Copyright Board Canada

Commission du droit d’auteur Canada

[PUBLIC VERSION]

Date	2023-12-01
Citation	Re:Sound Tariff 8 (2013-2018), 2023 CB 12
Members	The Honourable Luc Martineau René Côté Nathalie Théberge
Proposed Tariffs Considered	Re:Sound Tariff 8 – Simulcasting, Non-Interactive Webcasting and Semi-Interactive Webcasting (2013, 2014, 2015) Re:Sound Tariff 8 – Non-Interactive and Semi-Interactive Webcasts (2016, 2017, 2018)

Approval of Proposed Tariffs

As
Re:Sound Tariff 8 – Non-Interactive and Semi-Interactive Streaming (2013-2018)

REASONS FOR DECISION

I. overview

[1] Re:Sound is the collective society that administers the rights to perform in public and to communicate to the public by telecommunication published sound recordings for performers and makers of sound recordings.

[2] Re:Sound filed with the Board the following Proposed Tariffs:

- Re:Sound Tariff 8 – Simulcasting, Non-Interactive Webcasting and Semi-Interactive Webcasting (2013, 2014, 2015);

- Re:Sound – Non-Interactive and Semi-Interactive Webcasts (2016, 2017, 2018).

[3] The Proposed Tariffs cover semi-interactive and non-interactive online music services. Online music services provided by non-commercial entities other than the Canadian Broadcasting Corporation (the “CBC”) are subject to a different tariff.[1]

[4] On November 30, 2021, Re:Sound, Stingray Group Inc. (“Stingray”) and the Canadian Association of Broadcasters (the “CAB”) jointly requested that the Board approve a tariff based on a specified set of royalty rates, terms, and conditions (the “jointly-submitted text” or “JST”).

[5] The JST maintains a per-play rate expressed in cents. This fixed amount is owed each time a music digital file is communicated to a single person.[2]

[6] The JST is the result of an agreement between Re:Sound and the CAB, and Re:Sound and Stingray (collectively, the “RCS Agreements”). In fact, the JST is part of, and annexed to the RCS Agreements.

[7] We approve the proposed tariff, based on the JST, with modifications. The per-play rates that we approve are as follows:

Table 1: Approved per-play rates

Period	CBC	Non-interactive	Semi-interactive
January 1, 2013 to August 12, 2014	$0.000131	$0.000102	$0.000089
August 13, 2014, to December 31, 2017	$0.000193	$0.000193	$0.000193
January 1, 2018 to December 31, 2018	$0.000208	$0.000208	$0.000208

[8] The approved rates are an increase from the last-approved tariff, and essentially account for (i) inflation and (ii) repertoire-use adjustments due to statutory modifications (as of August 13, 2014). This change increased the number of foreign sound recordings eligible for equitable remuneration in Canada and represented by Re:Sound. No change was proposed compared to the last-approved annual minimum fee of $100, so it is accepted as proposed.

II. background

[9] The joint request was that the Board approve Re:Sound Tariff 8 (2013-2024) based on the JST.

[10] Alternatively, the parties to the RCS Agreements requested that, should the Board decide it cannot approve the JST for the full requested period of 2013-2024 at this time, the Board approve it for 2013-2018 (with the inflationary adjustment pro-rated to the end of 2018 instead of 2020) and establish a separate written proceeding for approval of the 2019-2024 period.

[11] The Board ruled that it would first deal with the 2013-2018 period and deal through a separate process with the 2019-2024 period.[3]

[12] The JST is based on the last-approved tariff, but includes new administrative provisions for the purposes of greater clarity and consistency with more recently approved tariffs.

[13] The JST rates are significantly lower than the initial rates proposed by Re:Sound in the Proposed Tariffs.

[14] Pandora and SiriusXM—respectively participating objector and intervener—do not dispute that legal changes (as of August 13, 2014) should have an effect on the rates. Pandora and SiriusXM, however, disagree with the scale of the proposed increase from the last-approved tariff.

[15] Table 2 compares the rates approved in Re:Sound - Tariff 8 (Non-Interactive and Semi-Interactive Webcasts), 2009-2012 (the “Last-Approved Tariff”),[4] and the 2013-2018 JST rates.

