Facebook “went missing” from the Internet for a few hours on 4th October (UTC). All their services such as Facebook, WhatsApp, and Instagram were all unreachable to users. While normal services have resumed by the time you read this blog post, we'd like to explain the situation from our perspective, and how it might have affected you and your ad campaigns.
Facebook has noted some details on their published statement found on their blog. From our understanding, there was an issue with their border gateway protocol (BGP). This is the mechanism that allows autonomous systems (AS) to share routing information with each other. This is essential as it is what is required to create a path for information to reach its determined destination. Due to configuration changes there was an error in Facebook’s backbone routers effectively breaking the link between servers, causing them to go offline.
It is usually important to understand the root of the problem because this also allows us to estimate the impact of the issue. For example, Facebook has indicated that this was not a malicious attack. Because of this we do not have to worry about any confidential information leakages that can result of such attacks. Since it was a network traffic issue on backbone routers then we can assume that it affected the communication between their data centers. This means most—if not all—services stopped communicating with each other and so we all need to check our ad campaigns to see if there were any syncing issues that could affect performance. This does not necessarily mean that the numbers are inaccurate or that it could negatively affect your business. That being said, hygiene is important. When looking at numbers, it is important to keep this incident in mind
I am sure most of you will have checked your ad accounts to see if your active ad campaigns were affected. We did a check on all of our clients’ ad accounts under management the day after the outage, and noticed that there were no drastic changes to the performance trends—some campaigns maintain good results, while some don’t. It is important to know that even before the outage, Facebook ad results have always been volatile. A mere 5-hour incident would not have been the main cause in case your ad performance declined.
The SoMin platform did pick up some patterns, and our technical team confirmed its findings: Your Ad Account’s time zone could affect your ad performance amidst the Facebook outage
When Facebook went down, people were not able to see your ads during that period. As most of your campaigns were being charged by impressions, it meant that you were not being charged by Facebook when the outage happened.
Take a look at these screenshots from three ad accounts in three different time zones.
Singapore Time (GMT +8): Outage happened between 12am and 5am on 5th October.
The breakdown in time shows that ads were not delivered during the outage, and was even excluded from the reporting. Rest assured that you were not charged any ad spend by Facebook during those times. After Facebook’s services resumed, ads managed to regain impressions after a few hours.
Moscow Time (GMT +3): Outage happened between 7pm and 12am on 4th October.
As this might coincide with the peak period of Facebook users, there were some risks of losing potential results (e.g. purchases, leads, app installs, etc.) during these times. Then again, no amount was spent as well. It also took a couple hours for the ads to regain impressions after midnight.
US Pacific Time (GMT -7): Outage happened between 9am and 3pm on 4th October.
Similarly, no ad spend was charged during these times, but there were risks of losing potential results during this period, which includes the peak lunch time. For this particular client in the US, their campaign recovery was much slower, getting only single-digit impressions until the end of the day.
As the issue affected Facebook, there was no detectable problem on the SoMin platform. Your ad campaigns are still being optimized in the same manner. However, a drop in performance (e.g. higher CPA and CPC) were expected, especially if the Facebook outage happened during your audiences’ peak usage hours. This affected advertisers across the board, whether or not they are using an optimization tool on their campaigns.
We hope that this gives you a much deeper understanding on the Facebook issue. The SoMin team is always here to help, and feel free to write to us if you face any further problems with your ad campaigns.