Facebook shares were down 1.4% to $170.96 on Thursday as the company has also become the target of a criminal probe into its data-sharing deals, according to The New York Times.
On Thursday early afternoon, Facebook tweeted that a "server configuration change" was responsible for the problems; it said the issue had been resolved and that its systems "were recovering."
The outage problems across Facebook, Instagram and WhatsApp, which together claim billions of global users, began on Wednesday afternoon E.T. According to BusinessWeek, Facebook generates an estimated $189 million in daily revenue and the company is considering offering refunds to its advertisers because of the outage.
Facebook posted two tweets on Wednesday saying it was aware of the issue and working to resolve it as soon as possible, with the most recent tweet coming at 3:03 p.m. ET.
But on Thursday, DownDetector.com continued to show access problems to Facebook in the northeast United States, northern Europe and parts of South America, while similar geographies were experiencing problems for Instagram. WhatsApp was out in many parts of South America and northern Europe, according to DownDetector.
People around the world started reporting the outage on Twitter on Wednesday with the hashtag #facebookdown.
Facebook said in its second tweet on Wednesday that it didn't believe the issues were related to a DDOS (distributed denial of service) attack, a common way of bringing down websites by overwhelming them with requests.
At 1:49 pm. ET on Wednesday, Facebook first acknowledged the problems, saying that "we're aware that some people are currently having trouble accessing the Facebook family of apps."
Outage maps on DownDetector.com showed difficulties on Wednesday for both Facebook and Instagram being reported in the U.S., Mexico, South America and Europe. According to DownDetector, outages on Facebook were first reported starting at 11:56 a.m. ET, while those for Instagram began at 12:01 p.m. ET.
Facebook users reported getting a message stating that the site was doing "required maintenance" and that service would be restored in a few minutes, which never happened.