Spotify: how a free, ad-funded tier fed a paying audience

The situation

At the end of 2017, Spotify had 159 million monthly active users, a rise of 29 per cent on a year earlier. Of those, 71 million paid for Premium (up 46 per cent) and 92 million used the free, ad-supported service, according to the company’s F-1 filing with the US Securities and Exchange Commission, made in early 2018. Revenue for the year was 4,090 million euros. The company also lost 1,235 million euros, and its accumulated deficit stood at 2,427 million euros.

Those figures sit awkwardly together. Spotify was growing fast, it was the largest music streaming service in the world by subscribers, and it was still losing money on a scale that would worry most investors. At the centre of the business was a choice made years earlier: let anyone listen for nothing, funded by advertising, and hope enough of them would later pay. The filing said the free service drove more than 60 per cent of all gross new Premium subscribers, a figure tracked since February 2014.

The question for this case is whether giving the product away was a sound way to build a paying audience or an expensive habit that the music labels were underwriting. The answer shaped how Spotify set its prices, how it negotiated with Universal, Sony and Warner, and how it explained itself to public-market investors.

Background

A Swedish start-up with a licensing problem

The F-1 dates the company’s inception to April 2006 and records its incorporation in Luxembourg on 27 December 2006. The founders were Daniel Ek and Martin Lorentzon. The service itself launched in 2008. From the start, the product depended on rights that belonged to other people. A streaming service has no catalogue of its own. It licenses recordings and compositions, and pays the owners every time a track is played.

That dependence is the single most important fact about Spotify’s economics. The F-1 states that the three major labels, Universal Music Group, Sony Music Entertainment and Warner Music Group, together with the independent licensing body Merlin, accounted for around 87 per cent of streams in 2017. The company told investors it had paid more than 8 billion euros in royalties since launch. Whatever Spotify charged, and whatever it gave away, a large share of what it earned went straight back to rights holders.

The 2011 squeeze on free listening

The free tier was not always as open as it later became. In April 2011 Spotify announced limits on what free users in its home markets could do. According to a report carried by Phys.org, each track could be played free up to five times, and total free listening was capped at ten hours a month after an introductory period of six months. Users who had signed up on or before 1 November 2010 faced the play limit from 1 May. Spotify said the changes would hit mainly heavy free users, because most people used the service to discover music, and that the average user would not reach the play limit on seven out of ten tracks after a year.

The same report gave some scale for the time. Spotify had reached one million paying customers, a number that had doubled in under a year, and paying subscribers made up about 15 per cent of users, which implied a total of roughly 6.5 million. A separate account from Music Row said the free allowance had previously been 40 hours a month and fell to ten. It also described figures from a report prepared for a rights holder, seen by the trade publication Music Ally rather than released by Spotify. For the period from March to June 2011, those figures showed Spotify adding 520,000 paying subscribers while losing 1.6 million free users. The article itself noted that other numbers it quoted for departing users did not match, so the exact scale of the loss is uncertain. The direction is still clear: tightening the free tier pushed some people to pay and others to leave.

How the money was made by 2017

By 2017 the revenue mix was heavily weighted to subscriptions. The Irish Times, summarising the F-1, reported that Premium produced about 90 per cent of revenue, and that nearly 45 per cent of users were paying, against roughly 25 per cent some years earlier. The standard Premium price was 9.99 euros a month. Revenue grew from 1,940 million euros in 2015 to 2,952 million in 2016 and 4,090 million in 2017. Net losses over the same three years were 230 million, 539 million and 1,235 million euros.

The industry

Recorded music had been through a long period of falling sales before streaming. Services such as Spotify offered a legal alternative to copying, and the labels agreed to licence catalogues on terms that gave them a large share of each pound or euro earned. The F-1 shows how concentrated that licensing was. A company that wanted a full catalogue had to deal with the same few owners as every rival.

That concentration shaped the rules of competition in three ways. First, the biggest cost was fixed by negotiation, not by efficiency. A streaming service could not simply run leaner to lift its gross margin, because most of the cost of sales was royalties. Second, rivals sold almost identical catalogues, so differentiation had to come from price, discovery, features and distribution. Third, the labels held leverage over free tiers in particular, since a free listener generates far less money per play than a paying one. In 2017 Spotify renewed its licence with Universal, and the Irish Times reported that the deal included a two-week period when Universal’s new releases would be available only to Premium subscribers. The same article said Spotify avoided exclusive content of its own for paying users and relied instead on personalisation. Algorithmic personal playlists accounted for 17 per cent of listening time, and editorial playlists for another 15 per cent.

The nearest rival was Apple Music, which the Irish Times reported had 36 million subscribers at the time, about half of Spotify’s 71 million. Apple sold its service through a very large installed base of phones. Spotify had no hardware and no operating system, which meant that it had to win listeners on the strength of the product and the price of entry. A free tier cost the labels and Spotify real money, but it removed the main barrier to a first try.

