Bumble reported second-quarter revenue of $210.5 million, down 15.2% year over year. Paying users fell 16.4% to 3.2 million. Average revenue per paying user rose 1.2% to $21.96. Adjusted EBITDA of $73 million came in above the high end of the company's own guidance.
The stock fell 5% to $3.04, then another 4.6% after hours to $2.91. It now trades at $2.83 against a 52-week high of $7.37.
The quarter
| Metric | Q2 2026 | Change |
|---|---|---|
| Revenue | $210.5M | −15.2% |
| Paying users | 3.2M | −16.4% |
| ARPPU | $21.96 | +1.2% |
| Adjusted EBITDA | $73M | above guidance high end |
| Diluted EPS | −$0.84 | improved from −$2.45 |
| Q3 revenue guidance | $205–213M | ~2.4% below consensus |
| Q3 EBITDA guidance | $56–60M | down from $73M |
Read the two lines together
Revenue down 15%, paying users down 16%, ARPPU up 1%. Put plainly: Bumble is losing customers roughly as fast as its revenue is falling, and squeezing marginally more out of the ones who remain.
That is the least attractive shape a subscription business can have. Revenue declining because of price is fixable — you change the price. Revenue declining because the customers left is a demand problem, and it compounds: fewer users means a thinner network, a thinner network means worse matches, worse matches means more churn.
The EBITDA beat sits on top of that. Seventy-three million dollars against guidance is real, but it came from cost control on a shrinking base. That is a good quarter inside a bad trend.
The guidance is the actual news
Management guided Q3 revenue to $205–213 million, about 2.4% under consensus, and EBITDA down to $56–60 million from $73 million. They are explicitly choosing to spend: expanding the free experience, simplifying subscriptions, and raising brand marketing through the back half of 2026 and into 2027.
I take that at face value as an admission. You do not widen the free tier of a paid product while your paying base is shrinking 16% unless you have concluded that the paywall itself is throttling the top of the funnel. It is the correct call if the diagnosis is right, and it means the next several quarters get worse on purpose before they can get better.
What the price says
Market capitalisation is $433.6 million on trailing revenue of $893.2 million, down 13.3%. Trailing net loss is $480.1 million — dominated by write-downs rather than operations, but it tells you what the accounting has already conceded about earlier valuations. Analysts sit at Hold with a $3.90 target.
A company trading at half of trailing sales with $73 million of quarterly EBITDA is statistically cheap. It is also a company whose customer count fell by a sixth in a year.
How I read it
This is the clearest value-trap setup in the sector, and I mean that as a description rather than a verdict. The cheapness is real, the decline is real, and which one dominates depends on a single variable: whether the free-tier expansion re-fills the funnel before the cash generation degrades.
What would make me interested: one quarter where paying users decline by less than 5%, or where total users grow while paying users fall. The second would mean the free-tier bet is working and monetisation is simply lagging. That is a rebuild. Anything else — revenue stabilising purely on price, EBITDA held up purely by cuts — is the ice cube melting more slowly.
At $2.83 the market is not asking you to believe in a turnaround. It is offering an option on one, priced at a level where being wrong costs real money because the business still burns credibility every quarter it shrinks. I would rather own the stabilisation story at Match, where I get paid a dividend to wait, than the same story here at a lower price with no coupon.