2 The situation in the UK
2.1 Reducing the number of nuisance calls made to UK recipients
2.1.1 The current position
As a starting point, we need baseline estimates of the level of nuisance call attempts that are made. Because of much cheaper calls and call centre technology, this level is much higher than five or more years ago. Unfortunately, an actual number is hard to pin down, though we are sure that it exceeds 5 billion. In Annex C we collect such relevant information as we have found. We use 2016 as our base year, because it is the most recent complete calendar year; it happens to predate most of the network suppression activity which we are now starting to see.
An important fact is that the distribution of nuisance calls received is very uneven across the public. We summarise in Annex C the Ofcom surveys that help to demonstrate this. Most relevant among them are the landline nuisance call surveys: starting in 2013, Ofcom has commissioned annual surveys in which around 800 diarists (a representative sample) record details of all the nuisance calls they receive on their landline over four weeks. While the composition of the nuisance calls has varied somewhat from year to year, their numbers and concentration have remained remarkably constant (with very few significant differences at the 99% level being noted from year to year). We have therefore combined these survey findings over the five years 2013-2017 to arrive at the distribution illustrated in Figure 3  .
Figure 3: Numbers of nuisance calls received in four weeks, 2013-2017
Source: Ofcom landline nuisance call surveys
To understand Figure 3, imagine all Ofcom’s diarists lined up from left to right in order of how many nuisance calls they receive in four weeks, each holding a placard showing how many they receive (so there are many people at the left with placards saying ‘0’, and a few at the far right with placards saying ‘84’ or ‘100’). The numbers on their placards are added up, and the total divided by 10 to show how many nuisance calls constitute a decile. Suppose the total is 5,000, so the resulting decile size is 500. Then, starting at the left, the numbers on the placards are added up until they make 500; all the users so far are members of the first nuisance call decile. Then we continue with the next set of placards until we reach 500 again; the people concerned are members of the second nuisance call decile; and so on  .
Figure 3 shows that almost half the landline users receive 4 or fewer nuisance calls in four weeks, with an average of 1.5 calls - a level which may be considered tolerable and be overlooked; while at least a fifth receive 12 or more nuisance calls in four weeks – a level which may well be felt as a problem. A twelfth may be thought of as having a serious problem, with more than 20 nuisance calls in four weeks, receiving between them 30% of all nuisance calls to landline networks. Later we shall look at what we call the “worst affected” group, defined as the people contributing to the highest decile. These people get 35 or more nuisance calls in four weeks, with an average of 46 calls each, or 6 times the average for all diarists of 7.6. This group accounts for under 2% of adults who receive calls on landlines.
A further step is to consider how this level would change over the next few years, independently of actions taken with the aim of reducing harm. This is discussed in Annex D. We are aware of influences in both directions, but cannot yet assess their relative strength. We therefore assume that, independent of harm-reducing actions, the level of calling will remain roughly constant.
A major plank of the regulatory framework to counter nuisance calls in the UK is the Telephone Preference Service ( TPS). The Privacy and Electronic Communications Regulations ( PECR), enforced by ICO, prohibit unsolicited live telemarketing calls to phone numbers registered with the TPS (and prohibit recorded telemarketing calls and text messages without specific consent).
Figure 4: TPS registrations, 2016-2017
Source: Telephone Preference Service
Figure 5: TPS registrations of mobile numbers, 2016-2017
Source: Telephone Preference Service
Figure 4 and Figure 5 show how TPS registrations continue to grow, though with mobile numbers still far behind landlines  . Annex B provides more detail on the current regulatory and enforcement regime; this is mainly shared between ICO and Ofcom, with the CMRU playing a role for claims management companies.
2.1.2 Actions intended to reduce the number of nuisance calls being made
1. Network or individual call suppression  . If sufficiently widespread, in principle call suppression could reduce the likelihood of call attempts being answered, thereby raising the unit cost of answered calls and ultimately the cost of commercially successful calls. However, unit costs of calls are currently so low that take-up of call suppression would have to be very high to have much effect. As long as network suppression is offered on an opt-in basis, the large proportion  of people who have few nuisance calls, or are not troubled by them enough to think about avoiding them, are unlikely to opt in. So for the time being, we do not expect suppression to have a material effect on calls made (though it can have a material effect on calls received, which we discuss below).
2. Consumer behaviour in aggregate. Calling levels will be affected by the perceived likelihood of answered call attempts resulting in success (maybe a “sale”, or a step towards a “sale” such as the called party agreeing to a follow-up call). This in turn depends on the receptiveness of the called party. Public education on how to handle these calls may have some effect here, particularly on specific mass scam calls (e.g. the “Microsoft support” scam) – though others seem to spring up to take their place. We suspect that this effect is no more than a few percent overall, as repeated publicity campaigns to date have not made noticeable inroads on the problem.
3. Consumer behaviour at the individual level. Consumers can try to avoid being targeted by nuisance calls by:
a. Registering with the TPS. In 2014 a randomised control trial commissioned by Ofcom showed that registering with TPS cut nuisance calling to the registered individuals by around a third. This finding pointed towards around a third of nuisance calling being by companies who comply with the rules, a figure which is confirmed by the data assembled by trueCall in Annex A. Greater compliance could lead to this figure rising, but we suggest it is more likely to fall as TPS registration rises further (from its now high level, which we estimate at over 75% of households) and the number of unregistered prospects becomes so low as to make compliant telemarketing to the residue barely worthwhile  .
b. Taking care with their personal data: for example, going ex-directory, or avoiding sharing their phone number when entering competitions or making online purchases. This is standard advice, which sounds good, and is in line with people’s expressed concerns  , but evidence on its effectiveness is lacking  .
