Spotify Q3 2026 Earnings: More Ads Sold, at Softer Prices

Three days before Spotify reported, in Netflix Q3 2026 Earnings: The Ad Forecast Cut Is About Who Owns the Pipe, I used Spotify as my example of an ad seller without purchase data, and I wrote that owning its own ad exchange "didn't save it." Spotify's third quarter was the first chance to test that, and I got part of it wrong. Ad revenue grew 9% in constant currency, three times the first-half pace, and Spotify credits its own auction for it. What hasn't moved is pricing. The update says music advertising grew on more impressions sold, "partially offset by softness in pricing," the same sentence it has used every quarter since at least the middle of 2025.

The stock rose about 11% on October 22, mostly on gross margin and subscribers, and a lot of the coverage read the ad line as the start of a second engine. I'd read it as Spotify filling more of its inventory at prices it still doesn't set.

Through the first half Spotify added free users at 14% a year while ad revenue grew 3%, so revenue per user fell by about a tenth. In Q3 the gap shrank to about three points. Spotify doesn't disclose impressions or price, so I can't split the gain exactly, but its own description of what drove it hasn't changed in six quarters.

That's the useful part of the table. Pricing was soft when ad growth was zero and it was soft when ad growth was 9%, so whatever brought growth back, it wasn't price. On the call CFO Christian Luiga said the progress was "driven by biddable, which is now up to 40%, and also podcasts." Biddable is the auction, through the Spotify Ad Exchange and Spotify's own Ads Manager, and he was describing its share of ad revenue, which was already close to 40% in Q2. An auction is very good at selling inventory that used to go unsold, and more impressions sold at softer prices is what that looks like.

So here's what held from the Netflix piece and what didn't. Owning the exchange did help, and I was wrong to say it hadn't. But it helped volume. The pricing part of the argument, that an ad seller without proof of purchase doesn't get paid like one with it, is exactly where Spotify's own updates say it is.

Amazon's announcement says buys on Spotify through its DSP are powered by Amazon's "trillions of shopping, streaming, and browsing signals" with "full-funnel measurement and attribution." Amazon DSP has been live on Spotify since October 2025, through the flat first half, and the direct connection to Spotify's exchange came in July, which fits the Q3 timing better. Walmart's DSP followed on October 6, after the quarter closed, according to a PPC Land report I couldn't match to a primary release. Neither Spotify's update nor its call names any of these buyers, so the link between them and the Q3 growth is my inference.

Here's why I think buyers like these get volume without paying up. A buyer with purchase data does bid more for an impression it can match to a shopper, and that's the whole pitch. But it bids against its own alternatives. Amazon can show the same shopper an ad on its own store, Prime Video or Twitch, where it keeps the whole price, so a Spotify impression is only worth paying for when it reaches someone Amazon can't reach as cheaply itself. The impressions it can't match clear at whatever the rest of the auction will pay. That's a hypothesis Spotify's disclosures can't confirm, but it fits a business that sells more and gets less for each. Meta shows the other outcome: in the same quarter its impressions grew 10% and its average price per ad grew 14%, because the conversion happens inside its own systems.

Spotify's proprietary data is rated intact rather than strong, because Apple Music, YouTube and Amazon hold the same kind of listening data, and Q3 fits that. Listening data is great for picking the next song and it fills an auction, but it can't tell an advertiser that someone bought the shoes. Netflix is the exception in its group: its data is rated strong, but that rating rests on viewing data that tells it what to commission, which is a different thing from proof of a purchase. The refresh of Spotify's analysis that ships with this piece leaves every pillar where it was and moves the ad driver from stable to accelerating, because the growth is real.

The best objection is that I'm reading too much into a boilerplate sentence. Auction impressions usually sell for less than direct-sold ones, so a business shifting toward biddable will report softer average pricing even if every channel holds its price. On that reading "softness in pricing" is a mix effect, and the number that counts is revenue per free user, which improved a lot in Q3. Management made that case on the call: Alex Norström said the rebuilt ad stack is "tracking really well against our expectations," and Luiga said Spotify expects "to continue our progress towards double-digit growth" in Q4.

The second objection is that renting the proof is fine. Spotify's audience is huge and attentive, and if Amazon brings measurable outcomes to it, advertisers spend more on Spotify than they would otherwise. Spotify still owns the exchange, the inventory and the listener, and most TV networks did well for decades without measuring a single sale. Ads are also only about a tenth of Spotify's revenue, as Luiga pointed out, so none of this decides the stock.

I agree with a lot of that. The mix effect is probably part of the story and Spotify doesn't disclose enough to rule it out, and I don't think ads decide the Spotify thesis, which rests on Premium pricing and the label deals. Where I still disagree is on ads as a second engine. A mix effect should fade once biddable is most of the business, so if pricing is still soft then, it's not mix. And the buyers bringing the proof sell ads on their own properties with the same data, so they've got every reason to pay Spotify as little as the auction lets them.

I'd hold Spotify for the reasons its stock page gives: Premium pricing, the bundle, and a gross margin that keeps rising. The Q3 ad number is good news for revenue, and I've credited it in the growth driver. I wouldn't credit it as a moat yet, because Spotify filled more of its inventory by letting outside buyers in, and on its own account the price per ad is still slipping.

The correction to the Netflix piece makes its lesson a bit more hopeful than I wrote: an ad seller's own exchange and partner integrations can bring growth even without purchase data. The question for both companies is the same one, and the cheapest place to check it is the gap in the chart above, ad revenue growth against free-user growth, in Spotify's next update.