State of Paid Acquisition: Q2 2026

In the second quarter of 2026, Meta delivered 14% more ads than a year earlier, and charged 12% more for each one on average.¹ That single pairing is the quarter for anyone buying traffic.

This is my read of the latest numbers as an advertiser, not an investor. Everything below comes from Meta's own Q2 and Q1 2026 results releases, plus a few Meta sources on how delivery and measurement have changed. Where I do arithmetic or give an opinion, I say so.

What did Meta report for Q2 2026, and how does it compare with Q1?

Start with the two headline numbers an advertiser should care about.

Here is the side-by-side, taken from the two releases.

Sources for the table are the Q1 and Q2 releases.²¹ Quarter over quarter, advertising revenue rose about 7.9% by my arithmetic ($59,363M against $55,024M). Q2 total revenue also landed near the top of the $58 billion to $61 billion range Meta guided in April.²

So the business is still growing fast. The interesting part is what it is growing on.

Where did growth slow, and does it matter for advertisers?

Yes, and the reason is the useful part. Meta's own wording is the same in both quarters on price, and different on volume.

Volume slowed. Price did not.

Impressions growth fell from 19% in Q1 to 14% in Q2.²¹ Price per ad stayed at 12%. Multiply them out, rounded, and you get roughly 1.19 times 1.12, or about 33% growth in Q1, and 1.14 times 1.12, or about 28% in Q2. That is my arithmetic, but it lines up with the reported ad revenue growth of 33% and 27%.

Put plainly, the whole six-point slowdown in ad revenue growth came from fewer additional impressions, not from cheaper ones. If advertising demand were softening, I would expect price growth to give first. It did not.

Two readings, and I cannot separate them

One reading is a more crowded auction: more money chasing a supply of attention that is growing more slowly. The other is that Meta is delivering more valuable ads. Meta said in November 2025 that its GEM ads model delivered a 5% increase in ad conversions on Instagram and 3% on Facebook Feed in Q2 2025.³ Better results make advertisers willing to pay more.

The release cannot tell me which effect dominates. For planning, it does not matter.

Ads are growing faster than people

Family daily active people averaged 3.60 billion in June, up 3% year over year, while impressions rose 14%.¹ By rough arithmetic (1.14 divided by 1.03, comparing a quarterly total with a monthly average), that is about 11% more ads per person than a year ago.

More ads competing for the same attention makes this a creative problem as much as a bidding problem. When people see more ads, the ones that stop a thumb are worth more. Either way, the cost of reaching the same person is not falling.

What does a 12% rise in price per ad mean for my account?

Less than the headline and more than a shrug. It is a company-wide average across every advertiser, placement and country, so it will not equal your CPM, your CPA or your competitor's. Do not paste "+12%" into a spreadsheet and call it a forecast.

What you can take from it is direction. Two quarters in a row at 12% says the baseline for acquisition cost is drifting up, not down, and that a plan built on cheaper traffic is a plan built on hope.

I would also check the unit economics before touching the ads. If a lead or a sale only pays back when traffic stays cheap, the quarter just told you how fragile that is.

What should you not read into the Q2 profit drop?

Not a weaker ad business. Operating margin fell from 43% to 31% year over year, and operating income dropped 8% to $18.8 billion, but the cost side explains it.¹

Total costs and expenses were $42.0 billion, up 55%, and included $2.40 billion in legal charges and $1.18 billion in severance tied to a May 2026 headcount reduction.¹ Capital expenditures including finance leases were $31.08 billion in Q2, up from $19.84 billion in Q1, and Meta narrowed its full-year capex outlook to $130 billion to $145 billion.¹²

You do not need to follow Meta's balance sheet to run ads. The point is narrower: the company is pouring money into AI infrastructure and saying it speeds up the advertising business. Those improvements reach your account through delivery, not through a discount.

What has changed in how Meta delivers and how you measure?

Two shifts matter for comparing this half-year with the last.

Delivery is more model-driven

GEM is described by Meta as its most advanced ads foundation model. It is built to improve other ads recommendation models, and it learns from inputs including advertiser goals, creative formats and measurement signals.³ Meta also notes that meaningful signals such as clicks and conversions are very sparse.³ That puts a premium on what you feed it: distinct creative and clean conversion events.

Your reporting windows changed

On January 12, 2026, the 7-day view and 28-day view attribution windows stopped returning data in the Ads Insights API.⁴ The change was framed as an API change, and the 1-day click, 7-day click, 28-day click, 1-day engaged view and 1-day view windows remained.⁴

If your dashboard reads from that API, comparing 2026 cost per result with 2025 is no longer like for like. Q1 and Q2 of 2026 are comparable with each other, so set your H2 baseline from them and stop measuring against last year's numbers.

How should you set budget and creative priorities for the back half of 2026?

Here is the order I would work in. These are my priorities, not Meta's.

Budget: volume before cleverness

Meta's guidance is that learning tends to end at roughly 50 optimization events in a week, and that editing during learning starts the clock again.⁶ With prices rising, wasting budget on constant resets or on too many thin ad sets is the most expensive habit you can have. Combine similar ad sets, hold your edits, and let data accumulate.⁶

Creative: the lever you actually control

You do not set the auction price. You do control what the system gets to work with. When delivery is model-driven, a handful of truly different concepts (the playbook is in Media Buying in the Andromeda Era) gives it more to choose from than a pile of near-duplicates. I would rather launch four different ideas with a clear hypothesis each than twenty cosmetic variants.

Dara Denney reached the same place in her Andromeda and GEM explainer, posted on December 21, 2025, which carries a Particle sponsor read: "Andromeda and [GEM] really does reward brands who are supplying great creative volume, but more importantly creative diversity."⁷

Signal: clean beats clever

Pair the browser pixel with server-side events, deduplicated, so each lead or sale is counted once and arrives reliably. That is a core part of every build I do for clients, because every optimization setting downstream is only as good as the event underneath it.

Targets: leave room

Meta's guidance for goal-based bidding is to start 10 to 20% above your recent average CPA for a cost per result goal, and to use a highest volume campaign first if you have no baseline.⁵ More in Bid Strategies for Lead Gen in 2026. In a pricier auction that headroom matters. A target tighter than your history is a spending cap in disguise.

The page: your cheapest discount

If acquiring a click costs more, then converting a bigger share of those clicks is the one improvement that is entirely in your control. Check load speed on a phone, and check the page says what the ad said (landing page speed and message match covers both), before you commission new ads.

What am I watching going into Q3?

Three things, in order.

First, Meta's own guidance. It expects third quarter 2026 total revenue of $61 billion to $64 billion.¹ The midpoint is about 2.8% above Q2's $60.8 billion by my arithmetic, so Meta expects the business to keep climbing. Guidance is a forecast, not a promise.

Second, whether impressions growth holds near 14% or slips further. If volume keeps cooling while price per ad holds at 12%, the squeeze on advertisers gets tighter.

Third, whether price per ad stays at 12%. A third quarter in a row would make it a trend I plan around instead of a data point I watch.

I also plan as if the fourth quarter will be crowded. That comes from experience, not from Meta's release, and it is a reason to have your signal, your targets and your pages sorted by September, not in November.

The takeaway is not that Meta ads got worse. It is that they got more expensive to be sloppy with. The accounts that do well in the second half will not be the ones with the cleverest hack. They will be the ones with clean data, honest unit economics, and creative worth showing to a stranger.

Keep building.Ekai