Three numbers landed in the last two weeks.

→ Retailers booked billions in tariff refunds and spent it on shelf price.

→ A six-year-old mac and cheese brand that raised $19.4M sold to Barilla.

→ Meta's average price per ad went up 12%.

Scroll past any one of them and you miss nothing. Put them together and they say the same thing: every cost on your P&L is moving, the two levers you've used since 2022 to absorb that are producing less, and what's left is how cheaply you can buy a customer who wasn't already yours.

August 20. Walmart's CFO tells investors the company was "eligible for approximately $2.9 billion of tariff refunds," about half a percent of annual US net sales, and that they've received nearly all of it.

He wasn't alone. Target booked $994M pre-tax. That's $1.65 of EPS and 3.7 points of margin in one quarter. Gap booked $512M.

And all of them dumped it into price. 11,000+ rollbacks at Walmart, up from 7,200 last quarter. 10,000+ items cut at Target.

Every outlet wrote the consumer version. Tariffs struck down, refunds landed, prices fell, good news at the register.

If you sell into those retailers, that's not your story.

Listen to how they described the exit.

Target's CFO said the same thing from the other side, listing "collaborating with vendors to find offsets" ahead of refunds.

So a retailer set a promo price with money it will never see again, and the plan for keeping that price is a cost-down conversation. With you.

Make them say what funded the rollback. The answer is a refund they already spent, and your landed cost didn't move.

Credit where it's due: Gap actually shared. It paid out $95M to vendors, roughly 19% of what it booked. It's the only one we found.

A tariff refund belongs to the importer of record.

Buy DDP from an Asian supplier, or import through a distributor, or let your 3PL be the IOR? Then the duty sat in your landed cost and the refund went to somebody else's bank account. Nobody's calling to tell you.

And they're not automatic. A July 15 CIT order had CBP refund finally-liquidated entries for roughly 3,700 plaintiffs: importers who actually sued. Everyone else is waiting on class certification.

We're not telling you you're owed money. Certification is unresolved, the mechanics are trade lawyers reading a court order, and entries more than 180 days past liquidation without a protest may be gone. We're telling you there might be a five or six figure number sitting in your supply chain, nobody in that chain wants to raise it, and finding out costs one email.

→ Email your supplier or broker: who was the importer of record on our entries since 2025, and was a claim filed?

→ Count your entries more than 180 days past liquidation with no protest. That's your expiry list.

→ Going into a line review? Bring the slide showing your landed cost didn't fall.

P.S. from Parker: this one's close to home so I'll just say it. I've spent the last year building Evana, which does exactly this: pulls your actual CBP data and finds the duty you overpaid and the refunds nobody claimed. We work with Fresh Clean Threads, SandCloud and a bunch of others. If tariffs have made your COGS unreadable, reach out. Do the free version above first though. It costs nothing and you might not need us.

September 2. Barilla agreed to buy Goodles.

Look at the shape of it. Founded October 2020. 73 people. Santa Cruz. About $19.4M raised total, a seed plus a $13M Series A led by L Catterton. No Series B, ever. Gal Gadot is a named founding partner. Team stays, HQ stays, CEO stays, brand runs standalone.

No price. No revenue. We looked hard and there's no real number out there, so we're not making one up. The only tell is the banker: Centerview ran the sell side, and Centerview doesn't show up on small deals. That's a signal, not a number.

Not velocity. Everyone in that aisle has velocity.

Goodles sells more than three units per second, about 260,000 a day. Fine. Here's the number underneath it.

At one mass retailer, 84% of Goodles buyers were new to the category. At another, 81%. At a club account, 80%. Goodles is "close to 90% of the growth in adoption in macaroni and cheese."

(All of that is company-provided and unaudited. It's also the most specific thing anyone disclosed about this deal.)

Here's why it matters.

Most brands' volume is substitution. You win a unit, a competitor loses one, the shopper spends the same money in the same aisle. Real business. But the category didn't grow, so the retailer is neutral and a strategic won't pay up for it.

Goodles' volume was new people walking into the category. That's the one thing a category leader can't build from the inside. So it buys it.

Meta's average price per ad rose 12% year over year in Q2, impressions up 14%. That's Meta's own filing, not an agency benchmark.

You can't price around it either. Circana's July read: total US retail -1.0% in dollars on -2.0% in units, but food and beverage +0.5% in dollars on +1.6% in units. Units growing faster than dollars is price deflation, about a point of it.

And discounting is tired. Circana's Kiara Barrett, on the record: "Promotional activity remains important, but broad discounts alone are generating less incremental demand."

(Circana is measured-channel POS. It under-counts DTC and chunks of Amazon, so treat it as the shape of the market, not your number.)

So you can't take price. Discounting does less. And the customer costs 12% more than a year ago. The only variable left is whether the demand you're buying is incremental or just switching, and almost nobody measures which.

A brand that exits on under $20M of capital is a brand whose marketing made demand instead of catching it. That's not a branding line. It's a page in a diligence file.

→ Ask your retail buyer for the category incrementality read on your item. Most can run it. Almost nobody asks.the

→ Recalculate contribution margin per new customer at today's CPMs. If your model is built on 2025 media prices it's off by about a tenth.

→ Put total capital raised next to where you actually are. Goodles got to an exit on $19.4M.

