You're viewing our new website - find out more

Publication - Report

Effectiveness of actions to reduce harm from nuisance calls in Scotland

Published: 19 Mar 2018
Part of:

Research commissioned to analyse the impact of actions set out in the Nuisance Calls Commission action plan, and to examine the outcomes of past interventions.

144 page PDF


144 page PDF


Effectiveness of actions to reduce harm from nuisance calls in Scotland

144 page PDF



1. The work has been commissioned by the Consumer, Competition and Regulatory Policy Unit of the Directorate for Economic Development of the Scottish Government from Antelope Consulting, with trueCall providing access to their nuisance call database and technical support on its use.

2. This Action Plan was the outcome of the Scottish Nuisance Calls Commission, which is explained and whose papers are available at

3. This part of the report builds on an earlier working paper that was circulated among interested parties for comment.

4. According to the survey findings, on average fewer than 1% of diarists received more than 40 calls in four weeks and no diarist received more than 84 calls in four weeks. We note here, and illustrate in Annex H, that these surveys may lead to underestimates because of incomplete recording.

5. In fact we have approximated this procedure by saying that everyone receiving the same number of nuisance calls will be in the same decile, so that the limits of the deciles are whole numbers.

6. The jump in mobile registrations in June 2016 followed the well-publicised new facility of registering a mobile number by text message. Registration, once carried out, does not need to be renewed, and TPS is not routinely informed about ceased numbers. The total of over 20m landline and 3.5m mobile registrations therefore includes an unknown number of phone numbers that are no longer in use. Registered numbers in use must, however, cover a far greater proportion of the ~20m UK households than the 21% saying they had registered when asked by an Ofcom survey ( Annex E). We can only think that many people have forgotten their registration, never knew about it (as it was done by another household member), or assume it has expired as they are once again getting lots of nuisance calls.

7. We use the term “suppression” to include both blocking and diversion. The term “blocking” is also commonly used with this wider meaning, but the distinction can be useful.

8. As shown above, combining Ofcom’s five landline nuisance call surveys of 2013-2017 shows 47% of diarists receiving at most 4 nuisance calls in four weeks and 68% of diarists receiving at most 8 nuisance calls in four weeks. And Ofcom’s omnibus survey published in March 2015 (waves 11-12) found that of the 70% who did receive nuisance calls (on their landlines or mobiles), 71% were not prepared to pay 50p a month to be free of nuisance calls – though a small proportion would pay varying amounts, up to £10 a month. The same survey found that 51% of landline users and 56% of mobile phones users did nothing to avoid getting nuisance calls because they got few or none, didn’t mind them or hadn’t thought about doing anything. More details are given in Annex E.

9. A consumer survey in 2008 (carried out by Which?) suggested that TPS registration cut nuisance calls by 54%, much more than the 35% found in 2014. See p 9 of TPS report on unwelcome calls 2008, Brookmead Consulting.

10. For example, in the 2016 ICO Track consumer research, 72% of respondents feared their data being used for nuisance and cold calling – second only to the 75% who feared it being stolen by criminals.

11. In 2013, already two-thirds of residential landline numbers were said to be ex-directory, but ex-directory numbers were still called, whether via other sources for the numbers, or via random or sequential dialling. The proportion has gone on growing: in November 2017 BT said that 75.7% were ex-directory. Ofcom’s 2014 diarist panel survey report (on p 14) states that there was little difference in the overall incidence of nuisance calls between those who did and did not take care about releasing their phone number, or opting out of marketing information, though those who did take care had fewer recorded sales calls and were less likely to receive over 20 nuisance calls during the four weeks of the study. The same survey also showed that online behaviour (shopping, entering competitions or using price comparison websites) made no significant difference to the number of nuisance calls received.

12. Claims Management regulation is the most successful example to date. In August 2017 the UK Government undertook to ban cold calling about pensions, but concrete steps are yet to come.

