Effectiveness of actions to reduce harm from nuisance calls in Scotland

Annex C: Estimates of the level of nuisance calling

C.1 Summary

As explained in the main text, to assess the effectiveness of actions to reduce nuisance calling, it is essential to know the level of nuisance calling in the first place, but currently there is no clear information on this. Some bases for estimates of the number of nuisance calls per year are:

• Complaint figures. These provide a lower bound, suggesting that the number of nuisance calls per year to landline and mobile networks is at least 3.2 billion.
• Survey results. Ofcom used the landline nuisance call surveys to estimate that the number of nuisance calls per year to landlines is 4.8 billion; an adjustment mentioned below would reduce this estimate to 4.4 billion.
• Network measurements. Ofcom’s statements suggest that the number of nuisance calls per year to landline and mobile networks is up to 9.4 billion. Various figures published by operators point to considerably higher totals, if extrapolated using market shares.
• User device measurements. Records from trueCall units located in users’ homes can be used to estimate that, after adjustment as below, the number of nuisance calls per year to landline networks is 4.5 billion.

C.2 Estimation techniques

C.2.1 Complaint figures

In 2016 ICO, Ofcom and TPS together collected about 320,000 complaints about nuisance calls, some of which might be made to more than one organisation or in more than one way ^[120] . This total can be used to give a simple floor estimate: the number of nuisance calls per year to landline and mobile networks is at least 3.2 billion, given that, as shown in Figure 28, for the investigations that ICO has completed, the ratio of complaints to nuisance calls is less than 0.01% (and this is probably an over-estimate, as ICO are more likely to investigate cases that attract higher levels of complaint).

C.2.2 Survey results

Ofcom estimated the number of nuisance calls per year to landline networks to be 4.8 billion in 2015 ^[121] . Underlying the calculation were assumptions that each diarist in the 2015 landline nuisance call survey received 8.4 nuisance calls in four weeks, that 84% of the 52.02 million adults in the UK in 2015 had access to a landline (in accordance with the 2015 technology tracker survey), and that a diarist received on average the same number of nuisance calls as any other adult; then the number of nuisance calls per year to landline networks was 8.4*52/4*52.02*84% or 4.8 billion.

The assumption that a diarist received on average the same number of nuisance calls as any other adult is questionable. The landline nuisance call surveys do not determine what proportion of calls to the landline in question a diarist receives in comparison with other household members. Of course, people in the vulnerable groups on whom intensive nuisance calling tends to be focused (as discussed in section 2.3.3) are unlikely to be recruited as diarists.

The Ofcom calculation that led to 4.8 billion can be adjusted to limit the lines to those used for receiving calls (which were 92% of the total in the 2015 technology tracker survey); it then leads to 4.4 billion. The assumptions underlying the calculation, adjusted like this and adapted to other years, can support analogous estimates of the numbers of nuisance calls derived from the other landline nuisance call surveys and from the 2013-2014 TPS effectiveness surveys. Figure 40 summarises the estimates. The estimates from the TPS effectiveness surveys are higher than those from the landline nuisance call surveys of the same years; they can be regarded as estimates of the numbers of nuisance calls that there would be if half (in the 2014 TPS effectiveness survey) or all (in the 2013 TPS effectiveness survey) of the landlines were not registered with TPS.

Figure 40: Estimates of nuisance calls per year to landlines

Year	Survey type	Mean number of nuisance calls per adult in four weeks	Numbers of adults (millions)	Proportion of adults with landlines	Proportion of landlines on which calls are received	Estimated number of nuisance calls to landlines per year (billions)
2017	Landline nuisance call diary surveys	6.8	54.09	82%	83%	3.5
2016		7.3	53.26	86%	89%	4.0
2015		8.4	52.02	84%	92%	4.4
2014		7.3	51.59	84%	94%	4.0
2013		6.9	51.26	85%	92%	3.7
2014	TPS effectiveness survey	7.9 [122]	51.59	84%	94%	4.2
2013		11.9 [123]	51.26	85%	92%	6.2

Sources: Ofcom landline nuisance call and TPS effectiveness surveys, other Ofcom data, ONS, this study

The estimates in Figure 40 might suggest that the numbers of nuisance calls rose over the last few years but are now are falling. However, this conclusion is suspect, because:

• There is a month-by-month variation in the number of nuisance calls (superimposed on a seasonal variation), as shown in Figure 46. The TPS effectiveness surveys themselves demonstrate the seasonal variation rather markedly: the mean number of nuisance calls recorded was 11.9 in November 2013 and 10.0 in March 2014 for the landlines not registered with TPS.
• The diarists for the TPS effectiveness surveys were selected from among people whose landlines were not initially registered with TPS; they may therefore have been atypical of the whole population somehow, and their nuisance call pattern likewise. In any case, though in the second effectiveness survey half of the landlines had been registered, in the first TPS effectiveness survey none had been registered and the mean number of nuisance calls recorded would exceed that appropriate to the whole population, given that TPS registration does indeed have an effect.

