[doc. web n. 3318874

VERSIONE ITALIANA

General Injunction Regarding the So-Called "Silent Calls" - 20 February 2014
Published in Italy´s Official Journal No. 79 of 4 April 2014)

THE ITALIAN DATA PROTECTION AUTHORITY

Having convened today, in the presence of Mr. Antonello Soro, President, Ms. Augusta Iannini, Vice-President, Ms. Giovanna Bianchi-Clerici and Prof. Licia Califano, Members, and Mr. Giuseppe Busia, Secretary General;

Having regard to the Personal Data Protection Code (legislative decree No. 196 of 30 June 2003, hereinafter the "Code"), in particular to Section 11 thereof where some key principles are laid down, among other things, in respect of the processing of personal data including fairness and lawfulness principles;

Having regard to the decision by the Italian Garante (No. 474 of 6 December 2011), published at www.garanteprivacy.it as web doc. No. 1857325, whereby the Garante required Enel Energia S.p.A. and Reitek S.p.A. to take several measures including those relating to "silent calls";

Having regard to the judgment by the Court of Rome No. 18977 of 26 September 2013, whereby the appeal lodged by Enel Energia S.p.A. and Reitek S.p.A. against the aforementioned decision was rejected and the arguments made therein were upheld in full along with the measures issued vis-à-vis the relevant companies, which were awarded legal costs;

Having regard to the decision by the Italian Garante (No. 482 of 30 October 2013 – web doc. No. 2740497), whereby the Garante adopted a draft general decision on silent calls and jointly launched a public consultation on the measures outlined therein by publishing the relevant notice on Italy´s Official Journal (No. 274 of 22 November 2013);

Having regard to the outcome of the public consultation, which was aimed at "acquiring comments and observations on adequacy of the proposed measures and the respective implementing mechanisms along with additional operational proposals";

Having regard to, in particular, the contributions submitted within the 60-day deadline by several industry associations and business entities as well as by consumer associations and individual data subjects;

Having regard to the records on file;

Having regard to the considerations submitted by the Office via the Secretary General under Section 15 of the Garante´s Rules of Procedure No. 1/2000 of 28 June 2000;

Acting on the report submitted by Mr. Antonello Soro;

PREAMBLE

A considerable number of complaints were lodged with the DPA by data subjects against the reception of unsolicited calls; such complaints have concerned a substantial number of the so-called "silent calls" ever since the second half of 2011 in addition to operator-assisted marketing calls. "Silent calls" are calls where the called person, after picking up the receiver,  is not put through to any speaker. The reiterated, continued reception of such calls, even 10 to 15 times in a row and often over a long time span, has been a severe nuisance to the recipients, who have been unable to rely on whatever protection or remedy exactly  because of the lack of whatever counterpart.

Silent calls, as also shown by several complaints and many blogs addressing this topic on the Internet, may give rise to anxiety and alarm in the called party, who wonders about the caller and may be dismayed both because one is naturally bound to make an association between such calls and illicit conduct (undue surveillance, harassment, reconnoitering by crooks prior to the commission of thefts or attacks, etc.) and because one feels unpleasantly prevented from getting in touch with someone who may have high-import information to convey. It was not seldom the case that the complaints mentioned significant circumstances like the fact that the called party´s children were not at home when the call was received, or the calls had targeted the complainant´s elderly parents living on their own, or they had been taken by family members who were sick at that time, etc. .

At all events, anxiety and nuisance are compounded by the frustration arising from powerlessness and the inability to react.

Silent calls have been spotlighted in other countries as well, where basic measures have been taken to regulate them. Reference can be made in this regard to the specific requirements laid down in the document drafted by the UK Ofcom (i.e. the independent regulator and competition authority for the UK communication industries) on 1 October 2010, called "Tackling Abandoned and Silent Calls" (http: ….) and to those set forth in the document by the US FCC (Federal Communications Commission) of 15 February 2012 called "Report and order in the matter of rules and regulations implementing the telephone consumer protection Act of 1991" (http: ….).

