Deceptive Customer --- Deceptive Bottomlines

The exact figures are not available but experts believe that telecom fraud

across the globe costs telecommunication companies between $4 and $8 billion

every year. This fraud eats away into the final profits of telecom companies by

as much as three to five percent. One starts appreciating the magnitude of the

problem when one realizes that most of these companies have annual returns of

less than ten percent.

Fraud

in telecom networks has been a problem that has always plagued the bottom line

of operators. The lure of easy money turns many a man into a fraud. There are

many ways to defraud and abuse telecom networks–especially wireless networks.

Ingenuity and innovation are the hallmarks of this tribe of modern-day bandits.

Telecom companies have a long history of fighting fraud but the imaginative

methods employed by frauds call for continuous innovation of the solutions

deployed to check fraud. Telecom fraud may have grown in sophistication from the

days when pay phones were jammed with foreign coins and telephone wires were

tapped. But in some ways, telecom fraud is as simple and low-tech as ever. Fraud

used to be accompanied by great risk of detection but little risk of

prosecution; now perpetrators can avoid both. Today, as rapid advances in

technology have reduced the risk of detection and increased the number of

exploitable loopholes, telecom fraud is more attractive than ever.

Unaware or Ignorant?

In spite of the fact that many telecom companies worldwide have lost

significant amounts of money due to fraudulent activity in their networks, a

large number of operators are still not addressing this crucial issue. In many

cases, they still feel that fraud does not exist. Even though one wishes that

this be the case, this is never true. Losses due to fraud often get swept under

the carpet as bad debt. A recent study has proved that the portion of revenue

lost due to fraud could form forty to fifty percent of the bad debt component!

Another aspect is the belief that networks based on digital technologies are

secure. The networks that were rolled out earlier used analog technologies that

had several technical loopholes. Frauds exploited these opportunities to make

money. The advent of digital technologies like GSM put a damper on most of the

technical frauds. Perhaps this could be the reason why many GSM operators feel

that they are safe from fraud. However, innovative frauds soon managed to find

simple, non-technical ways to continue their nefarious activities even in

technically advanced digital networks. Catching the offenders is both difficult

and expensive. Most telecom providers prefer to cut their losses by removing

service rather than spending money on the uncertain process of criminal

prosecution.

How does Fraud Happen?

There are many methods employed by unscrupulous elements to defraud telecom

networks. Technical frauds like cloning and tumbling were the major cause of

worry for analog network operators. Such technical frauds cannot be perpetrated

in digital networks (even though there are unconfirmed reports that cloning

could be possible in GSM). A rollout of digital networks, in countries like

India, has resulted in the evolution of non-technical frauds.

Analog phones provide a striking example of how a technology can present an

open invitation to fraud. With relatively inexpensive and readily available

equipment, a perpetrator can monitor signals between analog phones and the

network; picking off phone identity codes and reprogramming or cloning stolen

phones to imitate legitimate identities. A cloner can make thousands of dollars

renting the phone (or selling calls) before the fraud is discovered and the

identity shut down. In that time, hundreds of thousands of dollars in call

revenue may be lost. The prevalence of analog networks in the US may partly

explain the technology-driven nature of many US providers’ responses to fraud

prevention: sophisticated, real-time detection systems that analyze network

traffic. For example, such systems can flag as fraud an identity making a call

from one location and two minutes later making a call from another location on

the other side of the country.

Digital technology is more robust than analog. Nevertheless, GSM service

providers are vulnerable to people who use airtime with no intention of paying

for it. They simply move on to the next network and/or change their identities

when their service is removed. Extensive background checks on new customers are

often impractical and in many countries almost impossible. Iceland, for one, has

online links to a government registry of householders, but data protection

regulation and consumer pressure in many other countries rule out such

investigations. Furthermore, if service providers can’t give customers

near-instant connection, they risk losing business. So providers who have

invested heavily in GSM networks have a high exposure to risk, because they

cannot easily check whether the customers they are signing up are genuine. This

acquires gigantic proportions in a roaming scenario due to the delay in

accessing call detail records and subsequent detection.

Common Types of Fraud

• Subscription fraud:
  
  Subscription fraud is the most damaging amongst the non-technical frauds.
  
  The methodology adopted here is simple with the fraud obtaining a connection
  
  from the company using the normal, accepted procedure of the company. The
  
  fraud has no intention to pay, and rings up huge bills in very short time
  
  frames. The company comes to know of such high usage only after sometime
  
  (Orange, the cellular service provider in Mumbai, checks customer usage
  
  patterns every week), and by then it becomes impossible to trace the
  
  fraudulent subscriber. Ultimately, all the money that was to be recovered
  
  from this subscriber will have to be written off.

