maxpelle.me

The messaging world is seeing big changes. Perhaps the biggest is the rapid rise of A2P (Application to Person) messaging, a market that was valued almost $13bn in 2015, and is forecast to hit almost $60 billion by 2020. All that traffic is giving rise to another major change — an increase in fraudulent messaging behavior. This includes messages sent via “grey routes,” set up to avoid fees to network operators, or to hide the origins of spam or phishing texts. Lastly, the advent of Rich Communications Services (RCS), supported by Google, will lead to features such as chatbot interactivity, location-based services and payments, and richer branding — all creating a more dynamic messaging experience.

To date, the mobile industry has done a good job maintaining a “clean” ecosystem, one without a lot of spam, and where 90 percent of text messages sent are read within three minutes. That’s because of aggressive policies on the part of network operators to keep bad actors off their systems, and the use of firewalls and filters that snag much-unwanted traffic. “In the US market, there are growing complaints from developers and enterprises about their legitimate messages being blocked with no reason, creating a damage to the ecosystem as well.”

But as the text world becomes more complex, the system is starting to spring some leaks. Mobile Network firewalls primarily use an SS7 cloud connection to other mobile networks. They monitor voice and data traffic based on known, static identified threats, usually via a shared, open-source database, but with the ecosystem becoming richer and more complex those are not completely effective.

To maintain the integrity of mobile messages, what’s needed is a more sophisticated approach to understanding the intent of the message. One that learns from what it discovers on the network, conducts real-time message traffic analysis and applies policy controls to grey channels and other network interlopers.

Here at RealNetworks, we’ve taken on that challenge. We recently announced Kontxt (sounds like “context”), a solution designed to give network operators the insight and policy control tools they need to classify messages based on customer priorities, as well as control phishing and spam.

What do we mean by message “classification?” For example, companies sending two-factor authentication codes or customer support messages (such as changes to a flight) need that message to go through instantly — and have the budget to support that special service. Senders of bulk promotion messages, on the other hand, may be fine with a more relaxed delivery schedule and would be happy to pay a little less for slower delivery.

Spam or fraudulent messages, of course, need to be blocked right out of the gate. Similarly, if a sender is attempting to use a grey route to send a message, Kontxt can flag that.

The figure below illustrates how Kontxt relates to existing firewalls and anti-spam filters.

Kontxt will transform the messaging market. Eventually, nearly all mature markets offer differentiated services tailored to customers with different service needs and different budgets. Think about airlines, which evolved to offer first class, business class, premium economy, discount economy. Or take courier delivery services, which have evolved to include next day delivery, two-day delivery, three-day delivery, and so on — each with a different price structure. We believe strongly that mobile messaging is ready for such an approach, and that Kontxt enables it.

So how does Kontxt work? It uses highly advanced machine learning algorithms to unpack the semantics of a message so network operators can understand in which class it belongs including if it is malicious.

In my next post, I’ll outline the different classes of messages. Classes may differ depending on region and language so it’s a deep cultural problem that we’re working on.

The intelligence built into Kontxt, meanwhile, allows it to greatly reduce the number of “false positives,” instances where network filters mistake legitimate messages for spam or phishing attempts and block them. That makes for a healthier ecosystem with the value created for enterprises and brands as well as for consumers on the other side of the value chain.

In addition, Kontxt generates value for network operators, that can benefit in three ways:

1. Better quality, because Kontxt discovers different use cases and shows operators the value in classifying different types of messages so that they can handle them in a way that is best for each message or bundle of messages.

2. Increased revenue, because network operators will be able to show their customers the added value of messages managed with Kontxt, and develop a more sophisticated value chain.

3. Deeper insights, because Kontxt shows operators more clearly how their network is being used, where its weak points might be, and how to better serve customers while blocking fraudsters.

We’re enthusiastic about where mobile messaging is headed and see all kinds of bright innovations taking shape in the space. Do you agree? Do you have a different take? I’d love to know what you think.