13. Ofcom’s November 2016 NICC presentation slides 9 to 11.

14. Ofcom’s November 2016 NICC presentation slides 6 and 11.

15. Landline call termination charges are now very low (at around 0.03 pence per minute) and nuisance calls have a short average duration so the related revenues are not large. Mobile call termination charges are reducing, but still more than 10 times higher than the landline equivalent, making them more worthwhile for networks to terminate. Higher origination costs for calls from mobiles have historically helped to limit nuisance calls to mobiles, but will do so less in future.

16. This may mean diverting them to a voicemail box (as BT is doing), screening them (giving the called customer the choice of whether or not to accept them) or making them completely unavailable to the customer. Depending on the hardware and suppression technique used, the phone may ring briefly or not at all.

17. There can also be serious drawbacks to businesses, if their legitimate outbound calls get blocked. There is ongoing debate in the industry about how best to avoid this happening, and what action should take place if it does.

18. Sample prices are shown at the foot of this Ofcom guide:

19. In August 2016 the US regulator, the FCC, ordered network operators to take action against nuisance calls. Mobile users in the USA typically pay to receive calls, adding injury to insult when the calls are unwanted.

20. Chartered Trading Standards Institute, Stand Against Scams. (The figure of 300,000 mentioned in the 2017 Citizens Advice Scams Awareness Month Briefing is drawn from an earlier CTSI statement).

21. Budget 2015, amplified by DCMS handout at Round Table 23/03/2015 on call blockers for vulnerable consumers, reproduced in a DMA news release.

22. It has also been suggested that the default might be reversed just for customers whom the telco has reason to believe are especially likely to be vulnerable to nuisance calls. Such an approach would clearly have to be subject to appropriate data protection safeguards. If pursued, it might build on recent work by the UK Regulators’ Network on identifying customers in vulnerable situations as regards energy and water supply.

23. See Choose to Divert in the Wholesale Calling Features (analogue) section of BT’s price list

24. The survey sample sizes are too small to say whether those who had chosen to use blocking were those who received the most nuisance calls.

25. All the results are obtained by searching on “nuisance” on each operator’s website. The hyperlinks embedded in the operators’ names give some indications of how easy or difficult it may be to find this information by other means.

26. We exclude from “calls received” those which have been blocked by home call blockers, or mobile settings or apps. We include those which go to a home answering machine or voicemail, or which ring unanswered, as well as all those which are answered. This distinction is intended to reflect the fact that a ringing phone, along with the decision whether or not to answer it, is a distraction and takes time.

27. And reproduced here with the kind permission of Steve Morrell of ContactBabel.

28. The mean proportion of calls to mobile phones (rather than landlines) among survey respondents was 58%.

29. Quoted in Ofcom’s Call for Inputs preceding its Persistent Misuse review.

30. See, for example, the item Commercial alternatives to TPS, Bogus TPS & scams at

31. Pages 34 and 36 of Landline Nuisance Calls W5 presentation, GfK UK for Ofcom. These and similar distinctions can also be valuable for enforcement purposes.

32. The psychology of scams, University of Exeter for the Office of Fair Trading, 2009, OFT1070

33. Research on impact of mass marketed scams, Office of Fair Trading, December 2006, OFT883

34. See Silence of the Scams slide pack (from an autumn 2016 conference at Brunel University) and related publications listed at

35. Such as the TrueLink Report on Elder Financial Abuse 2015 and the FTC survey mentioned below. See also Annex F.8.

36. One reason for the rise in landline nuisance call incidence with age is that people in older age groups are more likely to be at home.

37. For the US, see for example and the FTC survey of consumer fraud 2011 (a repeat survey is now in progress). For the UK, see the Silence of the Scams slide pack.

38. Victims often feel embarrassed that they fell for a deception that becomes clear with hindsight; or in some cases they lack the mental capacity to recognise what is going on.

