State of Paid Acquisition: Q3 2024 and the Election Squeeze

Meta is showing more ads to a few more people and charging more for each one. That is not a crisis. It is a squeeze, and squeezes reward the businesses that did their math early.

This is the next read in my quarterly State of Paid Acquisition series, following the mid-2024 edition. Each quarter I take the public numbers from the platform that matters most to my clients and boil them down to what an operator should do about it.

The election is now behind us, and I am staying away from politics. I am only interested in what it did to the auction.

What did Meta's Q3 2024 report actually say?

Meta released its third quarter results on October 30, 2024. The numbers that matter to advertisers are few.

The headline figures

Total revenue was $40.59 billion, an increase of 19% year over year. Ad impressions delivered across the Family of Apps increased 7%. The average price per ad increased 11%. Daily active people averaged 3.29 billion for September, up 5%.¹

What the CEO said

Mark Zuckerberg called it a good quarter, driven by AI progress across the company's apps and business.¹

It is a company-level statement. For advertisers, the lines below it are the ones to study.

What is the squeeze, in four quarters of data?

A single quarter can mislead you. A run of them rarely does. Here is Meta's year-over-year change in ad impressions and average price per ad across the last four reports.

The trend

• Q4 2023: impressions up 21%, price per ad up 2%.⁴
• Q1 2024: impressions up 20%, price per ad up 6%.³
• Q2 2024: impressions up 10%, price per ad up 10%.²
• Q3 2024: impressions up 7%, price per ad up 11%.¹

Read the columns, not the rows. Impression growth has been falling, from 21% to 7%. Price growth has been rising, from 2% to 11%. In plain English, the company has been growing revenue less by adding ad volume and more by charging more per ad. Total revenue growth over the same stretch: 25%, 27%, 22% and 19%.¹²³⁴

What I am not saying

I am not saying your CPMs rose 11%. The average price per ad is an average across every advertiser, placement and region on Meta. It moves with the mix of who is buying and where ads show. Your account can be better or worse. Use the figure as a direction of travel and then open your own Ads Manager and check.

Did the election squeeze October's auction?

There is some evidence that it did, with a caveat.

What AdRoll saw

AdRoll's Q4 2024 report, drawing on data from over 20,000 online businesses, says Meta's, TikTok's and Pinterest's ad costs rose toward the end of Q3, and that Meta's rise came earlier. AdRoll ties that to Meta being the only one of the three that allows political ads. The same report put display CPMs up 39% year over year in Q3, and says the peak shopping season is five days shorter than last year's.⁵

The caveat

AdRoll's figures come from AdRoll's customers, who skew toward ecommerce. I would not read the 39% as a Meta number, and I would not claim that politics alone explains Meta's rise. Meta's own report does not break out political spending. What the two sources support together is a modest conclusion: the auction was more crowded in October than the headline growth in impressions would suggest.

What does this mean for a small account?

It means that the cost of being sloppy is higher than it was a year ago.

The arithmetic of a rising price

If the price of reaching someone goes up 11%, your cost per acquired client goes up 11% unless more of those people become clients. Put differently, you need roughly 11% better conversion just to stand still. That is not a creative problem or a bidding problem. It is a funnel problem.

So the places to find the 11% are the places most owners ignore (the same handoffs I mapped in leads are not the product):

• Speed to respond. Every hour a new lead waits is an hour for someone else to reach them first.
• Show rate. A booked call that does not happen is an ad dollar burned.
• Offer clarity. A page that tells the buyer exactly what they get and when.
• Follow-up. The people who did not buy the first time, contacted in a way that is not annoying.

Why this favors the prepared

The scale of the market is large. The IAB's report put 2023 U.S. internet advertising revenue at $225 billion, up 7.3%.⁶ Money keeps flowing into this channel. When prices rise inside a growing market, the winners are rarely the biggest spenders. They are the ones whose math still works at the new price.

A word on signal

When the auction gets pricier, the quality of what the platform learns from you matters more. If your account only reports a page view or a form fill, the system optimizes for people who do that. If it hears about booked calls and real payments, it can look for people who do those. I covered the plumbing in my Conversions API baseline, and I will go deeper on signal soon. Put it on your list now: after you check your break-even, check what your ad account is actually being told counts as a win.

How should you plan the holiday stretch?

Meta's own guidance tells you what to expect, and I laid out a Black Friday pacing plan around it last week. The company said it expects Q4 2024 total revenue of $45 to $48 billion, with foreign currency approximately neutral to year-over-year growth.¹ Against $40.59 billion in Q3, that is a seasonal step up, and a step up in revenue means more advertisers in the auction.

A four-step plan

What I would hold off on

I would hold off on large changes to account structure in the middle of the busiest weeks. Meta's Help Center says that editing an ad, ad set or campaign during the learning phase resets learning and delays optimization.⁷ A holiday is a poor time to experiment with a structure you have not tested.

What does the break-even math look like with real numbers?

Here is an illustration. The figures are made up and round so the arithmetic is easy to follow. They are not a benchmark for your business.

A worked example

Say a program brings in $4,000 of collected cash per client. Delivery costs $800 and you want to keep $1,200 as profit. That leaves $2,000, the most you can pay to acquire that client.

Now work backward through the funnel. If 70% of booked calls show up and 25% of the calls that happen turn into clients, then only 17.5% of booked calls become clients. Multiply $2,000 by 0.175 and you get $350. That is the most you can pay for a booked call.

What an 11% price rise does to that

Suppose your cost per booked call was $320. A rise of 11% takes it to about $355. You were inside the line and now you are $5 over it, on every call, before you spend a dollar more.

That is why I keep saying know your break-even before you scale. The squeeze did not make the offer worse. It moved the line, and only the owner who had drawn the line can see it moved.

The three levers

You can respond three ways: lower the cost per booked call, raise the show rate, or raise the close rate. The first is the one the platform controls. The other two are yours. A five-point improvement in show rate does more to this example than most creative tests will.

What mistakes are common in a squeeze like this?

The same few, year after year.

• Cutting the budget in a panic, then missing the cheap weeks. The reserve exists so you can spend when pressure eases, not so you can hide from the quarter.
• Chasing the cheapest lead. A low cost per lead means little if nobody shows up.
• Rebuilding the account mid-peak. Large structural changes in the busiest weeks reset what the system has learned.
• Reading one platform average as a verdict on your account. It is the forecast, not the window.

What should you watch next quarter?

Three things.

The watch list

• The impression and price trend. If impression growth keeps falling and price keeps rising, the squeeze is structural, not seasonal.
• Your show rate and close rate. These are the numbers that tell you whether you can absorb a higher price.
• Your signal quality. The more accurately the platform sees who actually buys, the better it spends your money. That deserves its own piece and I will write it.

In every squeeze I have worked through, the businesses that came out ahead were not the ones that found a trick. They were the ones that knew their numbers before the squeeze started.

If you want your break-even math checked by someone with no stake in your ad spend, my door is open. Next quarter I will do this again with the Q4 numbers and tell you what changed.

Keep building.Ekai