In tackling the cases at issue, the sole purpose aimed at by the Italian Garante is an institutional one and consists in taking due account of the claims made by data subjects and their legitimate expectations to be protected; still, this DPA believes that, in so doing, it will also protect the effective operational management of telemarketing operators, since "silent calls" do often result into jeopardizing whatever chance for the data subject to listen to and accept future marketing proposals.

This is why, especially in the medium to long term, implementing measures and arrangements that can bring silent calls back into physiological tolerance limits serves the interest vested in all the parties towards adopting marketing practices that are more virtuous – i.e. less invasive and more efficient.

Ultimately, the Garante is stepping in to regulate and limit the recourse to certain mechanisms for processing the data of telemarketing addressees that can account – as will be shown below – for the occurrence of "silent calls". The mechanisms in question are such as to give rise to the highly intrusive events mentioned above and may be in breach of Section 11 of the DP Code. The latter requires – under penalty of a ban on using the data – that personal data be processed "fairly and lawfully". This view was also endorsed by the aforesaid judgment of the Court of Rome, rejecting the appeal by Enel Energia S.p.A. and Reitek S.p.A. against the DPA´s decision of 6 December 2011; the Court stated in this regard that "using personal data to perform a silent call rather than to submit a marketing proposal is in breach of the fundamental fairness standard mentioned in Section 11 of the Code, given that the whole system for selecting and placing calls... is geared towards enhancing the success rate of the calls switched to the operators whilst all the risks and the inconvenience resulting from silent calls are shifted over to the called parties."

An especially aggressive marketing policy that is only aimed to ensure effective operation of a call center to the data subjects´ detriment by using their personal data indiscriminately and unfairly may hardly be considered in line with the above legal requirements – partly on account of the major impact caused by such a policy and its possible degenerations. This holds true also if the contacted party has given his or her prior consent to the use of their data for marketing purposes or – as regards phone directory data – if the user at issue has not signed in to the opt-out register.

The data subject is unquestionably the master of the personal information relating to him or her, as shown, inter alia, by his or her being empowered under the law to enable others to use such information in accordance with pre-defined mechanisms and arrangements for explicit, legitimate and specific purposes; however, this does not mean that the data subject may allow, via his or her consent, processing operations of his or her personal data that are carried out according to unlawful or unfair arrangements.

Partly because of the above reasons, the latter operations include  the processing related to the use of the data subjects´ personal data when performed through arrangements that - at least based on the findings gathered in some cases - do not produce the effects they are meant to achieve, i.e. making operator-assisted phone calls for marketing purposes, as they may actually give rise to different events that currently fall outside the scope of whatever regulations. Reference is made exactly to "silent calls": these are especially intrusive nuisance calls that, though intended for the purposes typically underlying telemarketing, i.e. enabling a conversation between an operator and a prospective customer for the marketing of goods or services, are ultimately stripped of such features because they turn actually into attempts to get in touch with the called party and are not followed by the striking of whatever conversation with an operator or the submission of a marketing proposal.

As already pointed out, this phenomenon showed a marked, significant increase in terms of frequency especially over the past few months and required complex investigational activities to be performed – first and foremost in order to establish whether some of the calls at issue could be traced back to the actual originators, that is to specific entities.

The investigations, including inspections, carried out by this Office were of a fact-finding nature as well as being focused on monitoring and assessing, also from a technical standpoint, and verifying compliance with personal data protection legislation. Specific attention was paid to the taking of measures such as to potentially generate silent calls and to compliance with Section 11 of the Code as for the processing of personal data relating to recipients of operator-assisted marketing calls.

The checks performed highlighted, in the first place, that the calls complained about had been carried out by call centers for marketing purposes with the help of automated call routing systems – which is by now standard practice.

This DPA could establish that in most cases the list of possible recipients are uploaded to the IT platform used by call centers – this may be either a proprietary platform or one that is supplied by a third party; the platform then dials phone numbers via ad-hoc software and allocates the calls to the individual call center operators also based on their being available and physically present at the respective workstations.