39. Annex F provides an old OFT estimate of 5% for scams and a recent Crime Survey estimate for fraud of 17%. Neither applies specifically to phone scams.

40. See for example Protecting older and vulnerable consumers in Scotland from nuisance and scam phone calls, Final Report, COSLA, August 2015

41. Presumably, in the main, those who are still living independently, including many whose dementia has not yet been diagnosed.

42. Based on a report provided to this study by CPR Call Blocker, who provided equipment to a trial of call blockers in the Wirral in 2015, 6% of triallists made a purchase from an unsolicited call and 67% of those making a purchase were dissatisfied. Average spend was ~ £1200/18=£70 before, and £25 (in a tiny sample of 3 people) after the device was installed.

43. The widespread practice of outsourcing telemarketing is an important factor here. See the discussion in 3.

44. Although the actual profile of calling during the day is fairly flat. For example, the 2013 landline nuisance call diary survey recorded that between 8% and 11% of nuisance calls were received during each hour between 9:00 and 19:00.

45. This action is in accordance with the theory of Ethical Business Regulation ( EBR), as advanced by Professor Chris Hodges. Our earlier discussions on the difficulties of regulating nuisance calls point to limitations of EBR here. The practical approach to EBR in these circumstances will be collaboration among all well-meaning market participants to achieve best possible outcomes for consumers, as outlined in a case study of ethical regulation in the book Ethical Business Practice and Regulation by Chris Hodges and Ruth Steinholtz, Hart Publishing, December 2017.

46. It should also improve the job satisfaction of the call centre agents.

47. Figure 50 of the same ContactBabel report says that 22% of answered debt collection calls are hung up by the consumer, while only 10% of cold calls end in this way, and only 0-3% of other types of calls. This may be a fair indication of how unwelcome the calls are.

48. See Annex G. Echo Managed Services provide a useful research-based presentation / good practice guide which highlights that treating debtors as valued customers, and in particular offering them a choice of communication channel, is more likely to lead to debt recovery and long-term profitable customer relationships. Echo found that 40% of customers with experience of debt recovery said they would react best to a reminder phonecall, with 30% preferring a letter, and smaller proportions citing other channels. Thanks to StepChange Debt Charity for identifying this resource.

49. According to Ofcom research summarised in Annex E, about 1% of consumers usually complain when they get unwanted calls, with a further 10% or so doing so occasionally.

50. Please see Annex H for more detail.

51. For example, by automating them. The new individual suppression services now being provided by network operators appear to offer an obvious opportunity for blocked numbers to be passed straight on to regulators and counted as complaints. The new TPS Protect mobile app includes a capability for the act of blocking a number to be turned into a complaint, to be passed directly to the relevant official complaint point, although full implementation awaits agreement by all the official complaint points.

52. By City of London Police, who are the national policing lead in this area, and responsible for Action Fraud. The report is at See also this news item on how the Banking Protocol (while still in pilot) can stop fraud attempts in their tracks.

53. The Annual Fraud Indicator report 2016 of the Centre for Counter Fraud Studies at the University of Portsmouth estimates total annual UK losses to fraud at £193 billion, of which “only” £10 billion is fraud against individuals.

54. Documented by the industry anti-fraud organisations CIFAS and Financial Fraud Action UK.

55. As offered by (for example) Pindrop in the USA and UK.

56. Ofcom says that more than half of the complaints it gets about silent and abandoned calls have spoofed CLIs. (Information about CLIs in Ofcom casework is currently held internally; Ofcom plan to increase transparency in this area).

57. As discussed below, these observations do not apply if Scotland is compared with other regions of the UK rather than with other countries of the UK.

58. These findings need to be treated cautiously, as with sample sizes of around 1,000, and an aim to be representative of the UK as a whole, the consumer issues survey have confidence intervals that overlap around the percentages for individual nations and regions. Relevant details are in Annex H.

59. Though each survey is shown in a different colour, there is no clear overall trend among the surveys. Variability occurs through time as well as by region.