4. The regulatory regime and its enforcement. In principle this affects both making calls and sourcing lists for targeting calls. In Annex B we present data on the enforcement actions taken by the main relevant regulators, which range from offering advice to imposing fines. Compliance cost and reduction in opportunities (for scrupulous companies) or the deterrent prospect of “naming and shaming” (for companies with a reputation to protect) and fines (for less scrupulous companies) may encourage moves away from telephone marketing towards alternative marketing channels. Those who do get caught stop making illegal calls (at least until they set up “phoenix” operations), but the numbers of calls that are thereby prevented, though large, can be only a tiny proportion of the total. For instance, Figure 28 shows 8 million nuisance calls having been “caught” by ICO during the period 2015-2017; on our estimates this is under 1 in 2,000 of those being made. Multiplier effects through deterrence (which are claimed but unquantified) would have to be implausibly large for this to make a detectable difference.
Looking from another angle, regulators can be resourced to pursue only a limited number of cases. Figure 20 (based on trueCall data) shows that even the 1,000 most used calling numbers generate under a third of all nuisance calls.
The Fair Telecoms Campaign has long advocated more activity from sectoral regulators (such as those for claims management companies, financial services or energy providers) to help prevent nuisance calling. The case made has many merits. However, this approach requires specific legislative or regulatory change in each sector and to date, despite promises, little progress has been made  . We return to it in the Scottish context.
Unfortunately, the low probability of getting caught (especially if operating from outside the UK), and the delay between the offence and any consequences, greatly dilute the positive effects of regulation. Stronger regulation and enforcement may also have a counter-productive effect, of encouraging unscrupulous and criminal elements to move abroad or otherwise evade the regime. Similarly, a lower availability of qualified lists (which may be of dubious legality) for targeted calling can encourage poorly targeted or completely untargeted calling, which is arguably worse.
By saying that the identifiable effects of enforcement efforts on nuisance calling are low, we do not mean to suggest that these efforts are wasted. They have a clear value in helping to uphold the rule of law and maintain societal standards. ICO are optimistic that the promised personal responsibility of directors for the payment of fines will have significant impact, and (depending on the interpretation of “consent”) the UK implementation in 2018 of the General Data Protection Regulation may put live voice calls on the same footing as recorded calls. Regulatory influence and encouragement may also lead to constructive actions by others – this appears to be the case with Ofcom’s voluntary “nuisance calls MoU group” of network operators.
5. Provision by network operators of assured Calling Line Identity ( CLI). There have already been some improvements in this area, with BT providing full CLI on international calls from December 2014, and a manual call tracing system (via Ofcom) in place since 2014. But widespread use of Voice over IP technology has made number spoofing very easy, and nuisance calls are now more likely than not to arrive with spoofed and therefore untraceable numbers. Assured CLI provision, if and when successful, could eliminate number spoofing, thereby improving traceability, which could deter non-compliant nuisance calling. However, this development is not in prospect for the UK for several years  , so we do not aim to estimate its effect. Shorter term, new General Conditions on CLI will take effect in October 2018, with new Guidelines, and other suggestions have been made for using CLI differently. All this, if properly implemented and enforced, should both make call tracing easier, and help consumers to see who is calling.
6. Interconnect agreements. Ofcom and some network operators are working to stop the origination of mass nuisance calls at source, via clauses in interconnect agreements that would require each link in an interconnection chain to prevent such calls from entering their network. This appears to be another major challenge which cannot be expected to deliver results short-term, though long term it has the potential to be very effective, especially combined with CLI assurance  . Different operators however have different commercial incentives related to nuisance calls, with some benefiting from revenues for call origination, call conveyance or call termination  , possibly without the costs associated with unhappy customers.
7. Care with announcements. Observers have pointed out that government or major business scheme announcements (for example, to encourage energy efficiency, or pension freedoms) often stimulate streams of nuisance calls. Efforts could certainly be made to minimise disreputable exploitation of these schemes. However, traders are naturally alive to opportunities and it is hard to see such efforts having much effect if the public are to be properly informed about the schemes.
In summary, there appears to be little that can be done short-term that will clearly have the effect of preventing nuisance calls from being made, without risk of counter-productive side-effects. For actors with variable or dubious compliance, prompter enforcement could have a stronger deterrent effect than current long-drawn-out procedures; more transparency during investigations (“naming and shaming”) may also be effective. Longer term, sectoral regulation may help, and widespread network suppression, with reliable CLI and inter-operator contractual provisions, could lead to significant reductions. However, it may well be that commercial forces reduce nuisance calling sooner than that, as alternative marketing channels prove more cost-effective.
Similar remarks apply to scam calls as well as to other nuisance calls, though here the relevant authorities are the police rather than civil regulators. Penalties are more severe, including imprisonment, but the probability of being caught is even more remote. We suspect that to bring about big reductions in scam phone call origination, speaking to strangers who phone would need to become socially unacceptable, and viewed as unwise, in the same way as admitting unknown doorstep callers into the house. This would be an extreme position with undesirable side-effects, but if it came about, scam merchants might well move to another channel.
2.2 Preventing recipients from receiving nuisance calls made to them
Once nuisance calls have been made, it may be technically possible to identify and suppress them before they bother customers  . Figure 6 outlines the main features of currently available relevant technologies.
The overall effectiveness of any call suppression technology in reducing harm from nuisance calls depends on its availability and take-up as well as on its technical features. These are affected by wishing to minimise drawbacks to consumers  of using these technologies:
- Setting them up in the first place can be difficult for end users, especially if elderly or vulnerable. Maintaining up-to-date black and white lists is also a continuing chore.
