Fighting Spam in the the Wild, Wild West MMS

MMS has had quite the journey since its introduction in the UK back in the early 2000s. Largely considered an afterthought to SMS, MMS added the capability for subscribers to send images and videos from a device and have them appear uniformly within the same default messaging app of another subscriber. Not to be outdone, US carriers took that technology and turned it into a new marketing channel ensuring that brands could reach their customers. While groundbreaking at the time, 20 years later the flaws of its original design continue to test the networking capacity limits of carriers and aggregators.

One thing that hasn’t changed is its popularity amongst Americans. While attempts have been made over the years to introduce new channels such as RCS, WhatsApp, FB Messenger, and a plethora of other services, SMS and MMS continue to remain the “business” space that American consumers allow brands to reach out, resisting mixing “brand” messages with the “personal” space that apps like WhatsApp allow. Companies like Attentive, Braze, and Wunderkind know that, and continue to ensure that MMS is used as a tool in any brand’s marketing platform.

Unfortunately for consumers, while SMS and Voice get most of the attention these days from carriers, MMS has been left largely unchanged from a technical perspective. That also means the measures put in place to protect subscribers have also been left loosely guarded, and vary vastly from carrier to carrier, and aggregator to aggregator. For many brave souls that have ventured into the MMS space, this can feel a lot like the wild, wild west.

Scammers know this too, and as the clamps come down on SMS and Voice, many scammers have turned their attention to MMS. From sending larger chunks of text contained within an image, to reaching people with an e-mail address instead of a phone number, the number of permutations of spam continues to grow and thwart legacy spam techniques within the channel.

Luckily the KONTXT team here at RealNetworks has been working securing the MMS channels for a long time, and our long (and often lonely) journey continues to evolve our technology along with these changes in scams.

Take this example of perceptual hashing in non-compliant MMS messages:

Here we put together 17 different variations of the same MMS image sent by a sender using multiple phone numbers in an attempt to bypass legacy signature based filters.

Perceptual hashing is nothing new, having been around for over two decades and has been used in multiple forms, however its use and detection within carrier and aggregator networks for the MMS channel is nearly non-existent.

That’s where KONTXT for Messaging comes in.

As a counter to these tactics, KONTXT for Messaging has been deployed in production environments and uses perceptual hashing as the MMS message passes through the network to scan for images that match known spam or non-compliant categories.  The images can then be compared by our algorithm  to known bad image, and if a “fuzzy” match is discovered, the MMS message can be blocked or flagged for further investigation. In addition to identifying known malicious images, our perceptual hashing techniques can also detect images that have been altered, such as the example shown above, if it bears resemblance to the original spam image.

Fraud detection and message compliance won’t come overnight and is definitely not going to be easy, but our technology can help ensure that the MMS channel can keep up with the rest of the messaging channels.

Be sure to let us know if you’d like to schedule a demo of KONTXT for Messaging to see our MMS spam fighting technology in action.