60. A fuller explanation of the vulnerable/standard distinction appears in Annex H.3.

61. Other parts of the UK display variability with time, but less marked and more random, without clear sustained peaks or troughs.

62. This finding is consistent with detailed analyses of nuisance calls for East Renfrewshire Council: 9 out of the 10 top calling numbers were making calls related to energy efficiency. See Also, 46% of the nuisance call cases recorded by Trading Standards Scotland in the year to March 2017 related to energy efficiency.

63. The latest publicly available study of the UK call centre industry, commissioned by the DTI, dates from 2004.The industry specialist, ContactBabel, has been unable to make specific contributions to this study. Ofcom commissioned a study of UK outbound calling from ContactBabel in 2015, published as Annex 6 to its Persistent Misuse consultation. A published summary of ContactBabel’s 2017 UK industry overview shows that in Scotland and the north of England, over 5% of the employed population is estimated to work in call centres, while in other parts of the UK, the corresponding figure is 2% to 4%, and in London it is under 2%. They say that large call centres (with over 250 agent positions) employ around half of all call centre staff, despite only accounting for less than 9% of physical call centre sites. Further information has been provided by Nerys Corfield, Chair of the DMA Contact Centre Council.

64. See for example the recent book Working the Phones: Control and Resistance in Call Centres, Jamie Woodcock, Pluto Press 2017, which includes a 10-page bibliography.

65. ICO’s term for contacts from the public, also known as complaints. The web page for their monthly newsletter about action to combat nuisance calls and texts links to a downloadable file of statistics for the current year.

66. See the Fair Telecoms Campaign briefing Energy Efficiency and Effective Regulation of Marketing Activity at

67. See DECC press release at Green Deal is now being revived in a different form, under private management – see

68. Pages 14 to 16 of the March 2017 Citizens Advice report Frozen Out usefully describes the Energy Company Obligations and Government funded programmes across the UK, highlighting how Government funding continues in both Scotland and Wales. Up-to-date information on various schemes is held at

69. Key Code provisions include: 3.10. A Green Deal Participant must not engage in high-pressure sales techniques and must not accept sales leads from persons who are known to use such techniques or are suspected of doing so. A Green Deal Participant must take reasonable steps to satisfy itself about how other parties obtain sales leads before entering into arrangements with them. 3.11. A Green Deal Participant must not offer payments or other remuneration which incentivise staff or other persons to engage in inappropriate sales techniques, or to recommend specific products or services when these may not be appropriate for the customer.

70. Warmworks Scotland, a key implementation partner, has around 30 approved subcontractors.

71. East Dunbartonshire was also involved in earlier similar work.

72. See the press notice:

73. At an estimated direct technology cost of £100 per unit (as used in the East Renfrewshire cost-benefit calculations), this would enable about 700 people to be provided with call blockers across the three local authority areas.

74. Much of this information was supplied to the study by the late Brian Smith, Senior Trading Standards Officer at Angus Council.

75. Interview with Chief Inspector Ronnie Megaughin and Detective Inspector Frank McCann, Police Scotland, 13/07/2017.

76. Police resources are allocated in the light of a Public Interest Test incorporating such considerations.

77. Following this study, the Scottish Government has commissioned from EKOS Consultants a study on the economics of preventative actions in relation to scams in general (via any channel, not just the phone).

78. See

79. This is also in use in much of England and Wales, though other parts of England and Wales use a different system, IDB.

80. A 5% reporting rate is sometimes quoted, but no foundation has emerged for this other than educated guesswork.

81. Interview with David Clancy of ICO, 01/08/2017.

82. This allows consumers to register their desire not to receive fundraising calls. See

83. Interview with Laura McGlynn and Jamie Steed of the Scottish Government, 11/07/2017.

84. There could be some parallels with the situation in Norway, outlined in an annex to the 2015 StepChange report Combatting Nuisance Calls and Texts.