C.2.3 Network measurements

Monthly co-ordinated measurements by the signatories have enabled Ofcom to estimate the number of potential nuisance calls on the networks covered by the MOU ^[124] . That estimate suggests that there are 22 million nuisance calls per day; the MOU networks have at least 85% of all landline subscribers and 85% of all mobile subscribers, so there could be up to 9.4 billion nuisance calls per year to landline or mobile networks. However, the estimate is based on aggregating all the MOU returns; it could be an underestimate, because:

• It largely picks up bulk automated dialling.
• It provides information about only the most frequently presented CLIs in various categories.
• We have heard that some smaller networks that are not MOU signatories have customers who originate disproportionate numbers of nuisance calls; such networks may see low origination charges for large call volumes as part of their competitive appeal.

Though some calls included in the MOU returns might not be nuisance calls, in our judgement there are probably more that are not included that are nuisance calls, since the measurement method excludes many originators of smaller numbers of nuisance calls (see Figure 20).

Ofcom say that the MOU returns indicate that the number of nuisance calls per month has been fairly stable since measurements started in early 2016.

Operators are wary of making data about nuisance calls public. However, some inferences can be made from what they do publish and the numbers of potential users (on specific operator networks and for specific network technologies). For example:

• BT has said that if all its customers requested it to divert nuisance calls it would divert 65% of them, which would mean finding 1.6 billion nuisance calls per year on its network, including calls carried for other operators ^[125] . We do not know the proportion of traffic on BT’s network which is destined for other networks, so we cannot extrapolate from this figure. However, additionally BT has said that adults on its network aged between 16 and 24 receive about 3 nuisance calls per week while those aged between 55 and 64 receive about 5 nuisance calls per week ^[126] ; this points to an average of at least 16.0 nuisance calls in four weeks for each adult. Under the same basic assumptions in the Ofcom calculations there would be at least 9.1 billion nuisance calls per year to all landline networks or, adjusted to limit the lines to those used for receiving calls, 8.4 billion.
• TalkTalk has said that it blocks 100 million nuisance calls per month. If this represented 50% of nuisance calls (as has been suggested), then given its market share there would be 17.6 billion nuisance calls per year to landline networks, unless TalkTalk is disproportionately targeted (which could reflect its customer profile or data leaks). TalkTalk is now also suppressing all calls without CLIs, which it believes amount to another 8 million calls per month (though some of those might not be nuisance calls).
• Vodafone has said that it blocks between 2.0 million and 2.5 million nuisance calls per day. It has not said what proportion of nuisance calls this is; a proportion of 65% would imply that there are between 5.9 billion and 7.4 billion nuisance calls per year to mobile networks. A proportion of 100% would imply that there are between 3.8 billion and 4.8 billion such calls.

C.2.4 User device measurements

User device measurements of the numbers of nuisance calls are like network measurements, except that they are performed by equipment where the users receive calls. The trueCall “standard” and “vulnerable” units have been collecting data on nuisance calls for some years. Figure 41 summarises the results of using the data collected for the standard units to find the number of nuisance calls received. The vulnerable unit users are likely to be vulnerable people and typically receive many more nuisance calls than the standard unit users, so they are not suited to estimating the number of nuisance calls received by the population as a whole.

Figure 41: Nuisance calls per year to landlines based on standard trueCall unit data

Year	Mean number of nuisance calls per unit in four weeks	Number of landlines (millions)	Proportion of landlines on which calls are received	Estimated number of nuisance calls per year to landlines (billions)
2017	16.3	26.4	83%	4.6
2016	17.8	26.1	89%	5.4
2015	21.3	25.5	92%	6.5
2014	23.1	25.0	94%	7.1

C.3 The relative contributions of landline and mobile calls

In 2016 there were 26.4 million residential landlines, 7.1 million business landlines and 92.0 million mobile subscriptions. If outbound calls from UK call centres had been distributed evenly across all these, 73% of them would have been made to mobiles. In fact only 58% of them were made to mobiles (in 2015), so there is an imbalance of about 43%: for every 100 such calls received by mobiles, about 143 [that is, 100*(73%/58%)((100%-73%)/(100%-58%)] such calls are received by landlines.

The number of nuisance calls per mobile user appears to be lower than the number of nuisance calls per landline user in line with this imbalance, in that according to the only available relevant data, from the Ofcom consumer issues surveys, for every 100 nuisance calls to mobile users there are on average 145 (between 120 and 182) nuisance calls to landline users. This is illustrated in Figure 42, which also suggests that the imbalance has reduced somewhat since 2013.

Figure 42: Landline and mobile users receiving nuisance calls in the previous four weeks

Source: Ofcom consumer issues surveys