Table 2: Last-Approved Tariff / JST rates as $ per music file streamed

Period	CBC	Non-interactive	Semi-interactive
Last-Approved Tariff	$0.000131	$0.000102	$0.0000102
January 1, 2013 to August 12, 2014	$0.000131	$0.000102	$0.000089
August 13, 2014 to December 31, 2017	$0.000202	$0.000202	$0.000202
January 1, 2018 to December 31, 2018	$0.000222	$0.000222	$0.000222

[16] The JST rates constitute, in general, an increase from the Last-Approved Tariff, as shown in Table 3.

Table 3: Change from Last-Approved Tariff to JST rates

Period	CBC	Non-interactive	Semi-interactive
January 1, 2013, to August 12, 2014	0%	0%	- 13%
August 13, 2014, to December 31, 2017	+ 54%	+98%	+98%
January 1, 2018, to December 31, 2018	+69%	+118%	+118%

III. issues

[17] We have identified the following key issues:

1) What should the unadjusted royalty rate be?

2) What should the repertoire-use adjustment be?

3) Should the tariff account for partial plays and free trials?

4) Should the 2018 rate be adjusted for inflation, as proposed by Re:Sound?

5) Should proposed wording changes be accepted?

IV. analysis

A. Issue #1: What should the unadjusted royalty rate be?

[18] We must first determine a starting point for the royalty rate. The starting point is referred to as the “unadjusted royalty rate.”

[19] We use the same unadjusted royalty rate that led to the Last-Approved Tariff, being $0.00026222 per-play,[5] as a proxy for the royalty rate, before any adjustments are applied to account for sound recordings not represented by Re:Sound.[6]

[20] In this proceeding, the unadjusted royalty rate is not disputed by the Parties. For example, Re:Sound submits that the JST “reflects the same effective rate and rate structure established by the Board when it approved the [Last-Approved Tariff].”[7]

B. Issue #2: What should the repertoire-use adjustment be?

Finding

[21] The only significant change in the relevant market, of which there is evidence, is the change in Re:Sound’s eligible repertoire. Relying on the combination of two points of reference, we apply a 73.7% repertoire-use adjustment to the unadjusted royalty rate. This results in a royalty rate of $0.000193 per-play, before any other adjustments are made.

Background

[22] A significant portion of recorded music from other countries, including the U.S., became eligible for remuneration in Canada on August 13, 2014.[8] As a result, Re:Sound’s eligible repertoire in the Canadian marketplace increased since the Last-Approved Tariff. A new repertoire-use adjustment is therefore required.[9]

Re:Sound Submissions

[23] Re:Sound argues that the size of its updated repertoire (particularly as a result of the inclusion of U.S. sound recordings) should be determined by using the JST as a proxy, which results in a 77% repertoire-use adjustment. This method leverages the 2021 agreements with the CAB, representing Bell Media and its iHeart Radio streaming service, and with Stingray. (The RCS Agreements)

[24] While the RCS agreements do not specify a repertoire-use adjustment, Re:Sound explains that the rate increase is mainly attributable to a “repertoire increase.” Calculating the changes from the Last-Approved Tariff, it corresponds to a 77% repertoire-use adjustment for the period of August 13, 2014, to December 31, 2019.[10]

[25] Re:Sound further explains the rationale of the RCS Agreements as far as repertoire is concerned: “[T]he settlement […] on repertoire was part of an overall compromise intended to simplify this proceeding and avoid the costs of a full hearing including the costs of an in-depth repertoire study and audit, in light of the low revenues generated by Tariff 8 (||||||||||||||||||||| |||||||||||||||||).”[11]

[26] Alternatively, should the Board choose a different method, Re:Sound recommends a specific repertoire-use study it commissioned and filed.

[27] This study was performed in June 2021, by Benoît Gauthier (Circum Networks) and Doris Tay (Re:Sound’s Vice-President of Distribution). (The “RS8 study”)[12] The RS8 study suggests that 95.1% (non-interactive) and 92.6% (semi-interactive) of the content streamed by Tariff 8 users is represented by Re:Sound. For reference, a repertoire-use adjustment of 38.9% was applied in the Last-Approved Tariff.

[28] Re:Sound further claims that the results of the RS8 study (a 92.6% to 95.1% repertoire-use adjustment) demonstrate that the 77% adjustment from the RCS agreement is conservative. An audit is therefore not necessary to confirm that a repertoire-use adjustment of 77% is demonstrably reasonable.