85. See for example Offline and Left Behind , CAS , May 2013; Bridging the Digital Divide : Measuring the progress of digital inclusion amongst Scottish CAB clients, CAS, May 2016.

86. See

87. The study was commissioned by the Consumer, Competition and Regulatory Policy Unit of the Directorate for Economic Development of the Scottish Government, which has also produced the Nuisance Calls Action Plan. For brevity, we assume that this Unit will continue to lead on the implementation and monitoring of this Action Plan, and that future progress and measurement reports will be made to it.

88. Ofcom’s Statistical Release Calendar for 2017 includes the following note: “We are proposing to change the frequency of this study from annually to every two years. The next nuisance calls landline panel research will report in March 2019. Please contact by Friday 20 October 2017 if you have any feedback regarding this change.”

89. The mapping exercise has used the same intermediate zones as the Scottish Index of Multiple Deprivation ( SIMD).

90. In particular, some of the registrations attributed to the council areas closest to Glasgow City may properly belong to Glasgow City. By contrast, the figures for Shetland Islands and Orkney Islands at one extreme, and Eilean Siar at the other, should be free from confusion with any other areas,

91. These include complaints about nuisance calls to CMRU and to the mobile operators via ‘7726’.

92. See for example Understanding Consumer Experiences of Complaint Handling , September 2016, research by djs for Citizens Advice.

93. Claims management regulation review: final report ( HM Treasury and the Ministry of Justice, 2016),

94. The figures here are for the few cases where both the numbers of calls or texts and the numbers of complaints are available. The ICO annual reports for 2014-2017 together provide a total of 508,396 complaints (or “concerns”) made directly to ICO, with 88% due to calls.

95. More than two million now on BT's free service to crack down on nuisance calls ( BT, April 2017),

96. Tackling nuisance calls and messages: Update on the ICO and Ofcom Joint Action Plan (Ofcom, December 2016),

97. For example, by First Orion for TPS Protect.

98. See for example which offers the dummy number 0333 88 88 88 88.

99. Installing a domestic call blocker requires no more special knowledge or skill than installing a phone with an emergency button.

100. Notice of decision to impose a financial penalty pursuant to section 27A(5) of the Electricity Act 1989 and section 30A(5) of the Gas Act 1986 (Ofgem, December 2013),

101. ASA Adjudication on Data Supplier ( ASA, December 2012),

102. About the FCA Warning List ( FCA, August 2017)

103. Fake FCA emails, letters and phone calls ( FCA, April 2016),

104. William Howarth ( FCA, May 2015),

105. In section 2.2 of the ICO Plan, 2016-2019, dated February 2016,

106. A useful timeline is provided at

107. Which? worked closely with Mike Crockart, from 2010 to 2015 the MP for Edinburgh West, and founder of the All Party Parliamentary Group on Nuisance Calls, which in 2013 published its own report on this topic.

108. All periods in this annex are ranges of UK government financial years.

109. A different determination of the start dates of investigations suggests a mean length of 466 days.

110. The figures for lengths of investigations are available for only two investigations (both resulting in fines).

111. However, it is striking that the Ofcom period for monitoring silent and abandoned calls in an investigation is only 48 days out of the average 543 days taken by an investigation resulting in a fine.

112. CMRU had no powers to fine until the end of 2014.

113. Claims management regulation review: final report ( HM Treasury and the Ministry of Justice, 2016),

114. Freedom of Information request about Ofcom’s own initiative investigation into Verso Group ( UK) Limited (Ofcom, 6 January 2017),

115. Consumer protection test for telephone number allocation (Ofcom, 30 September 2008),

116. Consumer protection test for telephone number allocation (Ofcom, 30 September 2008),

117. ICO Disclosure Log – Response IRQ0628919 ( ICO, 8 June 2016),

118. ICO Disclosure Log – Response IRQ0681123 ( ICO, 19 May 2017),

119. See Section 8 of Enforcement guidelines for regulatory investigations (Ofcom, June 2017), Detailed criteria and procedures are laid down for Directions under General Condition 20.3.