- These difficulties are aggravated by concerns to avoid interfering with wanted calls (“over-blocking”). Many genuine callers (including for example government agencies, health services or banks) withhold their CLIs, and so would be rejected by some Anonymous Call Rejection settings. And requested robocalls (for example, providing online security codes) would be foiled by most call screening techniques, as indeed are some telcos’ own services like Reminder Calls.
However, for users who suffer most from nuisance calls, and especially those who are vulnerable to scam calls, on balance these technologies are clearly positive. Pilots of add-on boxes for vulnerable users in some local authority areas have been very successful; the National Trading Standards Scams Team installed more than 100 of these in 2015 and is now launching a new project using the DCMS funding shown in Figure 7. Call blockers provided for vulnerable users in several local authority areas in Scotland have also worked very well; these activities are discussed in the next chapter.
An amendment in 2016 to the Privacy in Electronic Communications Regulations requires all telemarketing calls to include a returnable CLI (which may identify them straight away, or to which a return call can be made which will identify them), and a recent ICO case has enforced this new rule. It sets a precedent for all genuine callers to provide returnable CLIs. For example, the Scottish Government promises in its Action Plan that all its outbound calls will provide a CLI. Widespread adoption of this practice will make CLI-based tools more useful to consumers.
Network call suppression has been slow to arrive, and to some extent this may reflect network operators’ mixed incentives: they want to avoid customer problems, but at the same time may derive some revenue from carrying the nuisance calls.
Figure 6: Classes of relevant call suppression technology
Class  1 call management technology, introduced in the early 1990s, relies on the caller’s number (known as Caller- ID or Calling Line Identity ( CLI)) or an alternative “presentation CLI” being made available when the phone rings. Users can choose to screen their calls based on this information, and can also use Anonymous Call Rejection network services to reject calls with unavailable or withheld CLI, to Choose to Refuse calls from certain CLIs (typically 10) or groups of CLI (such as international) , or further calls from the same caller (Last Caller Barring). The rise in “number spoofing”  has greatly undermined reliance on actual or presented CLI.
Class 2 technology blocks all calls from a much longer “block list” of originating numbers. It too depends on the availability of CLI. Its effectiveness depends on how many numbers are on the block list , how the list is compiled and how often it is updated. On current calling patterns, it can prevent maybe 40% of unwanted calls. Early call blocking devices including the BT6500 phone and some Panasonic phones work on this principle. Some UK network operators are also now employing this technique on behalf of all their customers, blocking numbers identified through the Ofcom Nuisance Calls MoU Group  or by other means, such as crowd sourcing about unwelcome calls, or observation of unwarranted call origination by its own customers.
Class 3 technology applies modern data analysis techniques at network level to a wider range of real-time data about calls, to identify traffic streams that have certain characteristics. This can lead to allocation of a trust score, such as scam or suspicious, providing customer choice on call acceptance. CLI remains an important element but the technique is not solely dependent on CLI. Some recently introduced network suppression in the UK is of this kind, and it is also used in some mobile apps.
Class 4 technology also uses CLI, but can function without it. It requires selected callers  to take some action (for example, saying their name or keying certain digits) to be connected, which dissuades unsolicited callers from continuing, and, if they do continue, helps the called person to decide whether or not to accept the call. Examples are network services in France and the USA, now followed by Sky in the UK, and some more recent call blocking devices. This approach can block many more unwanted calls, in some cases over 90% , but it requires customer agreement as it may affect the reception given to wanted calls.
Applicability of different techniques: Mobile phones can use apps but not add-on boxes. Networks can access the underlying “network CLI” as well as the “presentation CLI” which reaches end users and their equipment. Fuss Free Phones handles nuisance calls through a personal answering service which takes advantage of their special mobile network status.
Notes to Figure 6
1. This classification draws both on Allowing Consumers to Block Nuisance Calls in the Network, trueCall, July 2013 (which speaks of technology “generations”) and on BT’s three different “types” of call blocking telephone (described on their shop website). We have used the word “class” to show that our classification is not quite the same as either of these.
2. These service names are BT’s; near identical services are available from many UK landline providers through Wholesale Line Rental of BT landlines.
3. A technique for sending any caller ID of the caller’s choice, which is easy when using Voice over Internet Protocol ( VoIP) technology or when using a PABX that has been subverted for this or other purposes (such as making “free” international calls).
4. In recent years, storage limits have increased dramatically, allowing hundreds or thousands of numbers on a block list. The effectiveness of block lists depends both on their permitted length (blocking 1000 numbers should be more effective than blocking 10) and on how often callers change the CLI they are presenting. It is now easy for a caller to present a different CLI for every call. Figure 20 shows that even the top 1000 calling numbers may account for under a third of nuisance calls.
5. More information about this group is provided in Annex C, section C.2.3, on network measurements. Before blocking calls from a CLI, each operator should perform its own “due diligence” to satisfy itself that these are indeed nuisance calls. However, practices differ among operators on how this checking is done, how callers become aware that their calls are being blocked, and how callers can get mistaken blocking reversed.
6. They may be selected in different ways, for example by not being on a “white list” of pre-approved callers (with recognised CLIs or equipped with a pass code), or by having CLIs in certain categories (such as international or withheld).
7. This is also a rough estimate, depending on the specific actions requested of the caller and on callers’ behaviour. But it seems that Class 4 technology is generally more likely to block wanted calls than to fail to block unwanted calls. Messages that are left can also lead to call back scams, which may (for example) encourage recipients to make expensive international calls.
