Golden Data

Issue #14 · August 9, 2026

Snap Lost Users and Rose 10%. Pinterest Set a Record and Fell.

Paul BrownWrites Golden Data from Ogden, Utah4 min read

The last issue told you exactly what to watch. Snap on August 3. Pinterest on August 4. If either printed monetization up, engagement flat, and a compressed multiple, the quarter stopped being an anecdote and became the thesis for the category.

Both reported. Both beat. I got the tell wrong.

Snap posted revenue of $1.6 billion, up 19% year over year, against consensus near $1.54 billion. North American daily actives were down 7% year over year to 92 million, the worst engagement line in the group. The stock closed August 4 at $5.77, up 14% from the pre-earnings close.

The platform that lost users got paid.

Except Snap did not lose users in the frame the market was watching. North American dailies were flat sequentially at 92 million, after sliding from 99 million a year earlier. Spiegel spent the call on one word: stabilization. Every price target that moved on August 4 moved on the sequential line. I was reading the year-over-year number, which happened to be the number that confirmed what I already believed.

The number the market was reading instead was North American ARPU: $10.26, up 23% year over year. Snap monetized a shrinking audience 23% harder. That is the actual explanation for why the user line did not matter, and I missed it because I was watching headcount rather than yield.

Engagement was the tell. I was holding it up to the wrong light.

What’s actually moving in the market

Snap’s guide priced revenue and left margin to do its own work. Third-quarter revenue guidance came in at $1.70–1.74 billion, roughly consensus, with adjusted EBITDA of $300–350 million against the $250 million it had just printed. UBS, Wells Fargo, Citi and Guggenheim all raised Snap price targets and all kept Neutral or Equal Weight. Truist and Bank of America cut theirs to $7 from $8. That is the sell side saying it cannot tell a turn from a World Cup either. Snap gave most of the move back the next session, closing August 5 near $5.23.

Pinterest beat on everything and got sold on what it plans to spend. Revenue of $1.18 billion, up 18%, with a record 640 million monthly actives, up 11%, and adjusted earnings of $0.43 against $0.36 expected. Adjusted EBITDA hit $311.3 million, up 24% year over year, with margin expanding to 26% from 25%. The stock closed the regular session at $25.58, up 5.9%, then fell 9.2% after hours to $23.22 on third-quarter revenue guidance of $1.19–1.21 billion — 13 to 15% growth measured against the 18% it had just delivered.

The consensus read is that growth got repriced. The better read is that the market was pricing the cost of the AI buildout. Pinterest guided EBITDA above the $334.8 million estimate and got nothing for it. The market was not ignoring margin. It was looking through current margin to what GPU capacity, stock-based compensation, and the international growth gap would do to it. That connects to a thread this letter has been running since May: the AI budget getting a meter. Here is the public market applying one to an ad platform and not liking the number.

The measurement-window problem is the piece itself. Every percentage in this letter depends on which window you measure. Snap’s move was 7.5% at the August 3 regular close, 17.3% including after-hours, 14% at the August 4 close, and roughly 9% at the August 5 close when the give-back landed. Pinterest’s after-hours drop was 9.2% per Investing.com, 8.5% per the same outlet’s transcript piece, 8.9% per ChartMill, and 7% per CNBC. Same events. The spread exists because the window is a choice. I am stating this rather than picking the most convenient number, because picking the most convenient number is what the last issue did with the engagement line.

What I’d watch

Whether the Snap turn survives the comp. Third quarter has no World Cup and a harder base. Snap guided to $1.70–1.74 billion anyway. If Q3 revenue lands in that range with North American dailies still flat at 92 million, the sequential line is the number this category trades on and the stabilization read was right. If dailies resume falling, August 4 was a spending anomaly priced as a turnaround. November 4 settles it.

Whether Pinterest’s margin guide gets paid later. It guided profitability above consensus and received nothing for it. Either the market comes back for that once the growth comp normalizes in the fourth quarter, or the category has decided ad platforms are growth stories and margin is table stakes. That answer sets the comparable multiple for every private company benchmarking against these two.

Sources

Send me an email and we will talk. If something here landed close to what you're working on, the door is open. No calendar funnel, no pitch deck — I read every note that comes in.

Doing the work rather than deciding what to build? Crafting is the column for that chair.

Email meRead more Signal

← All Signal issues · Drafted with Claude · Edited by Paul Brown