120. Tackling nuisance calls and messages: Update on the ICO and Ofcom Joint Action Plan (Ofcom, December 2016),

121. Review of how we use our persistent misuse powers: Focus on silent and abandoned calls (Ofcom, December 2016) 2016),

122. This figure is the average of 5.8 and 10.0, which were the figures resulting from the March 2014 survey for respondents whose numbers were or were not registered with TPS (respectively).

123. This figure is for the November 2013 survey, when no respondents’ numbers were registered with TPS.

124. Nuisance Calls (Technical Measures) Memorandum of Understanding (MoU) (December 2015), The operators that signed the MOU are BT, Gamma, KCom, Post Office, Sky, TalkTalk, Virgin Media, EE, O2, Three and Vodafone. On a given day each month, each signatory collects Call Detail Records for all calls entering or leaving its network, and counts the most frequently occurring CLIs on these in various categories suggestive of nuisance calling, including malformed CLIs, Premium Rate CLIs, very short calls (<1 second), short calls (1-3 seconds), high ratio of unanswered calls and calls with no CLI digits.

125. More than two million now on BT's free service to crack down on nuisance calls ( BT, April 2017),

126. BT launches free service to crackdown on nuisance calls ( BT, January 2017),

127. See National Do Not Call Data Book FY 2016, p 5. The year as a whole showed record complaints, with around two-thirds of complaints being about recorded calls (often referred to as “robocalls”).

128. This trend is also taking place in North America, but relative costs to consumers of landline and mobile services make it less pronounced at this stage.

129. Ofcom’s Technology Tracker in early 2017 showed 81% of adults using smartphones.

130. Ofcom’s Technology Tracker in early 2017 showed 18% of those with a landline at home claiming not to use the landline for receiving calls. (They may even not have a phone plugged in to the line). The corresponding figure in 2015 was 8%.

131. The Fast.Map 10 th Annual Marketing GAP tracker, based on an online panel, showed small (typically well under 5%) and declining percentages of consumers who wanted to receive any kind of marketing phone call, for any purpose.

132. Until July 2013, the consumer issues survey results were similar to those of the landline nuisance call surveys, but their questions did not restrict attention to the previous four weeks (unlike those of the landline nuisance call surveys). After the questions were amended, fewer respondents to the consumer issues surveys reported that they had received nuisance calls.

133. The corresponding figure for the TPS effectiveness survey of March 2014 is 84%; it is the average of 77% and 91%, which were the survey results for respondents whose numbers were or were not registered with TPS (respectively).

134. There is no corresponding figure for the TPS effectiveness survey of March 2014, but 53% (which is the average of the survey results for respondents whose numbers were or were not registered with TPS) reported receiving at least six nuisance calls.

135. Fewer than 1% of the diarists reported more than 40 calls in any year except 2016, when fewer than 2% did so.

136. The survey answers reported until November 2014 related to receiving nuisance calls on landlines; after that, they related to receiving nuisance calls on landlines or mobile phones.

137. For instance, the 2013 landline nuisance call diary survey recorded that the proportions of nuisance calls per day were 20% on each of Monday and Tuesday, 18% on each of Wednesday and Thursday, 16% on Friday, 6% on Saturday and 2% on Sunday. Calling rates were fairly constant between 9:00 and 19:00 but fell off rapidly outside those hours.

138. The consolidation uses the full range of records available for standard trueCall units until October 2017 and the results of the landline nuisance call surveys covering four weeks in each year.

139. The “best” alignment in this case minimises the sum of the squares of the differences between the (Ofcom survey and trueCall unit) proportions of recipients of given numbers of nuisance calls.