[29] With respect to CBC, Re:Sound claims that the rationale adopted in the Last-Approved Tariff for applying a specific repertoire-use adjustment to the CBC (compared to commercial webcasters) is no longer relevant. The Board determined that, because of its mandate, the CBC played more Canadian music than commercial webcasters. This meant that more content from ReSound’s was played compared to the commercial services. Now, Re:Sound states: “With the expansion of Re:Sound’s repertoire and U.S. repertoire now being eligible, a separate repertoire-use adjustment for CBC is no longer necessary.”[13]

Pandora Submissions

[30] Pandora critiques both methods of calculating the repertoire-use adjustment. Pandora argues that the repertoire-use adjustment should be between 63.5% and 70%.

[31] Regarding the RCS Agreements, Pandora raises three main issues. First, the agreements are not representative of the market (|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||). Second, they are not agreements on repertoire-use adjustment, but rather on a royalty rate. Third, even if the Board preferred to use the RCS Agreements, the implied repertoire-use adjustment in those agreements is 70% once inflation is removed from the per-play rate.

[32] Pandora explains the latter point as follows: The unadjusted royalty rate was 26.2¢ per thousand plays through 2012. Inflation from January 2013 to June 2022 (the midpoint of the 2020-2024 extended settlement period) was 24.7%. Based only on inflation, the unadjusted rate in June 2022 would have been 32.7¢ per thousand plays. But the royalty amount agreed to for that period was 22.9¢ per thousand plays, only about 70% of the inflation-adjusted nominal rate. This implies a repertoire-use adjustment of 70% (the “implied repertoire-use adjustment”).

[33] Pandora claims that this implied 70% repertoire-use adjustment is in line with the Re:Sound repertoire-use adjustment of 63.5% for 2015 and 2016 from the recent pay audio services decision.[14] Pandora notes that the 63.5% figure followed the partial completion of the audit phase of a repertoire study.[15] According to Pandora, it is the only repertoire-use adjustment supported by an audit of Re:Sound’s claims regarding repertoire eligible under the 2014 framework.

[34] Regarding the RS8 study, Pandora argues that its results should be dismissed because the full audit of authorizations to represent a sample of U.S. sound recordings, initially ordered by the Board, was not completed.

Considerations

[35] To determine the extent of the repertoire-use adjustment, two methods are before us: a tariff-specific repertoire-use study, the RS8 study, and a proxy, the RCS Agreements.[16]

[36] We observe that the RS8 study, while not perfect, relies on a robust design. We further observe that the RCS Agreements were executed after the RS8 study was completed and, therefore, were likely informed by it. In other words, the RCS Agreements could account for the RS8 study imperfections, and resulted in a reduced repertoire adjustment.

The RS8 Study

[37] We first analyze the RS8 study. This method for determining the repertoire-use adjustment has both strengths and weaknesses.

[38] In terms of strengths, we note several highlights. First, the RS8 study relates directly to the sound recordings that are streamed by users of this tariff (non-interactive and semi-interactive webcasters). By contrast, other repertoire studies (such as the one in Pay Audio and the other in Commercial Radio[17]) relate to different markets (with potentially different content offerings).

[39] Second, the sampling size of the RS8 study, at over 1.4 billion plays, is the largest compared to the repertoire studies in Pay Audio and Commercial Radio. The sampling size of the Commercial Radio study is 1.5 million plays. While the sampling size for the Pay Audio study is not available, it is probably about 10% larger than that of the Commercial Radio study,[18] so approximately 1.65 million plays.

[40] Third, the RS8 study was conducted in 2021, so it is the most recent, compared to the Pay Audio study (2016) and the Commercial Radio study (2008).

[41] In terms of weaknesses, the main issue raised by Pandora is that it was not able to fully audit the extent to which Re:Sound is actually authorized by rights holders to manage the sound recordings it claims to represent.