• Call sell
  
  fraud: A variation of subscription fraud, here, the fraud misuses the
  
  call-forwarding feature of the networks and uses his connection to set up
  
  multiple, simultaneous calls for his "clients". The
  
  "clients" pay a less amount of money than they would normally pay
  
  the company. The fraud can, of course, afford to offer such
  
  "subsidy", as he has no intention of ever paying any money. By the
  
  time the bills are raised and the company comes around to collect the money,
  
  the fraud would have made a significant amount of money and got away. Call
  
  selling is one of the major contributors to losses due to subscription fraud
  
  and could account for more than fifty percent of the total fraud loss.

• Internal fraud:
  
  Employees within the operator or associates of the company aid outsiders in
  
  defrauding the network in this kind of fraud. Employees have access to all
  
  parts of the network and are sufficiently knowledgeable to know where to
  
  play dirty. This is one of the difficult to monitor frauds, as employees of
  
  the companies are themselves involved in the act.

• Social
  
  engineering: Here, thieves con someone into giving up their telephone
  
  calling card number by claiming to be from a law enforcement agency, the
  
  phone company or any number of other organizations. Social engineers also
  
  attempt to persuade company telephone operators to give them an
  
  "outside" line, even though they’re calling in on an outside
  
  line.

• Shoulder
  
  surfing: It is an unsophisticated and simple method of fraud where the
  
  thieves look over someone’s shoulder and memorize the calling card number
  
  as it is punched into the phone. More sophisticated hackers sit at airport
  
  terminals, with telephoto-equipped video cameras, and collect bunches of
  
  calling card numbers from afar.

• Toll free
  
  fraud: This is becoming a growing problem as the use of 800 and 888 toll
  
  free numbers is increasing. Thieves call into a company’s toll free number
  
  and hack around the phone system until they get connected to an outside
  
  line. This only works when a computer and not a person answers the line.

Operators in new markets such as India are gullible to such

practices, as the knowledge about fraud fighting is just being developed. Price

Waterhouse Coopers (PWC) recently conducted a study and reported that Asian

mobile operators lose about thirty percent of their revenue to fraud and bad

debt. There have been several reports about huge losses by Indian cellular

operators and it is assumed that operators are losing about ten percent of their

total revenue due to fraudulent activity in their networks. This is quite high

as compared to the worldwide average of three to four percent.

Detection Strategies

Europe has led the world in prepaid cards and services on GSM

networks. Prepaid cards are also popular in the US. European providers believe

that prepaid services reduce their exposure to risk and maximize profitability.



However, the primary motivation for prepaid cards and services is growing market
share, not preventing fraud. The new payment options are aimed at attracting

transient customers who cannot afford high connection fees or who do not want

subscription services.

But

encouraging more people to go mobile may result in more fraud unless steps are

taken to monitor and control prepaid cards and services very tightly. Those

steps need to be taken at the interface between the network that captures call

information and the operational systems that process it. Clearly, the longer it

takes call records to reach the billing system, the greater a provider’s

exposure to risk. If, for example, a customer has a $50 prepaid card, the

provider needs to be able to disable the card the moment that the customer has

used $50 of airtime–not a week or even a day later, when the customer may have

made $5 to $20 worth of free calls. Call records need to be released to a system

that can monitor spending as close to real time as possible, whether this is an

intelligent network platform or the billing system itself. Todays best companies

can process call traffic information for possible fraudulent use within fifteen

minutes of receiving it, reducing their risk factor on each phone to that

quarter of an hour–a considerable blow to those intent on fraud.

Good management processes will also stop insider fraud in its

tracks. Some providers have experienced large-scale fraud when detection systems

are out of service, e.g., during a switch software upgrade that delays call

records from being released to administrative systems for three or four weeks.

Such situations can be avoided if upgrades are planned and tightly controlled,

allowing the ability to roll back the change process if all does not go

smoothly.

Airtime fraud is a concern for service providers, and there

is a strong link between stolen equipment and attempts to use that equipment for

free. Although providers may not feel responsible for the theft of handsets

(many of them do not sell equipment, only network services), they are caught in

a difficult position. If they do not appear to be taking steps to combat theft,

they are viewed as contributing to the problem. But the steps they need to take

may conflict with their own interest–monitoring spending as close to real time

as possible.

A model has been developed for the GSM community where, when

a phone is stolen or taken out of service, the provider informs a central

register operated by the GSM Memorandum of Understanding (MoU) and the

information is distributed to all the MoU signatories. Ideally, each provider’s

administrative systems will evaluate the equipment identity from call records

and check it against the register to determine whether the equipment status is

valid. There are two downsides to this process: It imposes a huge transaction

overhead on the administrative systems, slowing down call data processing; and

no manufacturer can guarantee that each handset has a unique identification

code. In India, implementing this is even more difficult since more subscribers

purchase phones from the grey market, than from the company.

