclude those with a low propensity score from these campaigns.
You could also set up near real time dashboards for Black Friday or TV ads that are fresh as a daisy!
We could go on for hours...
Are there any concerns before I make the switch?
A few yes...
- No Ads data - The main downside is that Ads data won’t be included in the GA data if you enable streaming exports. This is intentional.
Integrated data in GA from platforms like Adwords and DoubleClick require additional processing (incl. fraud detection etc.) before it’s fit for purpose, so it is left out here by design.
- Not everyone is eligible - There are a number of prerequisite that your GA view must adhere to before you can use the new continuous export. These are mostly covered here:
Not listed is the fact that you must be a Google Analytics 360 customer - This has always been a requirement for the GA export to BigQuery.
- (Small) additional costs - The continuous export option uses the Cloud streaming service, which includes an additional $0.05 charge per GB sent to BQ. This shouldn’t exactly break the bank and will likely be covered by your $500 BigQuery credit per month awarded to you as a GA 360 customer anyway.
How do I get set up?
Assuming you meet the requirements and are ok with the concerns mentioned above you can enable the feature within GA admin.
Under Product Linking, go to All Products and Link BigQuery/Adjust Link (if you already have it linked) and amend/set stage four of the setup wizard before saving:
By the next day, you should see a new set of tables in BigQuery called ga_realtime_sessions_YYYYMMDD with an accompanying view set called ga_realtime_sessions_view_YYYYMMDD .
These tables include the same schema as the standard 8hr GA tables, however there will be three new columns added at the bottom which are specific to real time exports:
visitKey - An identifier for the session that the hit belongs to.
exportTimeUsec - The time that the hit was streamed into BigQuery (in Unix time).
exportKey - A combination of visitKey and exportTimeUsec that form a unique identifier for the row/hit.
For more information on how the real time export works or for advice on how your company can benefit from this new feature, please get in touch, we’d love to hear from you.