Thus, an automated system is used to generate phone calls to subscribers such as to keep on hold those calls that are answered by the recipients until an operator becomes available to take the calls. This is clearly intended to enhance the efficiency of call enters by shifting wait-time costs compared to non-automated call generation systems. Whilst wait-costs are borne fully by the caller in the latter systems, since it is impossible for the caller to foresee whether the call will be successful, automated systems allow shifting the costs to the recipients to the extent they answer the calls but then do not find anyone speaking at the other end of the line. By the same token, it is the user/subscriber that is exposed to the risk addressed in this decision – namely, that the call is not forwarded to any operator after been kept on hold for some time and although the recipient has already answered it, because no operator is available at that time, so that it will be ultimately "terminated" either by the user/subscriber or by the automated system.

Phone calls are made by the system dynamically, i.e. adjustments and corrections are made along the way to take account of several real-time parameters such as the number of operators working in the call center at a given time.

It was also found that several variants (algorhythms) of the standard model are available even though the operational arrangements are mostly those described above; it is thus possible to change the initial default parameters. It is exactly because of the setting of such parameters that degenerative effects may be (or are) produced such as the excessive frequency of silent calls or the uncontrolled repetition of attempts to call one specific user/subscriber.

Algorhythms are selected as a function of the  business model endorsed as part of a telemarketing campaign; however, it was found that a key role is played in the overwhelming majority of cases exactly by the call centers: since they are contractors to the company waging the campaign, they are highly keen on devising and implementing technical and operational arrangements that maximize the cost-effectiveness of the systems they manage.

Thus, practically speaking, it is the call center that sets the number of prospective contacts to be achieved by the system and can customize algorhythms  accordingly in many different ways; to that end, the call center can decide to perform and forward a number of phone calls that may be considerably in excess of its own processing capacity in order to make sure that its operators will always find a call kept on hold immediately they complete the previous call. The operators will thus take up the new call and will not remain inactive as no idle time (or downtime) will be generated. It should also be considered that not all the calls made by the system are successful, since the recipients might obviously fail to answer, be busy with another call, not be at home, have an answering machine or a fax in operation, etc. .

It is partly for the above reasons – namely, to prevent any obstacles to the generation and success of phone calls from affecting telemarketing operators in terms of downtime, i.e. inefficiencies – that the automated systems are devised technically to enable call centers to make their own selections by setting and configuring the algorhythms that are considered to be functionally best suited for achieving the pre-determined targets.

In other words, a call center may decide to rely on a statistical prediction standard based on the parameters selected in the given case, so that the system will forward a number of phone calls that exceeds operators´ processing capabilities. The ultimate effect, whether intended or not, is exactly that an operator will not always be available to take the call: in such a case, the recipient will get a "silent" call. Since the objective of setting up a communication between the two partners was not achieved, the call may be reiterated as a function of the recall policy implemented by the call center.

Furthermore, the investigation by the DPA showed that calls reported as "silent" in complaints were not always classed as such in the call records kept by call centers and accessed by the DPA during the inspections.

That is to say, whilst the reporting of events dovetailed in full – a given call had been actually made on the given day and time by a specific call center workstation and addressed to the complainant´s number – the categorization of the outcomes was not always fully superimposable. Thus, a call referred to as "silent" by the complainant corresponded, for instance, to a call made to a busy number in the call center´s report.

This is a key point, since possible inconsistencies in categorizing the outcomes of individual calls result, first and foremost, into preventing the objective assessment of fundamental markers – such as whether silent calls did occur and how frequently.

The criticality at issue is definitely a substantial one and has been encountered in several of the inspected call centers as well as in a non-negligible number of sampled cases. All these factors point to the likelihood that the inconsistencies are due to different categorization criteria as applied to the same event by the entities involved in the overall process.

There are actually two separate entities involved in categorization, acting independently of one another – namely, the public telephone services operator, which runs the telephone exchange where the subscriber´s line is terminated, and the call center making the marketing calls and relying on its own call management systems. These two entities are connected via the telephone line. The former attaches a code tag to each possible outcome of the calls according to standardized, internationally received coding standards; the code tag is then forwarded to the call center. The latter does not apply any pre-defined or standardized criteria in receiving and interpreting this piece of information. Accordingly, one cannot rule out that the call center construes the code differently from what was meant – without this being in any way intentional or due to negligence; thus, a certain event may be misinterpreted, which impacts significantly – as already pointed out – on the expected reliability of the assessment concerning "silent" calls.