Class 1 call blocking technologies have been available for a long time, sometimes included in package pricing but sometimes charged extra, at up to £5.80 a month  . Call blocking add-on boxes and phones have to be bought, for prices ranging between £20 and £120. The network suppression services that are now arriving are all, so far, being offered to customers at no additional charge.
Mobile apps for call management and suppression have taken off in the USA  and are arriving on this side of the Atlantic. In general, they use crowd-sourced information on calls that are unwanted by their user base; some also scrape their users’ contact lists. First Orion has developed call analytics (Class 3) technology for T-Mobile in the USA, and claims a very high level of effectiveness for this approach.
Which? has published useful articles on call blocking options, covering call-blocking phones as well as three stand-alone landline call blocking devices. Another (March 2016) article compares five mobile call-blocking apps (not including TPS Protect). In general, accessing full Which? reviews requires payment.
The case for providing call blockers to vulnerable consumers is so strong that in 2015 the government promised £3.5m of central funding with this primary purpose. Figure 7 summarises published information on the expected and actual uses of this funding to date. Outcomes of component 1a are still awaited; National Trading Standards are using component 1b to deploy call blockers to vulnerable users (but will only be able to reach a small number compared with the 560,000 names thought to be on “suckers lists” now circulating)  .
Figure 7: Uses of government funding
|Item||Use||Proposed, 03/2015 ||Actual, 08/2017||Remarks|
|1||Trialling the development and provision of call blocking technology through challenge funding.||£2,000,000||£1,100,000|
|1a||Organisations to bid for funding to innovate, design and operate safe, practical and more cost-effective call blocking technology.||£1,500,000||£600,000||Half awarded to 6 companies and half to 3 of them (in phase 2)|
|1b||For agencies, local authorities and charities to trial providing call blocking devices to vulnerable people.||£500,000||£500,000||Awarded to National Scams Team|
|2||Awareness raising campaign about existing mechanisms to reduce and report nuisance calls.||£1,000,000||-|
|3||Research to determine where Government interventions could be most effectively targeted, seeking to understand the prevalence of different types of nuisance calls, actions consumers take to minimise those calls, and why others do not take similar action.||£500,000||-|
Figure 8 summarises estimates (based on limited information from various sources, plus guesswork) of the availability, take-up and effectiveness of call suppression technologies. We rely heavily on service providers’ published claims, which have not been independently verified. The 2020 figures are all guesses on the high side.
Figure 8: Possible effectiveness of call suppression technologies
|Class of call suppression technology||Start date ||Suppression effectiveness (% of nuisance calls per user)||Estimated take-up now (% of potential users) ||Possible take-up in 2020 (% of potential users)||Number of potential users in 2016 (millions) |
|Network suppression – landline|
|BT ||1, 3||01/2017||65%||22%||40%||9.4|
|TalkTalk ||1, 2||2014||50%||100%||100%||3.0|
|4 ||H2 2017||90%||-||50%||3.0|
|Virgin Media||1, 2 ||2016||Not stated||100%||100%||4.4|
|Network suppression – mobile |
|User device blocking – landline |
|Add-on boxes||2, 4||2007||67%||5%||10%||26.4|
|Blocking phones||2, 4||2013||60%||10%||20%||26.4|
|User device blocking – mobile |
|Smartphone apps||2,3,4||2010||80%||10% ||25%||41.0 |
Notes to Figure 8
1. For device technologies, estimated date of when first widely available in the UK.
2. Technologies that are applied to all connections, without individual customers choosing them, are regarded as having 100% take-up.
3. Latest available figures from Ofcom and in some cases the operators. A proportion of the landline connections is used for broadband only (not for receiving voice calls) and a proportion of the mobile connections is used for machine-to-machine communication. We have no operator-by-operator breakdowns of these proportions, so have quoted the total numbers of residential landline and mobile connections.
5. https://help2.talktalk.co.uk/what-talktalk-doing-stop-scam-calls (with clarification directly from TalkTalk, that the blocking mentioned here refers to all nuisance calls, not just scam calls). TalkTalk is also now blocking calls from numbers that have no CLI.
6. Estimate for the CallSafe service, launched 17 January 2018.
7. Virgin Media plans to provide a Class 4 nuisance call handling service to customers using the IMS platform that it is currently rolling out, and to which it plans ultimately to migrate its TDM customer base. Current IMS customer numbers are low.
8. Mobile network suppression would affect customers of MVNOs on a network in the same way as the network’s own customers.
9. O2 has provided no input to the study, but as it is a member of the MoU group we suppose that its practices are probably similar to those of Three.
10. Estimates based on information provided in confidence by sector participants, together with the sources quoted in the next footnote.
11. Estimates based on inference and information extracted from Ofcom surveys (see Annex E).
12. Truecaller claims to have the largest app user base in the UK, with 2 million downloads.
13. Estimate of the number of smartphone owners (not of the number of smartphone subscriptions).
What stands out here is the importance of user take-up, where suppression is provided on an opt-in basis. We believe that opt-in applies to all entries in the table except for network blocking provided by TalkTalk, Vodafone, and Three, which applies automatically to all subscribers. Consumer take-up of opt-in services is unlikely to be high: Ofcom survey findings, summarised in Annex E, include that although 65% of landline users were aware of blocking technology, only 9% of landline users had chosen to use it. 10% of mobile users had used their mobile settings or downloaded an app to block unwanted calls. Scams research by Citizens Advice in 2017 suggests that 11% of respondents had signed up for call blocking services, rising to 15% for people who had been targeted by a scam within the past two years.