The high demand for mobile services in parts of the world is

also fueling equipment theft. The rapid growth in call volume in Asia has made

it difficult for providers to monitor usage, resulting in the need for large

cash deposits by subscribers as well as expensive service charges. However, the

argument is circular. The higher the value of the service, the greater the

incentive for fraud. Prepaid cards and services in this market will only prevent

fraud if the monitoring systems are in place to enforce them.

Today, many telecom service providers are still unaware of,

or unresponsive to, the level of exposure to serious business risk that telecom

fraud represents. They need to institute management processes and administrative

systems to limit this risk. This means integrating a range of technologies–switches,

intelligent network platforms, billing systems and fraud systems–in ways that

suit their particular business requirements.

There are systems for fraud detection that use artificial

intelligence capabilities to detect ‘abnormal’ usage behavior, which

indicates that fraud is being committed. The system creates a profile of each

individual subscriber’s ‘normal usage behavior.’ This profile is based on

an analysis of Call Detail Records received from the switch. These individual

usage profiles are continuously updated and analyzed on a call-by-call basis.

Behavioral deviations fall into different patterns, which are indicative of

specific types of fraudulent activity. Any deviation from the normal usage

behavior of an individual subscriber is closely monitored. When a specific type

of fraudulent activity is determined to be probable, a prioritized fraud alert

is generated for review by telecom fraud investigators. In addition to

individual subscriber behavior monitoring, these systems have the ability to

monitor the behavior of any network-related entity such as lines, switches, etc.

This ‘network entity’ monitoring capability significantly expands the scope

of fraudulent activities that can be detected by the system.

Although such systems help to combat fraud to some extent,

providers need to understand fraud prevention not merely as a technological

challenge, but as a risk management issue and a strong aid for turning bad debt

into profit.

The Indian Scenario

In December, the CBI unveiled a Rs 400-crore telecom scam

where an operator had leased international lines from VSNL on the pretext of

software exports. Instead, he was using these, as well as over twenty MTNL lines

in Mumbai and Delhi, to provide people with lower cost telephone calls to the US

using the call back facility.

It is very clear from the above that the task of curbing

fraud in telecom networks is a daunting one. The first and foremost task is for

operators to take a hard look at what they are writing off as bad debts and

separate the loss due to fraud from this. Unless fraud is acknowledged, it will

never be fought. Unfortunately, many operators, especially in developing markets

like India, are on an overdrive to attract more and more subscribers to the

network. Some operators are just starting to realize that more subscribers could

as well mean more fraud generation and not necessarily revenue. The next item on

the agenda is to develop a corporate fraud fighting strategy that defines all

the policies and the approach of the organization to this issue. All departments

within the organization will have to be committed to this strategy, as this is a

critical requirement for its very success. The strategy should be complete and

should address all issues like policies to be adopted while signing on new

subscribers, approach to employees, defining security aspects within the

organization, procedures to be adopted when fraud is detected, etc.

A fraud management system has then to be implemented to

detect and curb the fraudulent activity in the network. This is typically a

sophisticated software solution that intelligently profiles each subscriber and

alerts operators of possible fraudulent activity in a timely fashion. It is very

critical to understand the peculiarities of the market and the behavior of

frauds within the network before the solution is selected. An intelligent

selection has to be made after a thorough study of the various options. This

will lead to a steep decline in loss due to fraud.

A country like India does not have a strong system for

establishing personal identity and credit histories or ratings are also not

available. These are very crucial aspects in fraud fighting as the lack of such

systems makes it a lot easier for frauds to have lucrative operations. So, a

fraud management system, to be effective in India, must address these issues of

checking identity and providing credit ratings. Cellular operators like BPL-US

West are initiating a group where all members would contribute the names,

identity and details of defaulters for the reference of other members. Other key

issues are scalability, platform independence for the software, real-time

operation, user friendliness, etc. Support from the government is also

important, and unless the government has a tight procedure for punishment in

place, more and more companies would continue to write off the fraud losses as

bad debts instead of pursuing the offenders in the court of law.

The bottom line is that fraud accounts for a large amount of

money being lost by telecom companies and is an issue that warrants close

attention. It is very important that it be curbed to improve the profitability

of telecom companies and to ensure smooth, hassle-free operations. Ultimately,

it is the responsibility of operators to provide fraud-free services to genuine

customers, through the implementation of efficient fraud management policies and

systems.

The article is written under the guidance of Professor V

Sridhar, by Deven Shah, Vishal Kaul, Vivek Bajpai and Vivek Shah, all second

year MBA students at IIM Lucknow