A look at existing and emerging internet-age business models

Ben Robinson
Jun 7 · 9 min read
Are US subscription-based business models an innovation, or a compromise? | artwork by @morysetta (IG)

Ben Robinson, Practice Head of Strategy and Marketing at Pangea and co-founder of a p e r t u r e | Hub

Have you ever wondered why the social features on Spotify aren’t better? It’s not difficult to imagine a more engaging experience if, say, it was easy to co-create playlists or you got updates on what friends and artists were listening to. One explanation could be that it is hard to build siloed social interaction — outside of horizontal apps like Facebook or WhatsApp. But my hunch is that, while Spotify wants to maximize users, it doesn’t want to maximize usage. Why? Because unlike most aggregation platforms, its marginal cost grows with usage, meaning that — like MoviePass — its best customers are actually its worst customers. That’s why it’s struggling to generate profits, why it’s struggling to protect itself from rival streaming services and why it needs to become the pre-eminent platform for podcasts.

The move to subscriptions

There is a structural move underway towards subscription-based business models. According to Zuora’s Subscription Economy Index, subscription revenues are growing five times faster than regular transaction-based revenues. Part of this reflects changing consumer behavior towards renting, rather than buying, assets. But the other explanation is that with internet distribution, digital services have become abundant. As such, it makes less sense to restrict supply to maximize profits, as companies did in the past. Instead, companies are working with new business models that take as their starting point low or even zero marginal costs of supply and seek to generate revenues and profits differently. Some seek to make the service available for free and charge advertisers (like Facebook), others seek to drive down the cost of sales to maximize output (like Uber) while others seek to aggregate abundant supply in a way that brings convenience to consumers, for which they charge a subscription price in return.

Source: Zuora

Is a subscription model right for Spotify?

Prima facie, Spotify and, to a lesser extent, MoviePass look like good candidates for subscription pricing. Abundant, there is negligible cost and no deterioration in the quality of either music or movies with additional consumption (although for movie theaters, there is finite capacity, of course — not to mention that MoviePass doesn’t access the most profitable aspect of moviegoing, which is concessions). But the common business model challenge with Spotify and MoviePass is that marginal costs increase as consumption increases. In Spotify’s case, it pays the labels and the artists every time a song is played. For MoviePass, it pays the cinema every time a customer watches a movie. As such, they are offering for a fixed cost to consumers what in their case are uncapped marginal costs. And so, their business models become a gamble that 1) most consumers will stop consuming before marginal costs exceed average monthly revenue per customer 2) they can drive down operational costs to compensate for low gross margins 3) they can open up additional revenues streams over time, like selling data to artists or movie makers or 4) they can develop superior experience that drives down customer acquisition cost (CAC) and churn over time. But it is easy to see how this is precarious — and hence why Spotify isn’t profitable — and also how elements of the model are in conflict, such as developing superior experience without pushing up consumption.

Spotify vs Netflix

For Netflix, the model works very differently. Netflix buys the rights to its content, so that it turns variable costs to fixed costs that it spreads over higher revenues, meaning high gross margins. Moreover, the content producers are in general fragmented — Disney , for instance, has its own route to market through its own streaming service — such that their power is small and diminishing vis-a-vis Netflix, allowing it to extract continuously better terms. Contrast this with the music industry where three labels control practically all of the rights and thus have strong bargaining power to negotiate advantageous terms.

Another major difference between Spotify and Netflix is that Netflix is able to constantly improve its offering by investing in original content. Where people say that spending billions on original content creates the pressure for Netflix to keep growing, this is somewhat true — they want to recoup the costs. However, it doesn’t oblige them to up the efforts to acquire new subscribers, but instead allows them slowly but steadily to increase subscription prices over time without seeing an increase in churn, meaning gross margins continue to rise, more than offsetting the increase in operating costs. The issue with Spotify is that, for now at least, it is offering a largely undifferentiated service. Sure, it has better playlists than Apple’s or Amazon’s streaming services, but all streaming services allow you to access all music. This not only limits the scope to take up pricing (in contrast to Netflix, this would provide significant churn), but it also pushes up CAC since consumers have many direct substitutes (witness for example how Spotify is giving away Google speakers to induce new subscribers), further pushing down gross margins.

Spotify as Netflix

It is against this backdrop that Spotify’s move into podcasting is so interesting. In contrast to music, the supply of podcasting is highly fragmented — there are 700,000+ free podcasts available today, including our own :) — meaning that the aggregation dynamics are completely different. In addition, the audience is similarly fragmented making it difficult to monetize podcasts through ads — today, podcasts in the US generate around $300m in advertising or $0.01 per active user per hour, 10x less than radio and 70x less than newspapers.

Spotify has made three acquisitions in this space — Gimlet Media and Parcast, which give it a catalog of high-quality podcasts and the team to be able to commission and produce more of the same and Anchor, which is a platform for amateurs to make and distribute podcasts easily. So, in short, it is building original content à la Netflix as well as providing a route to audience for the long tail of producers, on top of which is can seek to increase advertising revenues or take up subscriptions. With podcasts, therefore, Spotify is a position to become a true aggregator, gaining control over supply such that gross margins increase and able to leverage its position to deliver increasing returns to scale that would see a rise in average per user revenues at the same time as reducing churn and lowering CAC.