An alternative approach worth considering is to reverse the default, switching on the suppression service automatically while giving customers the option of switching it off. This could have benefits both in reducing nuisance calling overall, and in boosting coverage of vulnerable customers; however it could lead to some wanted calls being suppressed, and would require telco systems to be dimensioned for high take-up of the suppression service. Getting customer communications right would be critical to the success of this approach  .
It is worth noting that BT’s Call Protect service is available on a wholesale basis, at a charge of £1.68 a year  , to companies who repackage and resell BT landlines (through Wholesale Line Rental). BT say that calls to their Nuisance Call Advice Line have been much lower since the launch of BT Call Protect.
Annex E also summarises consumers’ reasons for doing nothing to prevent nuisance calls, as explored in Ofcom surveys. It seems the main barriers to action are managing to think about it, together with avoiding hassle (accounting for around a third of responses); not knowing what to do accounts for another 10%. Price and over-blocking are relatively minor concerns, together mentioned by under 10%. Close to 30% do not regard nuisance calls as a problem worth bothering about.
It seems reasonable to suppose that early adopters of blocking technology are people who know about it and are most troubled by nuisance calls  . If this is so, we may expect that over the next few years, as network suppression technology at no extra charge becomes more widely available and known, then take-up will increase and harm from non-scam nuisance calls will decrease, roughly in proportion to the effectiveness of the suppression method(s) used. However, this expectation comes with some big provisos:
- The easiest way to sign up for most services is online, but many people who are troubled by nuisance calls are not internet users. Telcos need to provide easy alternative ways of signing up and make sure that all their customers know about these.
- As long as take-up is on an opt-in rather than opt-out basis, it is unlikely that suppression overall will get high enough to deter mass automated telemarketing.
- Serious scamming often uses a variety of CLIs, and (especially when high-value) may not display the distinctive traffic patterns that network suppression algorithms recognise and exploit. Suppression with very high effectiveness (say, over 95%, or at least Class 4 standard), probably applied at the individual level, will therefore be needed to protect vulnerable users from receiving scam calls.
- Once Class 2 systems are more widely deployed, call centres will start taking action to defeat them. Call centres can easily keep changing their calling number. Class 4 systems focus on an ‘allow’ list rather than a ‘block’ list, so changed phone numbers will be treated as ‘untrusted’ and therefore intercepted.
Responding to the 2013 All Party Parliamentary Group enquiry, in 2014 Ofcom found out that most customers wanting advice on how to handle nuisance calls would ask their operator. Ofcom therefore looked at operators’ websites and practices in this area, and found some good practice but also considerable variation and shortcomings. We have looked again at this, and found that most significant operators with personal customers do provide some advice on their websites on nuisance calls, but as Figure 9 shows  , this advice varies quite widely without clear reason. Bringing all these websites up to best practice looks like a quick win, both for operators and for their customers.
Figure 9: Operator website advice on nuisance calls
|Mobile operators||Landline operators|
|EE||O2||Tesco Mobile||Three||Vodafone||BT||Plusnet||Sky||TalkTalk||Virgin Media|
|Advice on reporting nuisance calls|
|Report to TPS||Y||Y|
|Report to ICO||Y||Y ||Y ||Y ||Y||Y||Y ||Y |
|Report to Ofcom (silent and abandoned calls)||Y||Y||Y|
|Report to police (malicious calls and texts)||Y||Y||Y||Y||Y||Y|
|Report to Action Fraud (scams)||Y||Y||Y|
|Report to operator||Y||Y||Y||Y ||Y||Y||Y |
|Report to PSA||Y|
|Report to Which?||Y||Y||Y|
|Options for protection against nuisance calls|
|Register with TPS||Y||Y||Y||Y||Y||Y||Y|
|Block calls, with white and black lists||Y||Y||Y|
|Bar a number in network||Y||Y||Y||Y||Y|
|Bar a number on a phone||Y||Y||Y||Y||Y||Y|
|Change your number||Y||Y||Y||Y|
|Instructions on barring a number on phone||Y||Y|
Notes to Figure 9
1. For spam texts
2. For malicious calls and texts
3. Charged service, not free
4. For all anonymous callers (Anonymous Caller Barring, ACB)
5. For the last caller (Last Caller Barring, LCB)
2.3 Minimising harm caused by nuisance calls received
Lastly, we consider the harm caused by nuisance calls which, despite preventative actions such as those discussed earlier, have been received  .
2.3.1 Types of nuisance call
Annex A offers a set of estimates of proportions of call types, based on data from Ofcom and trueCall. As Annex A highlights, nuisance calls can be classified in several different ways – for example by severity of nuisance, by originating sector and location, or by whether live agents or recorded or interactive voice technology are used. According to ContactBabel information provided to Ofcom  , in 2015 the outbound activity of UK call centres was divided roughly as shown below, according to the motive for the activity  .
Figure 10: Breakdown of UK outbound calling
|Outbound calling activity to UK consumers||Proportion of total|
Source: 2015 ContactBabel survey of UK call centres
The fact of taking part in the UK call centre survey behind Figure 10 points to respondents probably being aware of UK regulations and at least aiming to comply with them. Not covered by this survey is a large number of other UK call centres, and many more whose calls appear to originate outside the UK (for example because they present international CLIs, or the agents have foreign accents). Calls in these latter two categories are less likely to comply with UK regulations, and may well be scams. Because of widespread number spoofing, no reliable estimates are available of the proportion of calls in each category. However, an industry source has suggested that the three (i.e. compliant UK call centres, non-compliant UK call centres, and non- UK call centres) produce roughly equal volumes of unwanted calls. This is in line with two-thirds of complaints to Ofcom about silent and abandoned calls having missing CLIs  , and with the one-third reduction in nuisance calls resulting from TPS registration (presumed to reflect calling from compliant organisations), so we work on this basis for the time being.
