Telecoms fraud: Where are we now?
The telecoms industry is at something of a crossroads in regards to fraud threats, says Kevin Stanfield, is Analytics Product manager at MDS.
A survey from the Communications Fraud Control Agency (CFCA) shows an overall downward trend in estimated fraud losses from just over 5% in 2005 to 1.7% of global telecom revenues in 2015. However, the CFCA still estimates that global fraud loss came in at a staggering $38.1 billion dollars in 2015.
Whilst the slowdown in losses can be attributed to increased industry knowledge and experience, better computing power and faster strategic fraud tools, for many smaller operators and particularly new starters in the market, investment in fraud prevention is not always achievable due to perceived barriers to entry cost or fraud management simply not being in focus.
Typically, a new MVNO or service provider will leave the fraud monitoring in the hands of its hosting operator. But it is not a given that this operator will prioritise this task, or that they are investing suitably in future-proofing fraud threats.
Outsourcing fraud monitoring to a dedicated specialist should at least be considered to avoid the perfect storm of declining revenues and increased targeting by fraudsters.
Whilst almost 50% of respondents in the CFCA survey had fraud losses of less than 1%, they cannot afford to rest on their laurels. New threats are emerging that will inevitably gather pace, sophistication and impact, and it’s all happening in a landscape of large-scale digital transformation projects that are consuming focus and resource.
Evolving threats and challenges
There are a number of new and evolving fraud challenges facing today’s service providers.
OTT apps have increasingly flouted their ability to terminate calls within an application, decimating SMS message revenues and impacting traditional voice revenues. Whilst the caller expects to have the call quality associated with a mobile number, the called party can incur roaming and data charges higher than those associated with voice traffic. Both instances drive complaints and cost back towards the telco.
Premium devices are, and will continue to be, high-value items that are readily cash convertible, making them particularly attractive to fraudsters, with any associated airtime abuse merely the icing on the cake. More transactions are now conducted in non face-to-face environments, and the increasing competitive pressure to ship goods within hours puts greater pressure on fraud detection efforts.
Whilst it might seem strange to include in this context, digitalisation creates an environment that serves the digital native but also enables the digital fraudster, who typically will have refined new strategies long before they are recognised as such by DSPs.
Protect Your Customer:
Whilst telecom providers have now built defence mechanisms against customer subscription fraud, modern day fraud has evolved and fraudsters can now penetrate the customer base from outside the perimeters of the telco. Ransomware, vishing and account takeovers are particularly concerning as devices increasingly provide a conduit to personal data and access to financial institutions.
Know Your Customer:
Governments can and will impose fines that could bankrupt businesses if they fail to adhere to regulations regarding identifying the customer at the provisioning stage. The next development for telecom providers will not just be to securely link customer with device at the moment of purchase but towards ongoing monitoring and securing of that link.
Wangiri fraud has been around for some time already. Large numbers of customers are hit with a missed call and a percentage will, out of curiosity, call the number back, only to find they are calling an expensive, often overseas, premium rate number. Expect to see more of these types of technically advanced scams, including phishing, hacking and malware, artificially increasing traffic in the future.
Undoubtedly there is benefit in deploying algorithms against massive data sets but a pragmatic approach needs to be followed. Self-learning customer profiles will increasingly be utilised with near real-time alarming on exceptions to trends in the normal customer profile.
Utilising identified fraud cases to constantly amend and improve profiling and finding hidden links between applications are far better handled by AI, than people using traditional reactive alarm/rules based systems.
Do more for less
The industry is certainly in for a rollercoaster ride; known fraud types unfortunately cannot be excluded from continued fraud monitoring and the new threats and exposures that emerge require constant attention.
In a telecom landscape, where there is huge price pressure on both the telecom operators and their suppliers, more cost-effective methods need to be found to meet the challenge of preventing more types of fraud for less money.
Bundling fraud fighting efforts could be the answer, as outsourcing of monitoring brings the benefit of scaling where investments in hardware, skilled resources and knowledge can be shared. Knowledge sharing amongst communities will add further value.
The author of this blog is Kevin Stanfield, is Analytics Product manager at MDS
Comment on this article below or via Twitter: @ VanillaPlus OR @jcvplus