[42] We however note that Re:Sound provided documents supporting a partial audit. Re:Sound filed 208 documents to support its capacity to represent and manage U.S. sound recordings in general, as well as examples of sound recordings specifically listed in the RS8 study which are lawfully represented by Re:Sound.[19]

[43] As far as major record labels are concerned (which represent a large share of the market), we note that Re:Sound is authorized to represent their repertoires via a general mandate. The specific sound recordings within those repertoires are identified through ownership claims made at the time of royalties distributions. [20]

[44] While this system of good faith claims of ownership at the time of royalties’ distribution may be a standard practice, ownership title errors or deficiencies are a possibility. For example, in SODRAC v CBC,[21] the Board found that SODRAC’s claim for repertoire representation was inflated by 12 percentage points (para 88). In Pay Audio, the audit of the repertoire study was not completed. However, Re:Sound accepted a repertoire-use adjustment of 8.5 percentage points lower than its initial claim.[22]

[45] In any event, Pandora did not respond or comment on these aspects of the RS8 study, although suggesting that it did not review these documents to complete a partial audit, as we did.

[46] This underscores a proportionality issue in light of Re:Sound Tariff 8’s relatively modest yearly royalties (see Exhibit Re:Sound-2 at para 33). If a full audit entails gathering the documentation for a sample of 300 sound recordings, this would generate more than 2,300 documents, whereas the RCS Agreements may act as substitute without the costs of a full audit.

[47] Moreover, imperfect or partially audited repertoire-use studies are precedented. The RS8 study can therefore serve as a background for the RCS Agreements. Conversely, the RCS Agreements can be used to account for the imperfections of the RS8 study.

The RCS Agreements

[48] This method is criticized by Pandora, who argues that the RCS Agreements are not representative of the market in terms of users and have no precedential value.

[49] Regarding representativeness, Re:Sound provides a list of 38 users who paid under Tariff 8 from 2012 to 2020.[23] For the purpose of the Last-Approved Tariff, the Board assumed homogeneity in terms of webcasters’ content offerings.[24] We apply the same assumption in the present matter. There is no reason to believe that the services who signed the RCS Agreements and the other users listed offer significantly dissimilar content.

[50] Pandora also notes that some of the larger market participants, like |||||||||||||||||||||||||||||||||||||||| ||||||||||||||||, are missing from the list. However, these services only offer interactive services. As such they would not be dealing with Re:Sound, who we believe does not administer the rights applicable to interactive services. Their absence is therefore not a problem for representativeness.

[51] In any event, representativeness is not determinative of whether the RCS Agreements are meaningful proxies for a repertoire-use adjustment.[25] We note that— ||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| —they set a rate binding the signatories. As such, they are legally meaningful.

An informed bargain?

[52] Agreements on the scope of a repertoire without the benefit of a repertoire-use study are common and frequently taken into account by the Board.[26] However, a bargain informed by data from repertoire-use studies becomes an even more meaningful proxy.

[53] In this case, the bargain appears to have been rationally informed by the RS8 study and may have been informed by the Pay Audio study. Together, they constituted an interval within which to negotiate.

[54] First, the RCS Agreements (dated November 21, 2021) were negotiated with the benefit of the RS8 study. For reference, CAB and Stingray were consulted on the methodology used and the sample drawn by Circum Networks was provided to them.[27]

[55] Furthermore, the RCS Agreements explicitly provide that the “|||||||////////////////////////////||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||//////////////|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||.”[28] This indicates that the final results of the RS8 study were considered. (Specifically, the study found that 92.6% (semi-interactive) and 95.1% (non-interactive) of streamed sound recordings were represented by Re:Sound)

[56] Second, the Pay Audio repertoire-use study likely informed the bargain leading to the RCS agreements. We note that Stingray, one of the parties to the RCS agreements, participated in the Pay Audio repertoire-use study. While we do not know for a fact that CAB knew about this study, since the CAB-Re:Sound agreement is essentially the same as the Stingray-Re:Sound agreement, we can reasonably assume that the Pay Audio study likely informed both set of negotiations.

[57] Furthermore, the Pay Audio study was duly referenced in the Board’s May 28, 2021, Pay Audio decision. In that proceeding, the parties settled on a repertoire-use adjustment after a repertoire study was completed but without a full, formal audit of the eligibility and repertoire status of a sample of sound recordings. The Pay Audio settlement consisted of two figures: 45% for the period up to August 12, 2014; and 63.5% for the period from August 13, 2014, onwards.[29]

[58] With respect to the Pay Audio settlement, Re:Sound stated that: “The parties have agreed that the joint proposal outlined above completes the repertoire study including the audit stage and that no further documentary evidence, submissions or formal hearing process is required.”[30]

[59] While both the Pay Audio and RS8 repertoire-use studies were incomplete, they provide relevant information. The RCS Agreements benefited from this technical, relevant information, which consequently informed the bargain.