A major telco reports that callers purporting to offer debt management, PPI reclaim and car accident claim management services make up the great majority of unsolicited marketing calls on their network. Other published data on types of call (mainly by originating sector) are shown in Figure 11 and Figure 12, with our own analyses of call types from ICO complaints data in section 3.2 (see Figure 21), and from trueCall data in Annex J. Overall we conclude:
- Without consistent terminology when classifying nuisance calls, these data have little objective measurement value. For example, what is meant by the term “scam” clearly varies by data source (and Annex I shows that it varies also by complaints system).
- There are also real variations from time to time in the intensity of different types of nuisance call; this may best be illustrated by the combined Ofcom diary surveys shown in Figure 12.
- Claims management (including PPI) has consistently been a major source of nuisance calls.
- Call blocking services have been a source of nuisance calls, with companies making fraudulent offers to put consumers on supposedly superior Do Not Call registers or provide poorly performing, high priced call blockers  . Such calls can further confuse consumers.
Figure 11: Data on types of nuisance calls
|Data from Truecaller Insights Special Report, based on 2 million British users of their mobile app, 01/01/2017-31/05/2017 31% telemarketing
14% nuisance (prank calls through to harassment)
12% telecoms operators
10% financial services
8% scam calls
2% market research
| Five biggest categories of nuisance calls on BT’s network, 04-11/03/2017 Accident claims 41%
Personal details (scam) 18.5%
Computer scam 12.6%
Debt collection 7.5%
Total (29.5 m) 100%
|Which? surveys 01-08/09/2017 and 11-19/11/2015 The most common calls to landlines reported in 2017 relate to: silent calls (mentioned by 48% of respondents), PPI insurance claims (42%) and accident claims (44%). In 2015, the three most common types of calls were PPI (66%), silent calls (55%) and the Green Deal or energy efficiency measures, including boilers and double glazing (52%).|
Figure 12: Call sectors from Ofcom landline nuisance call surveys, 2013-2017
Source: Landline Nuisance Calls W5 presentation, GfK UK for Ofcom, p 21
As well as looking at the originating industry sector, both Ofcom and ICO aim to classify nuisance calls by whether they are live or recorded, silent or abandoned. These distinctions are important to the regulators because of their split responsibilities, with ICO responsible for regulating live and recorded calls and Ofcom for silent and abandoned calls. Differences in how consumers feel about the calls in these dimensions (e.g. silent vs live voice) do not however appear to be great – all these are found annoying by around 80% and distressing by 5%-10%  .
2.3.2 Types of nuisance call recipient
Both telemarketing and scams will, where possible, naturally target those whom the callers have reason to believe are likely to be receptive to their messages. In both cases, having been receptive before is a prime indicator of likelihood of being receptive again.
The 2009 University of Exeter report The psychology of scams  says:
“Our research suggests that there is a minority of people who are particularly vulnerable to scams. In particular, people who reported having previously responded to a scam were consistently more likely to show interest in responding again. Though a minority, it is not a small minority; depending on how it is assessed, it could be between 10 per cent and 20 per cent of the population.”
Earlier OFT research  found that 52 per cent of victims had been targeted again by a scam and that, on average, a victim had a 30 per cent chance of falling for another scam within the following 12 months. 2017 research by Citizens Advice shows that 72% of respondents had been targeted by a scam in the past two years, and being targeted once raised the probability of being targeted again to 83% or more. Scam phone calls seemed to have a higher probability of repeat targeting than online, text, paper mail or doorstep channels, though lower than email.
The Exeter report discusses how, as a psychological type, vulnerability to scams is not age-specific. But circumstances which are more likely among older people (such as isolation, loneliness, and diminishing mental capacity) boost the probability of vulnerability being translated into being targeted and finally into becoming a victim. More recent UK research is working towards a psychological vulnerability profile  , and US research  provides further insights in this area.
Ofcom surveys make it clear that older age groups do receive more landline nuisance calls than average; Figure 13, from the 2017 landline nuisance call survey, illustrates this  . Numerous other sources (some quoted in Annex F) confirm that this is true of telemarketing and scam calls. Recent research in both the US and the UK  shows that younger age groups are also at risk, particularly via mobile phones.
Figure 13: Landline nuisance calls by age group
|Mean number of nuisance calls in four weeks per diarist||4.1||4.5||4.5||6.8||9.28||10.28||6.8|
|Proportion of diarists receiving nuisance calls in four weeks||63%||74%||75%||82%||88%||94%||81%|
Source: Ofcom landline nuisance call diary survey 2017
Susceptibility to the less desirable aspects of some telemarketing (such as high-pressure tactics, incomplete information provision, low quality or over-priced goods and services) has not received the same attention as susceptibility to scams. However, it seems reasonable to suppose that many of the same factors will apply, so we can regard the same population as being at particular risk.
2.3.3 Estimating consumer harm caused by the calls
The only available estimates of consumer harm caused by nuisance calls are those offered by Ofcom as background to the 2015 Persistent Misuse Consultation. At around £0.1 per call, these estimates are based on cost of time wasted and take account of consumers’ willingness to pay to avoid the calls, but exclude both mitigation costs (e.g. the cost of call blocking) and, importantly for this study, the consequent harm caused by answered calls. We know that the last can be high, especially when vulnerable consumers answer scam calls.