P.S. from Graham: the acquisition half of that squeeze is my day job. I run Pixel Theory, and most of what we do right now is exactly this: getting more out of the same budget when clicks keep getting more expensive, and working out which spend is actually bringing in somebody new. Happy to look at an account and tell you if there's room in it. Sometimes there isn't and we'll say so.

Smallest number of the three, which is exactly why it'll slip past you.

On the August 5 call, Klaviyo reported non-GAAP gross margin of 73.4%, down 3 points, because of "higher carrier fees, which we chose to fully absorb in prior quarters."

Read the tense again. Chose to. Prior quarters.

They've stopped. Klaviyo changed pricing in Q3 to "pass through the carrier fee increases that we have seen to customers," landing gradually, at renewal, and neutral to Klaviyo's revenue and margin.

"Neutral to Klaviyo" is the whole thing. It's not a margin grab, it's a cost moving off their P&L and onto your invoice. Three points of it. No announcement. Just a slightly different rate on a renewal you sign in a hurry.

What we can't tell you: Klaviyo hasn't published a merchant rate card, so anyone quoting you a percentage is guessing. Go ask for yours.

Wrong reaction: send less SMS.

Right one: a higher cost per message re-ranks your whole program, and most brands rank it wrong to begin with. They rank by attributed revenue. The messages at the top of a revenue list are usually the highest-volume, lowest-margin ones.

What we see in our accounts: high-volume, low-AOV broadcast goes underwater first. Lifecycle and subscription messages hold up, because they defend revenue that already exists. An upcoming-order notice, a failed-payment recovery, a real skip-versus-cancel screen. Tiny send volume, and they save a subscriber who was already paying you.

Which means the brands most exposed are the ones sending the most and keeping the least.

→ Ask your CSM for the new carrier rate card and when it hits your contract. Not the pricing page.

→ Re-rank every flow by contribution margin, not attributed revenue. Cut from the bottom of that list.

P.S. from Parker: this is what Bylders does all day, email and SMS retention for subscription and repeat-purchase brands. Want a second set of eyes on what your SMS actually contributes once you net out the carrier cost? Reply and we'll go through it. No deck.

Quick one, because the last fortnight was strange. Six consumer acquisitions landed in 48 hours on September 1 and 2. Across the whole window since August 20 we counted eleven. Four disclosed a price. Two disclosed revenue.

→ Nestlé sold its mainstream vitamins business to Yellow Wood for $1.0B on $1.2B of 2025 sales. That's 0.83x, and it includes Nature's Bounty, Puritan's Pride, Nuun, and the US private-label operation with its plants. Nestlé kept Solgar and Pure Encapsulations. Everyone's running "paid $5.75B in 2021, sold for $1B." That's wrong: the 2021 deal included Solgar and they're keeping it. It's a sort, not a fire sale.

→ Hold that next to the deal we wrote about four weeks ago. P&G bought Thorne for $3.8B on reported revenue over $500M. Roughly 7.6x. Same category, 28 days apart, about nine times the multiple. (Both revenue multiples, not EBITDA, and the Thorne one is our arithmetic on a trade-reported number.)

→ Medici Brands raised $250M, co-led by Greenoaks and Valor. That's David Protein, HallPass and Epogee, which we spent a whole issue on eight days ago. Bloomberg reported a $2.25B valuation; the company's release doesn't state one.

→ Drake's OVO sold in two pieces. Authentic took majority of the IP. Vince Holding separately bought the operating business. Brand equity and operations, valued and sold to different buyers.

→ Keurig Dr Pepper took $925M off the table: $800M for its Chobani stake, $125M for a plant. It put $300M into La Colombe in 2023 and exited at $800M in under three years. And it kept the distribution agreement and the K-Cup license. Sold the shares, held the channel.

→ And two that rhyme: Regent bought Avon North America, Skyline bought Lumin and Meridian out of Pangaea. Both buyers are distressed-asset consolidators. Skyline's CEO says his box is $20M to $50M revenue at 3x to 4x EBITDA. No prices. The Lumin and Meridian teams weren't kept.

Two things worth taking from that.

Sellers publish flattering numbers. Seven of eleven going quiet in one fortnight tells you something about where consumer valuations actually are.

And the only clean multiple in the window was 0.83x. The one from four weeks earlier was 7.6x. That gap doesn't get built in your next quarter. It gets built in whether a customer has a reason to come back that has nothing to do with your price.

Here's the thing. We're obsessed with this stuff. Graham and I spend our days inside these accounts.

So: hit reply.

Not a form, not a Calendly, not a funnel. Reply to this email and tell us what you're working on, whether that's cracking the second purchase, working out what your media can afford, or just getting your attribution to stop lying to you.

We're not going to pitch you. We'd love to work with you if it fits, and we'll say so plainly if it doesn't. Mostly we'd like to be friends of your business. Tell us what's broken and we'll tell you what we've seen work.

Parker & Graham

Sources: Walmart Q2 FY27 · Walmart call transcript · Target Q2 2026 · Gap Q2 FY2026 · Holland & Knight on the CIT order · Goodles and Barilla · Goodles Series A · Food Business News on Goodles · Meta Q2 2026 · Circana July read · Klaviyo Q2 2026 · Klaviyo call transcript · Nestlé divests mainstream VMS · CNBC on P&G and Thorne · Medici Brands Series B · Authentic and Drake on OVO · Keurig Dr Pepper and Chobani · Regent and Avon North America