Who should pay for a 3 month trial?

imgresTaylor Swift took to her blog today to, very politely, explain her decision to keep her latest album out of Apple Music.

I’m sure you are aware that Apple Music will be offering a free 3 month trial to anyone who signs up for the service. I’m not sure you know that Apple Music will not be paying writers, producers, or artists for those three months. I find it to be shocking, disappointing, and completely unlike this historically progressive and generous company.

This is not about me. Thankfully I am on my fifth album and can support myself, my band, crew, and entire management team by playing live shows. This is about the new artist or band that has just released their first single and will not be paid for its success. This is about the young songwriter who just got his or her first cut and thought that the royalties from that would get them out of debt. This is about the producer who works tirelessly to innovate and create, just like the innovators and creators at Apple are pioneering in their field…but will not get paid for a quarter of a year’s worth of plays on his or her songs.

I have to say that, as a parent, I really do like how polite she is. The issue though is whether she is right.

What is happening here is that Apple Music will cost $10 per month which is something Swift is in favour of. But it will have a 3 month free trial for consumers. That isn’t the problem: the problem is that Apple won’t be paying artists or anyone else while it isn’t charging consumers. In other words, it is a freemium service but one based on a free trial.

Having a free trial is not an unreasonable thing to happen. In fact, it has a solid rationale that should be good for everyone. If the trial gets people hooked to Apple Music, they will become subscribers and that will lead to steady income for everyone involved; indeed, it is better for the group other than Apple since Apple will be returning 70 odd percent of the subscriber fees to them.

Here is what Swift is asking for as an alternative: that Apple foot the entire bill for the three month trial or, alternatively, not offer the three month trial. Let’s think about these using a model since I suspect just arguing the case without some notation is going to go nowhere.

Let’s assume the following:

1. That x percent of users who have a three month trial convert to a full subscriptions. Thus, the average subscription after three months is $10x.

2. The share that artists/publishers/publicists receive of that is a. So they earn $10xa per month for however many months the subscription lasts. Let’s call that m on average. (I’ll ignore discounting in this equation).

3. 100% of all users take the three month trial.

4. If there isn’t a three month trial, then y percent of users sign up for full subscriptions. While one can imagine that for certain songs, it could be that y > x, let’s assume here that the trial is demand-creating so that x > y.

Given all that we can calculate whether the three month trial is on aggregate a good idea. We just need to compare the three month trial revenue $10xam to the non-trial revenue of $10ya(m+3). The three month trial is, on net, a good thing under Apple’s current terms if xm > y(m+3) or (x-y)m > 3y; that is, the increased revenue over the subscription life exceeds the lost revenue for the three month trial. Importantly, this is independent of a and so there is a sense that Apple and everyone else are in this together.

What Swift is proposing — that Apple shoulders the cost of the three month trial — changes the “in this together” aspect. In this situation, the artists etc will receive $30a + $10xam if there is a three month trial while Apple will receive $10xam – $30a. So now Apple will only want a three month trial if $10x(1-a)m – $30a > $10y(1-a)(m+3) or (x-y)m > 3(a/(1-a)+y). By contrast, the three month trial is a better deal for artists etc because not only do they receive $30xa (which is their so-called ‘lost’ revenue) but they receive an additional $30(1-x)a for all the people who tried and did not buy. In other words, artists are being asked to be rewarded when their music isn’t liked!

To summarise then: if the three month trial is commercially a good idea for Apple, then it is surely a good idea for artists etc. To propose an alternative sharing of the costs of the trial other than the one Apple has proposed breaks that commonality of interest. You need something other than ‘fairness’ or equity to argue the case for that.

So let’s delve into this a little further. Not all music is the same. Some is more popular than others. Some is more time-sensitive than others — so will only be popular for a little while before it becomes unpopular. How do things like this change the equation?

Let’s start with popularity. One reasonable hypothesis is that popular artists will bring more users to Apple Music’s free trial and will also be anchors that expose those users to less popular artists. Since streaming’s value proposition is that it has wide variety (compared to buying your own music) then popular artists are more valuable to unpopular ones than the other way around. What this means is that unpopular artists stand to lose more if the free trial doesn’t include popular ones. That, however, is not what she said. Swift was arguing she was pulling her currently most popular album to help the less popular. In reality, however, if the desire is to help the less popular artists get discovered, then it is better to have the popular ones on the three trial than not and, if anything, the popular ones would be entitled to some compensation for their larger possible losses as a result of this. Thus, there is a case that popular artists should receive a better deal, overall, than less popular ones and especially so with regard to the three month trial. For all the dressing up of the argument in terms helping the weak, they would be helped more if Swift’s album was available.