[60] In any event, we observe that the repertoire adjustment from the RCS Agreements is within the interval set by the Pay Audio and RS8 studies. As a result, it constitutes a reasonable and relevant figure.

What is the correct implicit repertoire-use adjustment?

[61] Based on the RCS agreements, Parties proposed different ways to calculate the implicit repertoire-use adjustments (RUA), meaning the percentage of the agreed rate that in turn reflects the signatories’ agreed adjustment for repertoire use.

[62] To compare the Parties’ proposals, we present them as formulas:

- Re:Sound: 2018 agreed rate (20.2¢ per thousand plays) ÷ Nominal rate (26.2¢ per thousand plays) = 77% RUA

* Does not assume that the agreed rate accounts for inflation

- Pandora: 2022 agreed rate (22.9¢ per thousand plays) ÷ [Nominal rate (26.2¢ per thousand plays) x inflation rate]* = 70% RUA

* Assumes that the agreed rate accounts for inflation using the “mid-point of the settlement agreement period from January 2013-June 2022, namely 24.7%

[63] In essence, Pandora believes that the 77% repertoire adjustment put forward by Re:Sound is too high because it does not account for inflation, which Re:Sound, the CAB, and Stingray, as signatories, would have accounted for in their agreements. Therefore, to find the “implied repertoire adjustment,” we must remove inflation from the agreed-upon rates.

[64] We tend to agree with Pandora’s approach. Since the RCS Agreements are considered reliable proxies, they must be considered in their entirety. The Board assumes that all relevant factors would have been accounted for in any agreement, including inflation.[31] Inflation should therefore be assessed within the agreements’ timeframe, independently from any description of the agreement by the parties, where the context requires it.[32]

[65] However, we do not agree with Pandora’s inflation rate. Pandora’s approach ignores the fact that the settlement agreements were signed on November 23, 2021. At this point in time, only the inflation rate from 2013 up to the month prior to the date of signature (October 2021) could be determined, which was 18.63%.

[66] An 18.63% increase to the unadjusted royalty rate equates to $0.000311 per play, which is 73.7% higher than $0.000229. Our preferred formula is as follows:

- 2022 agreed rate (22.9¢ per thousand plays) ÷ [Nominal rate (26.2¢ per thousand plays) x 18.63% inflation rate]* = 73.7% RUA

* Accounts for inflation known at the time of signature (from January 2013-October 2021) (signature Nov. 2021), namely 18.63%

[67] We therefore approve a repertoire-use adjustment of 73.7% by accounting for the known inflation rate of 18.63%. Re:Sound’s approach ignores the past and Pandora’s assumes that the parties could have perfectly predicted the future. Our approach overcomes these shortcomings.

Repertoire-Use Adjustment for CBC

[68] Finally, we explain why the 73.7% repertoire-use adjustment is also applicable to CBC.

[69] First, as the Board noted in its reasons for the Last-Approved Tariff, CBC was expected to play much more Canadian content by its mandate, most of which was in Re:Sound’s repertoire, compared to commercial webcasters.[33] As Re:Sound explains, with the expansion of Re:Sound’s repertoire and U.S. repertoire now being eligible, a separate repertoire adjustment for CBC is no longer necessary.

[70] Second, as the Board also noted in its reasons for the Last-Approved Tariff, there is no reason to believe that CBC’s webcasting content is markedly different from its commercial counterparts.[34] Similarly, there is no evidence in this proceeding that would support an inference that CBC and commercial webcasters use sound recordings not represented by Re:Sound differently.

[71] In October 2019, CBC discontinued its participation in the proceeding and stated on the record that it was no longer opposing Re:Sound’s proposed tariff up to and including 2018. CBC may, in future proceedings, present evidence that its content warrants a different repertoire-use adjustment, even if it chose not to in this proceeding.

[72] We use the RCS Agreements as proxies for a repertoire-use adjustment of 73.7%, including for CBC rates.