Spotify vs Amazon and Uber

One of the things you realize when you read this fascinating oral history of Amazon Prime is that, at the time, no one knew what it would cost to launch and what kind of a black hole in the company accounts it could create with massive adoption. Instead, you get a sense of Jeff Bezos’ drive and leadership, the company’s focus on constantly improving customer experience as well as how innovation materializes in real life.

What is interesting about Amazon Prime in the context of Spotify is that it is analogous to the Spotify model: a fixed price subscription for a service where marginal costs are uncapped. In the same way, Amazon’s best customers could have become its worst customers, driving down profitability. However, not only did Jeff Bezos see this the other way around, as a means to encourage its best consumers to consume more (“ We’re not going to take our best customers for granted”), but because of all its different business lines, Amazon was able to capture that extra consumption across its estate and turn it into revenue. This has continued to be the Amazon philosophy (When we win a Golden Globe, it helps us sell more shoes”), but it’s not a strategic play for Spotify since, for now at least, it’s a single service offering.

The problem with the Amazon model — and the Uber model — is that it leaves a lot of money on the table. In the case of Uber and Amazon, both have low distribution costs but high other costs (drivers, fulfillment centers, logistics, etc) so the incentive is not to maximize profits by cutting supply, but instead by maximizing supply to maximize revenues (and network effects). This effectively pushes out the supply curve, which creates significant additional consumer surplus (a difference between the price paid and the price customers would have been willing to pay). A study by Freaknomics author Stephen J. Dubner estimates that, in a single year in the United States, Uber produced USD7bn of consumer surplus, with consumers paying USD4bn when they would otherwise have been prepared to pay USD11bn. This is money that would historically have gone to workers and to profits.

Spotify vs Tencent Music

There do, however, seem to be emerging business models — coming out of Chinathat conform to the platform dynamics of low distribution costs and high supply, but which manage to capture much of the consumer surplus generated. A look at Spotify compared to Tencent Music helps to illustrate this.

Tencent Music is a music subscription service like Spotify, but with a twist. Its best customers are its best customers.

Tencent Music operates a completely different business model from Spotify, one described by Eric Feng as shared-value transactions. For most users the service if free (less than 15% of sales are generated through subscriptions). This opens up the service to the largest possible audience (800m+), maximizing the potential for network effects and downstream monetization. This monetization comes from getting its most active and engaged users, those with the highest willingness to pay for the service, to pay more — in turn converting consumer surplus into revenue. Practically, it does so through social interaction, such as allowing users to sing karaoke together or send each other virtual gifts, gamification, such as fan leaderboards that put users in the running for prizes, as well as plenty of paid-for exclusives like live-streamed concerts or early access to new albums. As a result, 4% of its customers account for 70% of its revenues.

A new way?

When you read this excellent A16z article, you realize that these kinds of shared-value business models based on deep customer engagement are proliferating across Chinese e-commerce, from music to video to books. They offer an alternative to business models based on subscription, which aren’t the right fit for many businesses and which lock consumers into long duration contracts; an alternative to business models based on advertising, which create a conflict of interest between service provider and consumer (as well as plenty of negative externalities); and an alternative to business models based on transactions, which leave a lot of money on the table.

The question is why are these business models emerging out of China? Connie Chan suggests it’s because businesses in China thought about “content consumption in a mobile-first way”. This is certainly true, but it is also because China built its payments infrastructure to support a mobile-first world, one where business models based on micropayments can be viable. As we discussed in our first podcast with David Galbraith, the “US internet” is built on layers and layers of legacy technology that make payments extremely expensive and which effectively force companies into transaction, subscription and ad-based models.

One solution to this problem could be crypto. Crypto is almost infinitely divisible, making the most tiny of micropayments practically possible. In addition, because it circumvents the existing payments intermediaries and their fees, it can make micropayments economically viable. The problem, however, is that — for now at least — distributed ledger technology is nowhere near scalable enough to cope with the volume of e-commerce payments, which would of course multiply if transaction amounts got smaller and if micropayments opened up a larger addressable market (note, for instance, that ad spending is around 2% of GDP, creating a ceiling on ad-based revenue).

Is Blockchain the internet of finance?

Whether the solution comes through crypto or an infrastructure upgrade, not being able to execute on micropayments-based business models is bad for business and bad for consumers. Furthermore, it is likely to give a growing advantage to Chinese vs American internet giants as they battle to colonize the rest of the world.


As usual, words by me, graphics by Dan Colceriu

This article is part of the a p e r t u r e | Hub, a content hub and community platform.

a p e r t u r e

a p e r t u r e is built on the exchange of ideas around technology, strategy and the dynamics of the platform economy. A content hub and a community.

Ben Robinson

Written by

At Pangea, I help scale great companies | At Aperture, I help expose great ideas | Proudly European

a p e r t u r e

a p e r t u r e is built on the exchange of ideas around technology, strategy and the dynamics of the platform economy. A content hub and a community.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade