Spam-A-Lot: How Spam Filters Work
When half the world’s total email volume is spam, how do our internet service providers separate the junk from the genuine?
A decade ago, over 90% of global email volumes consisted of spam. Today, however, it’s estimated that levels of email spam are approaching the psychologically important 50% threshold. While this is still a long way from the halcyon days of 2001 when only one in twenty emails was categorized as junk mail, it appears that the war against spam is slowly being won.
A Thank You To Spam Filters
From the perspective of individual users, inboxes have been relatively spam-free for a while. While an occasional message can still slip through, it’s more likely to be a marketing communication than an invitation to purchase performance aids or view photos of scantily-clad ladies. Those emails are still being sent in their billions, but a relentless assault against junk messages has ensured that more and more spam is headed off before it ever reaches our webmail or inboxes.
Gmail
This onslaught has been driven by email and internet service providers, determined to secure email’s future following post-millennial fears that the entire industry could be killed off by an unrelenting tide of junk messages. Google’s Gmail markets itself as the industry leader in correctly separating wheat from chaff, following an inglorious period where its email service was the favored tool of junk mail senders. Google’s neural networking approach has helped to customize spam settings for individual users, while its proprietary Postmaster Tools allow approved agencies to bulk-send messages and monitor the rate at which they’re marked as spam by Gmail’s filters.
Internet Service Providers
ISPs have also adopted a more proactive approach to spam filtering than simply providing a button users can click when a rogue message arrives. Certain keywords and terms will be blocked as a matter of routine, while accounts associated with messages frequently marked as junk are often automatically blacklisted These lists are typically regulated by impartial bodies like Spamhaus, a non-profit pan-European organization founded in 1998. The reputational blocklists produced by Spamhaus’s forensic specialists and IT engineers are used to protect almost three billion mailboxes, including most of the world’s internet and email providers.
Government Aid
A third front has been opened against spammers by antivirus firms who have developed increasingly sophisticated software to tackle botnets – the armies of infected computers that distribute huge numbers of spam messages. The police have also become more heavily involved in botnet takedowns, helped by increasing cooperation and communication among ISPs throughout Europe. Much of the junk mail sent nowadays involves counterfeit goods being promoted by Vietnamese, Russian and Chinese spammers, with poor grammar and amateurish presentation making these messages relatively easy to detect. The vast majority of spam emails are less than 2KB in size, which is another tell-tale clue that can be recognized by automated filter algorithms, while today’s comprehensive lists of spam sites ensure the all-important hyperlinks carried in junk messages can be instantly identified and targeted.
A Downward Trend
Although some spammers have adopted workarounds such as embedded email URLs linking to compromised legitimate websites (which instantly redirect to malware sites), spam has undoubtedly entered the decline stage of its life cycle. It relies on near-identical mass mailings to achieve economies of scale, which are increasingly ineffective as spam filters become evermore refined and personalized to individual users. A swathe of reports last year concluded that spam is gradually being supplanted by phishing and other forms of cybercrime, such as malware. Online criminality is metamorphosing into new areas rather than being defeated, but at least decreasing spam levels are good news for anyone who regularly communicates via email. The next time your inbox pings, it’s very likely to be something you actually want to read.