C. Issue #3: Should partial plays and free trials be accounted for?

[73] Pandora claims that partial plays (i.e., streams of 30 seconds or less) and free trials should be excluded when calculating the royalties owing. In other words, plays heard during free trials and partial plays should not trigger royalties.

[74] We conclude that partial plays should trigger royalties but free trials should not.

i. Partial Plays (or “Skips”)

[75] Pandora and SiriusXM argue that partial plays should no longer attract royalties because of new evidence regarding changes in industry standards. [35]

[76] Pandora also suggests that partial plays are insubstantial, non-compensable acts. Moreover, Pandora argues that “an approach that requires payment for Partial Plays as some kind of interactivity premium is inconsistent with technological neutrality. Skipping is made possible because services buffer the next song “early,” and in parallel to the song actually being played, in order for it to be available if the song being played is skipped. Accordingly it is a form of reproduction (buffering), not performance, and the intervention of technology (at the service’s expense), not the work, that drives this interactive feature.”[36]

[77] Pandora suggests that a partial play exclusion should be implemented by either not requiring payments on partial plays (plays that last 30 seconds or less), or by a blanket discount to the per-play rate.

[78] If a blanket discount is used, Pandora suggests that the adjustment should be the average of three available data points: ||||||||||||| of Pandora’s plays being partial plays on its ad-supported semi-interactive tier, ||||||||||||| on Pandora’s paid semi-interactive tier, and ||||||||||||| based on a third-party study of Spotify user behaviour. This would result in a |||||||||||||discount.

[79] Re:Sound submits that partial plays must be subject to full royalties. It argues that skipping and its associated interactivity provides real (and added) value to both consumers and providers of online music. Re:Sound disagrees that excluding partial plays is an “industry practice”: only a handful of SOCAN’s licence agreements exclude partial plays. Furthermore, it notes that under the U.S. copyright framework, partial plays attract royalties in general and exceptions are very narrow. Comparatively, the Canadian per-play rate is low and would be even lower if partial plays were excluded.

[80] According to Re:Sound, excluding skips from the royalty base would cause a data problem. First, the data on skips is provided by Pandora through its fact-witness evidence filed in the OMS proceeding. Re:Sound therefore was not able to question the witness about the reliability of this data. Moreover, the data relates to a different geographical market. Second, the additional figures on skipping come from a third-party study relying on unverified data sources and relate to Spotify, which offers only fully interactive services. The latter are usually subject to a different tariff and rate structure.

[81] Finally, Re:Sound disagrees that skips relate to the reproduction right (unrelated to this tariff) through buffering. Skips can occur without buffering. Additionally, according to Re:Sound, it is not legally accurate to associate partial plays with fair dealing for the purpose of research. Society of Composers, Authors, and Music Publishers of Canada v. Bell Canada, 2022 SCC 36 (“Bell”), which related to 30-second previews before purchasing a music file, does not apply in this case.

Considerations

[82] Whether or not 30-second partial plays may trigger legal liability is irrelevant. We do not attribute a value to skips. We simply count skips as a means to value interactivity.

[83] Indeed, a service with interactivity has more value than a service without interactivity, that added value being the “interactivity premium”. In this regard, the data filed by Pandora suggest that skipping differs depending on the market: ||||||||||||| of Pandora’s plays are partial plays on its ad supported semi-interactive tier, ||||||||||||| on Pandora’s paid semi-interactive tier,[37] and ||||||||||||| on Spotify—a fully interactive service.[38]

[84] If we convert these percentage rates into a per-play rate by using data on the number of streams and the amount of royalties paid by each service from the OMS proceeding,[39] we find that the fully interactive per-play rates are, on average, |||||||||| times higher than the rates for semi-interactive services that we would approve. This suggests that increased interactivity commands a higher price.

[85] The fact that some industry agreements do not count partial plays for the purpose of royalties is not determinative on how interactivity is valued. Interactivity may nonetheless be reflected in the agreed rate, the rate structure, or specific financial arrangements.

[86] In a perfect situation, we would have set two different rates: one for non-interactive services and another higher one for semi-interactive services, the difference expressing the interactivity premium. However, we have no information on the precise value of interactivity to consumers, nor on the marginal value of skips from the consumers’ viewpoint.

