Between the 28th October and the 3rd November, many people found that they could no longer carry out a search on Google. Despite entering a search query, nothing happened once they clicked “Enter”.
The problem first came to light in a Google Community Forums post, which as of today contains 394 posts and over 10,000 views. A Google Product Expert at Google quickly took the number of complaints seriously and made a post on the 30th October asking for users experiencing the bug to post with their Internet Explorer versions.
AJ, Community Manager, followed up on the 3rd November saying:
Fortunately, the problem has now been fixed, but it may be useful to be aware of the issue to help interpret any anomalies in your Google Analytics data.
How Many Searches Were Impacted?
It is almost impossible to determine how many users may have been affected, and we cannot determine whether users gave up trying to carry out a search or just changed to a different browser to complete the task. What we can do is summarize a few stats, and then throw a wild guess at it.
- As of April 2015 the browser share for Internet Explorer was sitting at a healthy 55.83%.
- The browser share for Internet Explorer 11 was 24.36%.
- The global market share of Google for desktop searches was 69.24% as of October 2015.
- Google processes over 3.5 billion searches per day.
- In October 2015, it was reported that Mobile Searches surpassed desktop searches for the first time.
- The problem continued for approximately seven days.
On the community forums, Product Manager Byee said that the problem specifically affected desktops and laptops, and many of the responses in the thread referenced Internet Explorer 11.:
Taking everything into account we can wildly speculate that we need to calculate the following:
24.36% browser share for IE 11 X 69.24% of the Global share X 3.5 billion searches X 7 days
Doing the math, this comes to a total of 4,132,381,680 searches that were potentially affected.
User bluequoll in the Google Product Forum thread said:
We, therefore, need to multiply the result by an arbitrary and entirely unknown figure to get the number of users affected. It doesn’t look like we can go any further with this method.
Ok, Let’s Check Google Analytics
One sure way to test if the problem had affected many people is to check Google Analytics data. For this, we can load a report that shows visitors by Web Browser type and compare time the bug was active with the preceding week. This is a snapshot of one website:
Well this doesn’t help that much either. The volumes of traffic using Internet Explorer 11 is relatively small for both weeks, but the number of users using Internet Explorer went up which may indicate a relatively limited number of people affected.
This bug affected relatively few people as an overall percentage, and as such had very little impact on search volumes.
However, we can easily surmise based on the activity on the Product Forums that the number of people affected was in the thousands, and possibly tens of thousands.