Scam approaches and successful scams are believed to be grossly under-reported  , though by what factor is unknown  . This, coupled with a lack of statistics on the role of phone calls in successful scamming, makes it impossible currently to pin down the number of scam calls or the damage that they cause.
The numbers in Figure 14 are therefore notional, but they are influenced by the scattered and varied information that we have been able to gather (summarised in Annex F). We think it likely that worst-affected call recipients are more likely to engage in conversation with dangerous callers, as well as receiving far more than their share of approaches from these callers, resulting in a more than ten-fold greater exposure to the risks.
Figure 14: Notional distribution of harm from nuisance calls to landlines
|2% worst affected||Remaining 98%||All||Source|
|A||Number of adults using landline (million)||0.80||39.2||40||Rounded figure based on ONS and Ofcom data (see Figure 40)|
|B||Average number of nuisance calls per month per adult||35.6||7.2||8.2||Based on Ofcom diary surveys 2013-2017 |
|C||Proportion of nuisance calls that are scam calls||30%||15%||15.3%||Conservative assumptions drawing on BT and trueCall data|
|D||Proportion of scam calls leading to dangerous conversations ||30%||10%||10.4%||Conservative assumptions drawing on data from Money Advice Service, AgeUK, and Citizens Advice|
|E||Proportion of dangerous conversations leading to loss||15%||4%||4.2%|
|F||Proportion of scam calls leading to loss||4.5%||0.4%||0.5%||= D x E|
|G||Average scam loss (£)||350||350||350||Assumed |
|H||Average cost of scam calls per year per adult (£)||2,019||18||58||= (12 x B) x (C x F) x G|
|J||Total cost of scam calls per year to this group
|1,615||711||2,326||= A x H|
|K||Basic cost per nuisance call (£)||0.1||0.1||0.1||Ofcom|
|L||Total basic cost of nuisance calls per year to this group (£ million)||34||339||373||= A x (12 x B) x K|
|Total (basic and scam) cost per year to this group (£ million)||2,053||1,050||3,103||= J+L|
Notes to Figure 14
1. Figure 3 relates to a week and this table relates to a month, so the average here is 52/12 of the Figure 3 averages.
2. A “dangerous conversation” is one where the caller has criminal intent and the called party engages in conversation (rather than cutting the call short early on).
3. This round figure is based on an estimate (based on research) from the 2017 Citizens Advice report Changing the story on scams that the median phone scam loss is £693, and allowing for around half of losses to have been recovered. The total of £3.1bn attributable to phone scams is under 30% of an earlier Citizens Advice estimate of £10.9bn total personal losses due to scams.
Although many of the numbers have little empirical basis, we believe that this exercise correctly illustrates how harm is magnified by vulnerability to scam attempts, and how it is concentrated on a small minority of worst-affected recipients – with 2% of people bearing two-thirds of the cost. This picture should be taken into account when prioritising actions.
Here we are defining “worst affected group” as those people who receive the most nuisance calls each and together receive one tenth of all nuisance calls. As explained in section 2.1.1, according to the Ofcom data used in Figure 3, people in this group receive 35 or more nuisance calls in four weeks (with an average of 6 times the overall average) and amount to under 2% of all landline users.
Who are these “worst affected” people? As discussed in the previous section, we know that successful nuisance and scam calling attracts more calling, so it is reasonable to assume that many of them share the University of Exeter’s “vulnerable to scam” psychological type, which they thought might affect 10% to 20% of the population. We have already mentioned the 560,000-long “suckers list” of scam victims. We also know (from Trading Standards  and other sources identified in Annex F) that high levels of nuisance calling may afflict people living with dementia  , people with physical and sensory disabilities, older people, and those living alone. We suspect that people in our worst affected group will have two or three of these attributes. They are not easy to identify or help. Clearly they are vulnerable in more than one sense - simply getting that barrage of nuisance calls is quite bad enough, on top of which they may feel it is wrong to put the phone down on someone, be prone to fall for scams or to make unintended purchases, or risk a fall on the way to a ringing phone.
All this applies only to nuisance calls to landlines. A similar exercise should be carried out for nuisance calls and texts to mobiles, on which there is currently less information available. Given high and growing use of mobiles, we would expect the harm associated with nuisance calls to them to be of a similar order to the harm associated with nuisance calls to landlines. However, the groups who are worst affected will differ, in particular by age.
Harm caused by answered calls without criminal intent is very hard to assess. As well as the time wasted by the interruption (with associated annoyance), there may also be detriment associated with mis-selling, for example successful high-pressure sales that are not what the consumer really wanted. However, the risks in question should on average be lower – if an actual purchase is made, presumably on average some value will be derived from it. We have almost  no evidence of the incidence or size of such detriment, so unlike the cost of scams it is not added to Ofcom’s basic cost per nuisance call to landlines. For a complete harm assessment, further estimates would be needed.
Costs of the order shown in Figure 14 make a very clear case for efforts to prevent nuisance calls from reaching customers, and especially those who are worst affected. Accounts of the experience of sufferers (like Jessica of the Think Jessica campaign) show that this is a public health problem, in the same way as gambling; and like other public health problems, it imposes huge costs on public services as well as on individual victims. However, preventing scam calls without an overall scam prevention strategy may lead to some criminal activity simply being displaced; harm removed from scam calls may well pop up again elsewhere, with online scams being an obvious area to watch.
