Most online account hijacking capers are carried out using automated bots, but not all. In fact, some of the most effective and damaging heists result from targeted, carefully staged, manual attacks, a new study by Google shows.
Researchers at the search company recently reviewed manual account hijacking incidents involving users of various Google services from 2011 to 2014. For the study, the researchers looked at how criminals acquired a victim's login credentials to take over an account and how they attempted to exploit and monetize that access.
For the purposes of the study, the researchers defined a manual hijack as an incident where an attacker spends considerable time exploiting a single victim's account for financial gain. They discovered that such incidents are extremely rare. In fact, over the period of the study, the researchers observed an average of just nine incidents of manual account hijackings per million Google users per day.
However, manual account takeovers were far more damaging than automated ones. Often such attacks caused more problems and financial losses for victims than nontargeted mass account takeovers.
"These needle-in-a-haystack attacks are very challenging and represent an ongoing threat to Internet users," the researchers said in a report released this week.
Most of the account hijacking cases that the Google researchers reviewed appeared to originate from one of five countries: China, Ivory Coast, Malaysia, Nigeria, and South Africa.
In a majority of the cases, attackers used phishing emails and fake websites to get victims to part with the login credentials to their Google account. About 14% of the people who were lured to such websites were fooled into submitting their login information. Even patently fake sites managed to deceive about 3% of the people who visited them, the researchers found.
Interestingly, around 20% of the hijacked accounts were accessed in less than 30 minutes after the attackers obtained the login credentials. Once inside, the attackers spent an average of about three minutes profiling the account to see if it was worthy of further compromise.
Hackers typically used search terms like "wire transfer," "bank," and "investment" to determine an account's value quickly. For accounts they wanted to exploit further, the hackers quickly changed the password to lock out the legitimate user.
In many cases, the hackers would then attempt to scam the victim's contacts using the hijacked account. The most common scam consisted of an email describing a reasonably credible story of how the account owner got into a difficult situation, followed by a plea for money to help get out of the situation, the Google report said.
The research also showed hackers spending considerable time and effort to retain control of hijacked accounts long enough to be able to profit from them. In addition to locking out the victim and delaying recovery, hackers sometimes would set up a "doppelganger" account to divert communications from the victim's contact.
The research highlights the need for users to be vigilant about protecting their account, Google anti-abuse research lead Elie Burszlein said in a blog post. For instance, associating a backup phone number or a secondary email address with an email account is a good way to ensure an email provider has a way to get in touch in the event of a primary account takeover. Similarly, enabling two-factor authentication is another relatively simple way to mitigate risk of an account takeover, Burszlein said.
Previous studies have shown how surprisingly widespread and common account hijacking can be. In a survey of 300 people by researchers at Google and Carnegie Mellon University this year, 30% said at least one of their accounts had been hijacked. The number was more or less consistent with a Pew Research survey in 2013 that showed 21% of email account owners have had at least one of their accounts compromised.
This week's report is one of the first to examine the problem of manual account takeovers.