The decision

The decision at the heart of the case was how much of Spotify’s service to give away, and for how long. The F-1 and the press coverage around it show the options clearly, even though Spotify did not publish a formal list of them.

Option one: free for everyone, paid for by advertising

This is the model Spotify chose. Anyone could listen without paying, with advertising between tracks and some limits on features. The upside was reach. A person could try the product in minutes, build playlists and form a habit. The downside was cost. Every free stream carried a royalty, and advertising revenue per listener was lower than subscription revenue per payer.

The Fox Business summary of the F-1 gave the numbers. In 2017 the ad-supported business had a gross margin of 10 per cent, against 22 per cent for Premium. It had made losses for a long time and turned positive only in 2017, after Spotify negotiated better licensing terms with the labels. Management had treated the ad-supported segment as a marketing expense. The article noted that its losses in 2015 and 2016 were a fraction of sales and marketing spending, which was 219 million euros in 2015 and 368 million euros in 2016.

Option two: subscription only, with a short trial

A company in Spotify’s position could have charged from the first day and offered only a limited trial. That would have cut royalty costs on non-paying users and made every listener a paying customer or a lost one. It would also have narrowed the top of the funnel. The F-1 figure that more than 60 per cent of gross new Premium subscribers came through the free service suggests how much of the paying base would have been missing. A paid-only service would also have made Spotify look like a direct competitor to Apple, which could present its offering to every iPhone owner.

Option three: a free tier that is deliberately narrow

The third course was the one Spotify tried in 2011: keep the free tier but squeeze it so that heavy users find it uncomfortable. The 2011 limits, five plays per track and ten hours a month, were the clearest example. The Music Row figures show that the squeeze moved people in both directions, with more payers and fewer free users. The risk was that a narrow free tier loses its value as a showroom. If people cannot build a habit on it, they have little reason to stay long enough to pay.

What they did

Keep the free tier and tune it

Through the middle of the decade Spotify kept the free tier open and kept working on what paying customers got that free ones did not. By 2017 the F-1 described the ad-supported service as the main acquisition channel for Premium. The company reported that Premium subscribers streamed more than three times as much content per month as ad-supported users, a sign that the paid product was used far more heavily.

Use discounts and family plans to lift the paying base

The Irish Times reported that at least a fifth of new Premium subscribers in the preceding two years had come through discounted offers, such as a free month or three months for 99 cents. It also covered the effect on revenue per user. Average revenue per Premium user fell from 6.84 euros in 2015 to 6.20 euros in 2016 and 5.32 euros in 2017. Spotify linked the fall mainly to its family plan, which allowed up to six accounts for 14.99 euros a month, and to a student plan at half price. These plans lowered the average, but the article said they also reduced cancellations. Fox Business made the same point about families: moving a whole household to another service is hard, so family plans tend to keep payers.

Negotiate better terms with the labels

The most consequential move was commercial. Fox Business reported that the ad-supported segment turned profitable at gross level only after Spotify negotiated more favourable licensing terms. The 2017 Universal renewal, with its two-week Premium-only window for new releases, shows what Spotify gave in return: a way of rewarding paying customers that did not require Spotify to buy exclusive content itself. The window gave the labels something they wanted, which was a reason for listeners to pay.

Take the story to investors

In March 2018, Fox Business reported, chief financial officer Barry McCarthy told investors at an investor day that a free user who converts to Premium takes about twelve months to repay the losses the free tier incurred on that user, after which lifetime value keeps rising. That was the company’s case for treating the free tier as a form of marketing spend rather than a drain. The figures in the F-1 were the evidence behind it.

What happened

The results over the next few years fit the pattern the company described. The strongest evidence is the long run of subscriber growth and the move from loss to profit.

Spotify’s results for the fourth quarter of 2025, reported on 10 February 2026, were covered by TheWrap and Billboard. Monthly active users reached 751 million, a record, and Premium subscribers 290 million, up 10 per cent on a year earlier. Revenue was 4.53 billion euros, with Premium revenue of 4.01 billion euros. Operating income was 701 million euros, against 582 million a year earlier, and net income was 1.17 billion euros, up from 899 million. Billboard gave a gross margin of 33.1 per cent, compared with the 20.8 per cent implied for 2017 by the F-1 gross profit of 849 million euros on revenue of 4,090 million. The 2017 figure is a calculation from the filing’s gross profit and revenue, not a margin the filing states.

The ratio of payers to users is another useful measure. In 2017, 71 million of 159 million users paid, about 45 per cent. At the end of 2025, 290 million of 751 million did, about 39 per cent. The share of payers fell as the free base grew faster in absolute terms, even as the paying base more than quadrupled. Growth in reach did not come at the expense of a larger paying group.

The ad-supported side tells a more mixed story. Ad-supported revenue in the fourth quarter of 2025 was 518 million euros, down 4 per cent from a year earlier according to TheWrap. That is roughly 11 per cent of total revenue for the quarter. For a segment that drives most new Premium signups, it was a small and, in that quarter, shrinking source of income.