[87] We therefore follow the same approach as in the Last-Approved Tariff: requiring payment for partial plays in a per-play tariff that does not distinguish between non-interactive and semi-interactive streams is a practical way to give value to the interactivity in the latter.

[88] Furthermore, we do not accept the argument that partial plays are exclusively tied to the reproduction right. This argument ignores the fact that a communication has to occur in order for the end user to decide to skip or not. Even if the reproduction right were engaged by partial plays, we do not have evidence that would permit us to allocate the interactivity premium between the communication and reproduction rights.

[89] Because there are no agreements covering the reproduction rights for streaming on the record, we lack information on whether such an interactivity premium is already being paid in respect of these, and the amount of such a premium.

[90] In conclusion, we recognize that an interactivity premium is appropriate. Since we do not have information that would permit us to price that premium more accurately, we continue to use partial plays’ count as an indirect measure of interactivity. As a result, partial plays are part of the royalty base.

ii. Free Trials

[91] Pandora and SiriusXM propose that any plays that occur during free trial periods of 31 days or less per year should not be subject to royalties.

[92] The rationale is that free trials increase the number of subscribers and ultimately increase total plays and royalty payments to rights holders. Pandora cites in support the economic expert evidence it filed in the OMS proceeding[40] and the fact that free trials are common industry practice.[41]

[93] From Re:Sound’s perspective, this is purely speculation, relies on untested evidence, and cannot be a basis to fundamentally alter the economic terms of a tariff. It is just as possible—since many free trials appear designed for one service to compete against other services for listeners—that free trials draw users from one service to another with no net benefit to rights holders. According to Re:Sound, rights holders do not benefit from promotional measures designed by online music services to attract subscribers, nor should they be expected to subsidize those efforts by allowing their music to be given away for free.

Considerations

[94] While free trials may feel counterintuitive from a legal viewpoint, they are accepted in the market, which suggests that they are mutually beneficial to both licensor and licensee. Of the |||||| agreements SOCAN filed in the OMS proceeding (and discussed by the parties in this proceeding), free trial provisions appear in |||||||||| ||||||||||||.[42]

[95] Also, free trials are limited in time and, as such, they cannot be a substitute for a paid subscription model. We can reasonably assume that the impact on the rights holder’s market is therefore negligible compared to the potential benefits.

[96] This was implicitly recognized in SOCAN - Tariff 22.D.1 (Internet - Online Audiovisual Services) (2007-2013) [Redetermination],[43] where the Board approved a tariff based on a jointly-submitted text involving Netflix. The tariff provided that any singular, free-trial month within a 12-month subscription is royalty-free.

[97] Accordingly, plays during any 31-day free trial within a 12-month period shall not be counted for the purpose of royalties.

D. Issue #4: What should the adjustment for inflation be for the year 2018?

[98] Both Re:Sound and Pandora accept that an adjustment to the royalty rate for inflation is appropriate. Re:Sound asks for a single inflationary increase of 10% applicable to the royalty rates for 2018 (following Board practice, inflation for 2013-2017 would be given in 2018). The Board agrees that an inflationary adjustment is appropriate because it preserves the purchasing power of rights owners.

[99] The Board generally measures inflation as the percentage change in CPI (consumer price index) between January of the first year and December of the last full year of data available.

[100] Re:Sound asks for an inflationary increase of 10% starting in 2018, using the monthly CPI growth from January 2013 to December 2018. The CPI in January 2013 was 121.3 and the CPI in December 2018 was 133.4. This results in a 9.98% inflation rate, which is almost the same as Re:Sound’s calculation of 10%. Of note, generally, the Board does not round inflationary increases to the nearest whole number.[44]

[101] The issue at hand is whether to apply the inflation increase starting in January 2018, given that all of 2018 (January-December 2018) is used in the year of calculation. The Board has frequently said that it will not forecast inflation; however, in 2023, inflation for the period of 2018 is known.

[102] Nonetheless, allowing all of 2018 to be included in the adjustment starting January 2018 would not be fair or equitable. In such a case, users paying royalties for the period of January 2018 would be paying a higher rate based on an inflationary increase that had not yet occurred. For these reasons, an inflation adjustment using the period between January 2013-December 2017 is more appropriate.

