Google Analytics is Down
As Google Analytics has become more ubiquitous, disruptions in service tend to be felt far and wide.
Google Analytics Outages
There have been a handful of widespread outages since Google Analytics was first rolled out. They have typically been rare and brief, but they have been troubling for users nonetheless.
Outages have affected users in two ways:
- the Google Analytics interface has been slow or inaccessible
- traffic data has been lost
Some users report that page load times on their own sites became slower, too. Fortunately, outages do not affect the websites of users who have set up the code correctly.
Since the Google Analytics interface is hosted on Google's servers, an outage can affect your ability to log into your accounts. Or, it may affect the load time of the report pages.
If there is a widespread outage, __utm.gif requests won't be served or logged, and data for that time period will be irretrievably lost. Historically, this has been the exception. Typically, data is delayed but not lost.
Google Analytics' Reliability
Outages like these raise the question of Google Analytics' reliability. Or at least of the wisdom of relying too heavily on it.
Google has one of the largest and most redundant infrastructures in the world. Take the most recent outage as an example. Comparatively speaking, it was a huge one. It lasted one and a half hours. Let's say that three of those happened every year. That hasn't been their track record, but if it were, it would represent a 99.95% up-time rate.
Let's carry that further and assume that all the data was lost for the entirety of those three outages. (Data loss has been unusual during outages.) A .05% data loss would be a statistically insignificant (albeit annoying) blip. The integrity of your data would be unquestioned.
Google's up-time rate is unmatched by enterprise companies' internal infrastructures.
Avoiding Problems
In the end, Google Analytics is free software. An incredibly high up-time rate is a lot to get without paying anybody money.
However, if outages and delays are too much to risk, you can bring the same data in-house by using Urchin Software. Urchin was the foundation for Google Analytics, so it uses the same code and basic methodology, but it is installed on your own servers. It can be used concurrently as a backup. That puts you in charge of your own reliability. The only charge is a flat one-time licensing fee. No monthly usage fees or upsells.
It's also worth considering whether being unable to access reports for an hour or two is actually an issue, since all analysis is post-analysis and relies on historical data to be very actionable. For many companies, GA outages may in fact be much ado about nothing.