2.3.4 Actions intended to reduce harm from nuisance calls received
1. Improve warm sales calls. Anecdotally, warm sales calls (within an existing business relationship) may be as persistent and unwelcome as cold sales calls. The customer may not be aware of having agreed to receive calls from the company. It seems like basic business sense for companies to record and respect customers’ contact preferences, but clearly this is not standard practice  . At a minimum, agents could be instructed to ask whether this is a convenient time to speak – a common feeling being that calls “always arrive during dinner”  . We suggest that compliance with best practice  might reduce both the number of unwanted calls made and the annoyance that they cause, thereby reducing the harm caused by the one third of 18% of calls in this category  .
2. Support vulnerable consumers. As discussed above, even legitimate calls can be particularly risky for vulnerable consumers. If all call centres followed the DMA Guidelines on calling vulnerable consumers (mainly those with physical or mental disabilities), risks of harm to this group should be much reduced. Companies involved in fulfilling telesales can also help here, for example by querying duplicate or multiple insurance policies or magazine subscriptions, or (in the case of banks) unusual account activity.
3. Minimise debt collection trauma. Debt collection is a large category of call which risks being unwelcome, no matter how carefully it is carried out  . However, harm to recipients of these calls could probably be reduced somewhat by attention to guidelines  , and in particular by checking with recipients whether they would prefer an alternative contact method.
4. Educate identified vulnerable consumers. Answered scam calls cause the most harm, especially to vulnerable consumers. The only real protection here is through consumer education, with support for the most vulnerable from others who come into contact with them (as fostered by the Trading Standards Friends Against Scams initiative). People who have been scammed once are known to be at particular risk in future. The 2009 Exeter report offers a grain of hope:
“The likely existence of a subset of the population with enhanced vulnerability to scams is both a problem and an opportunity from a consumer education point of view. It is a problem in that it suggests that a high proportion of any general awareness campaign will be wasted on people who are relatively unlikely ever to fall for a scam. It is an opportunity in that if the more vulnerable group can be identified – or can be encouraged to self-identify – educational material can be targeted at them.”
5. Spread CLI display and educate consumers. Consumers can, of course, reduce harm by refusing to speak with unwanted callers. Their time has still been wasted, but at least they will not be making unwise purchases, far less being caught by scams. Much consumer education consists in getting across the message that it is better not to speak to unsolicited callers, even if this feels impolite. The 2015 consumer issues survey findings in Annex E show that 53% of home phone call recipients and 63% of mobile call recipients did consider the possibility of calls being unwanted and varied their answering behaviour accordingly. However, under half of home phone call recipients said they had a CLI display, which supported the decisions of over 60% of those who had it. Actions to increase take-up of home phone CLI display, together with education to avoid answering the phone to unknown callers and to cut short unwanted conversations, could significantly reduce harm.
6. Provide CLIs to which return calls can be made. Routine provision of CLIs that are recognisable, authenticated and allow return calls to be made to them, in particular by government agencies, health bodies and businesses making genuine calls, would foster confidence in the use of CLI and should encourage consumers to use nuisance call suppression systems.
7. Improve complaint systems. For a small minority of consumers  , complaining about nuisance calls may help them to feel a bit better, presumably slightly reducing harm to them, even if complaining takes up extra time. Despite some improvements in the last few years, nuisance call complaints procedures remain hard to find and navigate, and can easily take up many times longer than the call itself did  . More complaints may be of some value to enforcers, as they provide intelligence and strengthen the case against offenders. Complaints procedures could certainly be made much easier to use  , but it is hard to suggest that this would bring about any material reduction in harm to consumers.
8. Transform early warning of scams. On the other hand, sharing information about scam calls received could avert a lot of harm, especially if done (and reacted to) early. Prompt intelligence about new scams could alert other consumers, network operators and the authorities to the dangers, enabling protections to come into play. To get this to happen requires new levels of willingness to report among affected consumers, together with much improved systems for consumer reporting, and sharing information among the operators and authorities. It is probably too much to hope that victims of multiple scams can be changed in this way, but “once bitten, twice shy” consumers might be recruited to an early warning network, with new phone numbers.
This project looks at scam calls as the most harmful type of nuisance call. Scam calls are more often, and perhaps more helpfully, viewed as just one channel among others (such as email, websites, and paper mail) used by mass market fraud. Scam calls will therefore be addressed by new national counter-fraud initiatives like the Joint Fraud Taskforce and the Banking Protocol, both mentioned in the Annual Review 2016-2017 on Economic Crime  . These should impact fraud by any channel, but we cannot estimate their effectiveness specifically for fraudulent phone calls.
Fraud directly against individuals is especially distressing, and likely to lead to lasting personal financial loss. However we should bear in mind much larger scale fraud  that targets institutions like banks, often via identity theft  . The losses concerned tend to be borne by the institution, and accordingly, spread among all its customers. The telephone can play a significant part in institutional fraud, since fraudsters often impersonate bank customers phoning the bank’s call centre. The fraudsters gather information needed for impersonation by various means, which may include nuisance phone calls. Techniques used for countering telephone fraud against call centres, such as voice and data analytics  , may also help in countering calls like these, as well as direct telephone fraud against individuals.
2.4 Summary of possible effectiveness of actions
Figure 15 summarises the main actions that have been discussed in the three sections of this chapter, and tentatively assigns to each a rough level of potential reduction of harm by 2020. Some actions should have greater effects longer term, in particular those relating to CLI  . The effects only apply to the beneficiaries shown in the “beneficiaries” column – so, for example, call blocking apps can bring about major reductions in nuisance calls to smart phones, but this is confined to those people who download and use them. The indicative effect bands are colour coded as shown below.
Figure 15: Summary of possible effects of existing UK actions
There is a problem
Thanks for your feedback