Other views

The labels’ interest

The F-1 lists dependence on third-party licences and a lack of control over content providers among Spotify’s main risks. Rights holders have little reason to favour a free tier for its own sake, since the royalty per free stream is low. The Premium-only window for new releases in the 2017 Universal deal is evidence that the majors wanted the free tier held back relative to paid. Any reading of the case that treats the free tier as Spotify’s decision alone ignores that the labels set its terms.

The cost of free

The Fox Business figures give critics something to work with. A 10 per cent gross margin on the ad-supported service meant a very thin cushion, and the segment was only just making a positive gross margin in 2017. The same article said the company’s own estimate of a twelve-month payback was an average claim at a single investor event. It did not publish the free-to-paid conversion rate or ad revenue per free user, so outsiders could not check the claim from the filing alone. Neither figure appears in the Fox article.

Discounting and falling revenue per user

The decline in average Premium revenue per user, from 6.84 euros to 5.32 euros over two years, shows what Spotify paid for subscriber growth. Spotify accepted that trade. Whether the lower per-user revenue was a sound price for lower churn depends on lifetime value, which the filing does not give. Churn for 2017 was 5.5 per cent a year on the company’s own measure, according to the F-1 summary.

Limits of the evidence

The 2011 numbers come from a document that Spotify did not release, and the article reporting them noted an inconsistency in its own figures. The 2017 numbers are a company’s own account made at the point of a stock market listing. They cover a single period. They do not prove that free users convert at a stable rate, only that more than 60 per cent of new payers had come through the free service by then.

Reading it through a framework

Freemium economics as a funnel

Freemium is a conversion funnel in which the free tier acts as the top and the paid tier as the bottom. Spotify’s numbers fit this closely. The free service had 92 million users at the end of 2017, and it supplied more than 60 per cent of gross new Premium subscribers. The cost of the funnel showed up as a 10 per cent gross margin on the ad-supported side and as sales and marketing spend of 219 million and 368 million euros in the two years before the filing. The twelve-month payback reported by Fox Business is the funnel’s unit economics in one line. The framework also shows where it can fail: if conversion falls or the free segment costs more than it returns, the model is a subsidy.

Bargaining power and Porter’s five forces

Porter’s five forces help with the supplier side. The labels hold a near-monopoly over the key input, around 87 per cent of streams in 2017, so supplier power is very high. That explains why Spotify’s gross margin was 20.8 per cent on the implied 2017 calculation. Buyer power is moderate: consumers can switch to Apple Music, which had 36 million subscribers, but family plans raise the cost of switching. The threat of new entrants from large technology companies is real, since Apple was already there. Rivalry is intense because catalogues are the same. The framework suggests that Spotify’s real options for margin were bargaining with the labels, raising prices or changing the mix of products, and that is what the case shows it did.

Since then

The F-1 was the start of Spotify’s life as a listed company. The sources used here show the position by early 2026. Monthly active users were 751 million and Premium subscribers 290 million. Operating income for the fourth quarter of 2025 was 701 million euros and net income 1.17 billion euros, a long way from the 1,235 million euro loss of 2017. Gross margin was 33.1 per cent. Billboard also reported that Spotify paid out more than 11 billion dollars to artists in 2025 and was rolling out official music videos to Premium subscribers in the United States and Canada, a feature that executives said both subscribers and advertisers wanted.

The free tier remained the entry point, but the paying base was the profit engine. Ad-supported revenue was falling in the quarter reported, which raises the question of how long advertising can carry the top of the funnel.

What to take from it

• A free product can work as paid marketing if it brings in a measurable share of paying customers. Spotify’s free service supplied more than 60 per cent of gross new Premium subscribers in the period the F-1 tracked.
• When a supplier controls the main input, margin comes from negotiation. Spotify’s ad-supported segment turned positive at gross level only after it renegotiated terms with the labels.
• Growth in subscribers can lower revenue per customer. Premium revenue per user fell from 6.84 euros to 5.32 euros between 2015 and 2017 as family, student and discount offers spread.
• A free tier needs limits that are tuned, not fixed. The 2011 restrictions moved some users to pay and others away, so the setting has a cost on both sides.
• Company-reported conversion and payback claims are worth testing. The twelve-month payback came from one investor event and could not be checked from the filing.

Questions to think about

1. If the labels earn little from free streams, why would they agree to a free tier at all, and what did Spotify have to offer in return?
2. Spotify’s share of paying users fell from about 45 per cent to about 39 per cent between 2017 and 2025 while the paying base grew. Is that a sign of weakness or of success in reaching new markets?
3. What evidence would you want before deciding whether the free tier is marketing or a subsidy?
4. How might the answer differ for a company that owns its own content?

Concepts at work

This case uses freemium conversion, supplier power and customer lifetime value. Explore them in Insights.

Written from the public sources listed below. Figures are as those sources report them.