In factual terms, it may then be the case that a "silent" call, i.e. a call that was answered by the recipient, is interpreted differently by the call center´s reporting system – e.g. as a "busy line" call, or as a call terminated by the exchange because of technical reasons.

In this connection, one should draw a fundamental distinction between "successful" calls, i.e. calls that are answered by the called party and are usually charged to the calling party, and "silent" calls -  i.e. calls that, though answered by the called party and therefore being technically successful calls, are terminated by the called party or the call center´s management system before a call center operator becomes available. Hence, "silent" calls make up a sub-set of "successful" calls and may only exist as such; indeed, successful calls can also be said to include other categories such as the calls answered by fax or answering machines.

Since the intervention by an operator in "successful" calls can be easily established by the call center  - in fact, this is in the call center´s interest from a business standpoint – the DPA should focus on the measures required to streamline the categorization of call outcomes that might go missing in the relationship or, more specifically, in the communication between telephone operators and call centers as described above. Moreover, since the possible causes of "silent" calls as mentioned above include rare, statistically negligible events -  especially if these events have allegedly to do with technical factors such as the termination of a call by the public telephone exchange -  and given that such events do not impact on the estimated incidence of silent calls in whatsoever manner, the events in question will not be taken into consideration as concurrent causes in calculating the percent tolerance thresholds outlined in the paragraphs below.

The assessment in question may also be affected by the choice of the time span for measuring frequency (and therefore incidence) of "silent" calls. Finding a reduced frequency of "silent" calls over "long" periods does not rule out that the relative frequency of such calls during specific sub-periods may actually be above average – at times significantly so.

Similarly to what has been decided in other countries, it is therefore necessary for the quantitative assessment of this phenomenon to be performed by having regard to the arithmetical ratio between "silent" and "successful" calls, which ratio should be calculated by having regard to each individual telemarketing campaign. In this manner, it will not be possible for a high rate of "silent" calls to feature in a single campaign by a call center and to be "set off" by calculating the mean value of such calls over a larger number of campaigns. For the same reasons it is also necessary for the timeframe of the assessment to be defined beforehand.

In short, it is necessary to perform a substantial, difficult as well as necessary exercise in balancing the interests at issue – namely, the protection to be afforded to data subjects´ rights; the business and occupational efficiency to be achieved by the industry; and the effects produced by technological developments as applied to telemarketing. This means that it is necessary to regulate the mechanisms for processing the personal data of the addressees of telemarketing campaigns so as to make them lawful and fair, i.e. to bring them into line with the provisions contained in Section 11 of the Code. To that end, the appropriate measures should be laid down to allow, first of all, gauging and, secondly, preventing or at least reducing to a reasonably tolerable level the nuisance caused by "silent" calls. This is all the more necessary given the current regulatory gap in this sector, which allows "physiological" occurrences to degenerate into unacceptable pathological events; furthermore, one should also consider that in assessing – as part of its competences - whether the processing mechanisms are compliant with lawfulness and fairness requirements, the Garante has carefully evaluated the multifaceted issues related to this specific topic before issuing each of the measures set forth below.

BASED ON THE ABOVE PREMISES, THE GARANTE NOW

Under Sections 143(1), letter b), and 154(1), letter c), of the Code orders all the data controllers determining the mechanisms for the processing of personal data relating to the addressees of telemarketing initiatives to take, either directly or by the agency of their processors as duly instructed, all the measures that are both necessary and appropriate, including technical measures, to ensure that the processing in question is performed according to arrangements that are compliant with the fairness principles referred to in Section 11 of the Code. As regards specifically "silent" calls:

1) When acquiring the codes corresponding to call outcomes as transmitted by public telephone exchanges, call centers shall create a category including all the "successful" calls; the latter category shall be broken down into two sub-sets, of which one, to be possibly termed "Set A", shall include "silent" calls – i.e. the calls that failed to be taken up by call center operators within 3 seconds, after which the calls must be "terminated" by the system; the latter threshold is considered to be appropriate given that the interactivity of a telephone conversation is preserved for as long as the 3-second threshold is not overstepped. The other sub-set, to be possibly termed "Set B", shall include all the remaining types of "successful call".