[103] According to Statistics Canada, the CPI in January 2013 was 121.3 and the CPI in December 2017 was 130.8. This results in an inflationary increase of 7.83%.

[104] Accordingly, we approve an inflation adjustment of 7.83% to the royalty rates for the year 2018.

E. Issue #5: Should proposed tariff wording changes be accepted?

[105] As explained by Re:Sound, the JST contains revisions to the Last-Approved Tariff wording and new administrative provisions “for the purposes of greater clarity and consistency with more recently approved tariffs.” The main proposed revisions follow. Neither Pandora nor SiriusXM oppose these wording modifications.

Scope

[106] Re:Sound explains that for greater clarity, section 3(1)(e) has been added to confirm that the tariff does not apply to background music suppliers, which are subject to Re:Sound Tariff 3.A – Background Music Suppliers (2014-2018).

[107] We approve this change as it adds clarity to the tariff’s scope of application and does not constitute a circular reference.

Definitions

[108] Re:Sound explains the changes in definitions as follows: The terminology in the JST has been revised to replace the term “webcast” with “stream.” The purpose of this change is to reflect the more current terminology used in the industry and for greater consistency with SOCAN’s and CSI’s Online Music Services tariffs. For the same reason, the term “webcaster” is replaced with “service.”

[109] Furthermore, the definitions of “non-interactive stream,” “on-demand stream” and “semi-interactive stream” have been revised to provide greater clarity, including using examples. The definition of “file” has been revised for greater clarity, such as to include all plays, whether they are in Re:Sound’s repertoire. Indeed, since the rate is adjusted upstream on account of repertoire, there is no need to determine downstream whether a sound recording is part of the repertoire. The definition of “simulcast” has been revised to reflect the Board’s recent rulings clarifying that “whether or not an option for interactivity is actually used or not does not change the character of that particular service.”[45]

[110] Finally, the definition of “aggregate tuning hours” has been removed as it was part of the transitional provisions under the Last-Approved Tariff and is no longer applicable.

[111] We approve the changes for the reasons provided by Re:Sound.

Timing of Minimum Fee Payment

[112] Re:Sound explains that section 5(3) is added to provide for the payment of the $100 annual minimum fee by January 15 of each year, to be credited against the monthly amounts payable. Under the Last-Approved Tariff, there was ambiguity as to when the minimum fee was payable.

[113] We approve the changes for the reasons provided by Re:Sound.

Adjustments to the Amount of Royalties Owed

[114] Section 11(2) of the Last-Approved Tariff, [46] which addressed erroneous information reported by a user regarding music files streamed, has been removed.

[115] Re:Sound explains that it is not applicable given that (i) Re:Sound’s per-play rate is adjusted for repertoire (which means that any file streamed need not be identified in terms of whether it is eligible, in the public domain or in Re:Sound’s repertoire), and (ii) it imposes an overly onerous and complicated obligation with respect to how Re:Sound’s royalty distributions are conducted.

[116] We approve the changes for the reasons provided by Re:Sound.

Audits, Confidentiality, Delivery of Notices and Payments

[117] Re:Sound is proposing the following changes. Sections 10, 11 and 14 have been revised for consistency with more recently approved tariffs. Section 11(2)(b) has been expanded to allow for the sharing of information with CMRRA, SODRAC and CSI for the same reasons the Board already allows information to be shared with SOCAN. The sharing of information among collectives creates administrative efficiencies and can reduce the need for unnecessary, duplicative audits of services.

[118] We approve the proposed changes for the reasons provided by Re:Sound.

Transitional Provision – Interest Factors

[119] The Board recognizes the time value of money and, as such, typically provides for interest factors. We include interest factors that compensate for opportunity costs incurred during the 2013-2018 period. Opportunity costs arise because of the difference in royalties that may not have been paid out to Re:Sound during that period.

V. conclusion

[120] We approve the Proposed Tariffs as per the JST with the following adjustments:

• -Accounting for inflation in the agreed-upon rates to identify the implicit repertoire adjustment, and adjusting the rate accordingly;

 

• -Increasing the 2018 rate by 7.83% to account for inflation between January 2013 and December 2017;

 

• -Excluding free trials up to 31 days per year;

 

• -Including partial plays (up to 30 seconds); and

 

• -Making minor wording modifications as per the JST.