What of time sensitive music. Here we need to distinguish between the launch of Apple Music and free trials after that point. At the launch, because many millions will try it, those whose work is currently popular but may not remain so, stand to lose more than others who benefit from the paid subscriptions later on. For Swift, this would apply to her currently popular latest release. She actually has an interest in people not signing up to Apple Music during that launch phase because it will cannabilise her sales during the critical period before her fans shake it off. By contrast, the launch is a boon to the less popular artists. Indeed, with Swift not there, there is a chance that more users will discover them during that free trial. Note there is a nuance here to my earlier argument. Popular artists can bring more users to a platform but once they are there they compete with unpopular ones for attention. On net, it is not quite clear how this plays out. There is a chance that Swift pulling her album is actually helping less popular artists out. This will surely make for a good empirical study at some point.

One way Apple could insure against some distortions would be to do the following. Collect information on the music listened to by individual. Then if a consumer converts to a paying customer, compare their music with those who don’t. The assumption is that the music played during the trial played a role in causing the consumer to become a paid subscriber and so the artists etc responsible for that deserve some additional compensation if (and this is a big IF) the mix of music played during that trial differs from what those subscribers listen to later on. In this way, if Swift’s album is played a lot during free trials and those consumers do become subscribers then Swift should receive a bump in compensation thereafter.

Apple like simple pricing rules with, for instance, a one size fits all revenue sharing formula. But music streaming is a bundled service and not all components of the bundle are equal. The data is now available to work out the right compensation formula. However, I suspect that the right formula is one that pays more to artists like Swift and less to the artists she is claiming to defend.

8 Replies to “Who should pay for a 3 month trial?”

  1. If Apple wants to be “in this together” with the artists, they need to actually partner, which would imply being consultative with the artists on the appropriate length of the trial and sharing revenue with products that are part of Apple’s music bundle (e.g. iPhones). Since Apple isn’t going to do that, I don’t think they should be relieved from licensing costs for the trial period any more than they are of bandwidth costs.

  2. This makes sense for Apple, because we are talking about their music distribution system. Taylor Swift has several music distribution systems she can use to get to her fans. It seems fairly obvious to me that artists should not be expected to foot the bill when some music distribution company wants to get into the business or get a larger share of the pie. But that is all ignored in this analysis, as if Apple had just discovered a magical new revenue stream with a completely innovative service.

  3. “If the three month trial is commercially a good idea for Apple, then it is surely a good idea for artists.”
    Not so quick, Joshua. The free trial is Apple’s strategy to break into the market currently dominated by Spotify, the candy to lure customers away from Spotify. If the new restaurant offers a 2-for-1 deal to attract customers it still has to pay its suppliers, kitchen staff and waiters for their services. Why should the artists foot the bill for Apple’s market entry strategy (unless they have explicitly and voluntarily agreed to share the cost of a joint promotion offer)?

  4. Agree with Sven – Apple is not a monopolist. Assume that Apple is correct that they are better for artists than Spotify and others in the long run. Assume artists get a>b in lifetime revenue from Apple and Spotify every three months, respectively. Let v>w be the number of spotify users who will switch to Apple and eventually pay under free trials and no free trials, respectively, and s and t be the number of people who don’t subscribe to anything now but who will sign up and pay for Apple under free trials and no free trials.

    With a free trial, artists earn a difference from the status quo of (a-b)vd+asd-xvb(1-d), where d scales for not any revenue over next three months (given 2 year average subscription plus discounting, maybe d=5/6 is reasonable). The first term is increased revenue from the Spotify switchers, the second eventual revenue from new signups due to Apple and the third term lost short term Spotify revenue.

    With no trial, they earn a difference from status quo of (a-b)y+at. Apple only cares about v, s, y, t, and not the third term in the free trial artist equation, hence they will prefer longer trials (for business stealing reasons) than the artists would. This is easy to fix: Apple can offer a free trial, and pay artists xvb(1-d); that is, Artists can be paid what they would have been paid from Spotify for all switchers, which aligns incentives again.

    If I were Taylor Swift, though, I would operate under the assumption that the musicians have more bargaining power right now, but not later, and try to extract more than xvb(1-d); Apple is still better off with the free trial so long as the rents they pay artists during the free trial are less than the difference in profit they earn under free trial and no trial. That is, the artists should charge an “entry fee” to Apple exactly equal to the full expected string of Apple profits from their new music service. And indeed, it seems that Apple is offering a premium to artists over their lost Spotify revenue, in recognition of their current bargaining position. Who knew Tay was such an astute bargainer?

  5. It seems pretty clear from the comments that the original post gets this issue wrong. I enjoy reading the material on this site and I often learn from it, but I do wish the posters would occasionally respond to comments. In this case, I hope Joshua Gans acknowledges the mistake (hey, we all make them, that’s what blogs are for).

  6. Your article assumes that the Apple service enlarges the pool of listeners, such that all the revenue generated does not come at the expense of other paid music services. Furthermore, if Apple wanted to give away others’ work for free, simple manners would have required that they ask for permission before doing so. High-handedly offering other’s work for free the way Apple did would offend anyone, and I applaud Taylor Swift for calling them on it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s