2) The mean percentage of permitted "silent" calls (termed P) as computed in accordance with the arithmetical ratio between the number of "Set A" events (NA) and the sum of such events plus "Set B" ones (NB) pursuant to the distinction drawn under 1) above, shall in no case exceed 3%. In practice, the mean percentage of permitted "silent" calls shall be calculated as follows: P = NA / (NA+NB).

Additionally, the above percentage shall be computed by having regard to each individual telemarketing campaign, which shall be uniquely identified. At all events, the computation shall start at the beginning of the campaign and must be repeated at least at 10-day intervals; the latter time span is considered to be appropriate also in the light of the balancing of interests performed between the need for carrying out the assessment in question without overstretching the reference period and the need for enabling call centers to be operational. Accordingly, if the duration of a marketing campaign is above 10 days, the assessment shall be performed with regard to the first 10-day period and repeated thereafter at 10-day intervals or at shorter intervals where applicable – until conclusion of the campaign. In order to limit variability and oscillations of percent values and the occurrence of a number of silent calls well above the threshold mentioned in the foregoing paragraphs, the mean daily percentage of silent calls as calculated via the said formula shall not be in excess of 4%.

3) Data controllers are required to implement a technical measure called "comfort noise", either directly or by instructing their data processors accordingly: the call center shall transmit, to the called party, a prerecorded audio file reproducing synthetic environmental noise. In practice, for each successful call that cannot be taken up by a call center operator, the call management system must ensure that the said background noise is heard immediately instead of the operator´s voice – i.e. immediately the called party picks up the receiver. This is meant to mitigate the "silent" call effect and reduce the called party´s anxiety or disquiet. The comfort noise must actually be devised in such a way as to give the impression that it is produced in a workplace – e.g. it could include voices in the background, phones ringing, murmurings, etc. – so that the called party, albeit not speaking to an operator, can feel that the call is coming from a call center and thus rule out any malicious intent on the unknown caller´s part.

4) If a "silent" call takes place, no recalls of the given user shall be permitted for at least five days; the latter period is considered appropriate partly in the light of the fifteen-day term for using the data extracted from phone directories after matching them with the Public Opt-Out Register as well as by taking account of the mean duration of telemarketing campaigns  - which is usually 30 to 60 days  based on the inspections carried out by the Garante. Furthermore, the recall must be performed in such a way as to make sure that the call is routed on a priority basis – that is, availability of an operator must be ensured before the call is made.

5) Call centers are required to keep the statistics of the percent "silent" calls performed during each campaign -  which must be structured as explained under 2) above, i.e. they must include the unique identification codes of each individual telemarketing campaign – for no less than 2 years so as to allow such checks and verifications as may be deemed appropriate.

6) The deadline for implementing the measures set out under 1) to 5) above shall be 180 days as from the date of publication of this decision in the Official Journal of the Italian Republic.

As regards specifically silent calls, this general decision shall supplement and replace, where applicable, decision No. 474 of 6 December 2011 as issued with regard to Enel Energia S.p.A. and Reitek S.p.A..

This decision may be challenged under Section 152 of the Code and Section 10 of legislative decree No. 150/2011 by seizing a judicial authority, in particular the court having jurisdiction over the data controller´s place of residence, within thirty days as from the date it is notified, or within sixty days from the latter date if the appellant is resident abroad.

A copy of this decision shall be forwarded to the Ministry of Justice-Office for the publication of laws and decrees in order for it to be published in the Official Journal of the Italian Republic.

Done in Rome, this 20th day of the month of February 2014

THE PRESIDENT
Soro

THE RAPPORTEUR
Soro

THE SECRETARY GENERAL
Busia