The Speed Read

  • Meta CPMs are up north of 20% year over year on identical campaigns, with third-party platforms including Northbeam confirming it at aggregate scale.

  • The ANA published a retail media measurement standard with Walmart, Albertsons and Target, and Amazon did not participate.

  • YouTube starts counting a view the moment playback begins on August 24, which raises the headline number without anyone watching more.

  • AI disclosure requirements are now live in California, New York, South Korea and the EU, and the IAB has updated its framework for which creative triggers them.

Your BFCM campaign will probably work. That is the risk.

We are a paid media agency, so take this in the spirit it is offered: your Q4 campaign is not the thing most likely to cost you the season.

Every year at Acceler8, at least one brand on our roster runs the same play. The media team executes, creative is strong, campaigns go live on time.

Traffic spikes. Orders arrive at a volume the brand has never processed before. By every metric in the ad account, it is working.

Then the back end gives out.

Orders sit for two days because the warehouse was staffed for a normal week. The flagship SKU sells out on day two because inventory was forecast against 40% growth and demand came in at 65%. CX tickets stack up behind one part-time person. Customers start posting about shipping delays during the exact window when more new people are looking at the brand than at any other point in the year.

The refunds cost money. The reputational damage costs more, because it lands on the most expensive customers you will acquire all year.

This pattern has always been expensive. This year the math is worse.

Meta CPMs are running north of 20% above last year on identical campaigns, same objective, same budget, same targeting. Impression counts on that spend are down. Meta is funding a large AI infrastructure build, the AI revenue does not offset it yet, and advertisers are covering the gap. Northbeam and other third-party platforms see the same thing in aggregate.

So you are paying a meaningful premium to acquire each customer, and then handing a portion of them an experience that guarantees they never come back. At current media costs, operational readiness is the margin.

For a brand where six weeks carries something like 65% of annual volume, that is the whole year turning on whether a 3PL contract got signed.

The August leadership conversation is four questions. None of them are about campaigns.

Is inventory confirmed, with a date? Domestic sourcing makes August the planning window. International already closed. If nobody can give you an arrival date this week, you are forecasting Q4 revenue against inventory you do not have.

Is the 3PL signed and the warehouse staffed for 2x? October is late. Negotiating peak capacity during peak planning season means paying a premium for whatever is left.

Is CX sized above 2x? Every order generates post-purchase contact at a higher rate than normal volume. Shipping questions, order status, returns. A team built for a steady month does not absorb a 65% spike.

Are you certain about your agency? This one has the shortest fuse. Uncertainty in September leaves two bad options: go into Q4 with a partner you doubt, or hand a new agency six weeks to learn your brand during the most complex season of the year.

Ask all four this week and watch how long the answers take. A slow answer is the answer.

If you want a second set of eyes on the Q4 plan while changing it is still cheap, that is what we do.

Retail media has been grading its own homework and invoicing you for the grade

The ANA released its Retail Media Measurement Standardization report this week, built with Walmart, Albertsons and Target.

Until this week there was no agreed definition of what a retail media network was selling. No standard attribution window. No shared meaning for the word "outcome." Every network reported its own performance using its own math, and every brand paid the invoice.

Target's ad business grew 28.6% year over year last quarter. That money came out of somebody's budget, and at mid-market brands it is increasingly coming out of paid social, on the strength of numbers nobody independently checked.

The most useful recommendation in the report is a standardized 14-day loopback window. Networks with long lookbacks claim credit for purchases your email program and your prospecting already earned. At $30K a month into Roundel or Walmart Connect, the difference between 14 days and 30 changes which channel looks like it deserves next year's increase.

Before you commit Q4 retail media dollars, put three questions to your rep in writing.

What is the attribution window, and can we shorten it. Who validates the sales data besides you. What does "outcome" mean in this report, precisely.

Written answers, not a call. Methodologies firm up remarkably fast when someone has to type them.

Nobody's replying at 9pm. That's when they buy.

Your team can't be on every channel around the clock — but your customers still message after hours. Wati connects your channels into one inbox and its AI-powered automations reply instantly, day or night, so you only step in when it's time to close.

Quick Takes

YouTube changes what a view means on Monday Starting August 24, playback start counts as a view on long-form and live. Your video numbers will rise in Q4 for reasons unrelated to anything your team did. Name the person who owns that annotation and tell the board before the number moves, not in January when you are defending it.

AI disclosure in ads is now law in four jurisdictions California's SB 942 and Article 50 of the EU AI Act both took effect on August 2. New York's synthetic performer law landed in June. The IAB's updated framework draws the line at AI that changes what a viewer could believe is real, so synthetic voices, avatars and digital twins trigger disclosure while drafting copy with a model does not. Ask whoever owns creative whether they can tell you which live assets would trigger it and where those assets are running. If a retail partner's legal team asks first, that is a worse day.

Paid search growth slowed in Q2 for the first time in two years Queries are migrating to AI assistants, and the channel had already been absorbing an organic decline. If paid search is over 30% of your media budget this matters to you, and over 50% it is urgent. No action required in Q4. Put it on the Q1 agenda before performance forces it there.

The Last Word

There is a version of Q4 planning that never leaves the marketing function, where the only real decision is how large the November number gets to be. It feels like leadership because there is a budget attached to it.

The brands that break during BFCM almost always had good campaigns. That is what makes the failure so expensive.

Nobody puts "the warehouse was ready" in a board deck. It is usually the line that decided the quarter.

Know someone who should be reading this?

If this was useful, forward it to one person managing media budgets who doesn't have time to read everything. That's exactly who this is for.