140. Nuisance calls reported fall during the four weeks of each Ofcom landline nuisance call survey; this might reflect diarist fatigue or underlying fluctuations. The diary accommodates only 40 nuisance calls; anyone getting more than that is asked to record them on separate sheets or request another booklet, but this too could lead to under-recording. A diarist may also just overlook noting call details, especially when in the middle of doing something else, in a different room from the diary, or getting a series of calls.

141. In the North East (of England) the likelihood of receiving any calls was similar to that in Scotland, but the number of calls received was 6.6 per month.

142. This geometric progression in the proportions of units receiving calls is not followed at higher numbers of calls.

143. These are North East, North West, Yorkshire and Humber, East Midlands, West Midlands, East, London, South East and South West. Scotland has approximately (within 1/10) the same population as four of them, a higher population than two of them and a lower population than three of them.

144. The consumer issues surveys asked about nuisance calls received “in the last four weeks” only from July 2013 onwards. Earlier surveys asked the same question without a time limit, so with other things equal, answers should have been higher.

145. The surveys use samples, so though they have been weighted to be representative of the individual nations and regions they are subject to random variations. The results here are therefore expressed with ‘±’ alongside, conveying a 99% confidence interval. Essentially ‘±2%’, for example, means that if the survey occurred 100 times with different samples then in 99 of the occurrences the true value (which is assumed itself to be a percentage) would be within 2% on either side of the value estimated in the survey. The length of the confidence interval (which in this example stretches from -2% to +2%) depends on the size of the sample and the estimated value.

146. This would be so even for 95% and 90% confidence intervals.

147. Ofcom kindly provided this study with datasets enabling analyses going beyond the published results, which do not distinguish between Scotland, Wales and Northern Ireland, or between the nine regions of England, as in some cases the sample sizes are so small that the confidence intervals are unsound. However, Figure 70 does distinguish between them, for ease of comparison with our other figures. To a reasonable approximation Scotland, Wales and Northern Ireland contribute respectively 1/2, 1/3 and 1/6 of the population of the UK except England (and therefore weight the results of surveys accordingly).

148. The sample sizes for Northern Ireland are too small for the corresponding figures to be convincing or to justify identifying the ‘most affected’ and ‘least affected’ nations or regions.

149. The figures used are those for the months when the landline nuisance call surveys start, in 2014-2017, adjusted to cover four weeks (as in the Ofcom landline nuisance call surveys) instead of one month.

150. For the trueCall data the proportions for Scotland settle down to resemble those for the rest of the UK when the number of calls received is at least 26, while for the Ofcom data the proportions resemble each other only when the number of calls received is at least 36. This might reflect random variation in the Ofcom data: despite its aggregation of the data for 2013-2017 the Ofcom sample size for Scotland remains much smaller than the trueCall sample size for Scotland.

151. The consolidation uses the full range of records available for standard trueCall units until October 2017 and the results of the landline nuisance call surveys covering four weeks in each year.

152. For instance, the figures for Colonsay indicate that there are 89 TPS registrations (in 2017) and there were 124 inhabitants (in 2011).

153. These estimates are available at

154. Intermediate zones contain up to nine data zones. However, the algorithm is unable to match data zones to telephone areas more accurately than it can match intermediate zones to telephone areas, because the names of the data zones in an intermediate zone are just the name of the intermediate zone with numerical suffixes.

155. An alternative to using the number of TPS registrations in Scotland as a whole would be using the number of registrations in the council area of the main switching centre for the telephone area. However, there are no main switching centres in East Dunbartonshire and East Renfrewshire, so there are difficulties in constructing an algorithm that converges when iterated starting with this alternative.

156. The names of intermediate zones in West Dunbartonshire and East Lothian are completely unhelpful in this respect, as they are essentially just numerical. This does not appear to cause any seriously wrong assignments of place names to council areas, because the algorithm also uses the names of localities and settlements, for place names associated with each range of telephone numbers or, in the absence of matching with that range, for place names associated with a range enclosing that range.