Were you forwarded this email? Sign up here to get the next edition.

I've had a lovely few weeks with no travel. Bunkered down, lots of recording, and building our first Add To Cart Education course. It should be ready in the next month or so.

But the week after next, I'm back on stage at impact.com's iPX 2026 Sydney, where they're bringing together brands, retailers, agencies, publishers and influencers to unpack the latest partnership trends. If you're in Sydney and want to come along, use the code NATH50 for 50% off. See? Partnerships.

CHEERS, BUSHY

01

Meta lost a lazy $18 billion this week, settling the case that Facebook and Instagram harm and hook children. Derrr. The money goes to US states over ten years, no doubt funded by increased costs. The more interesting implication is the mandated changes for teens: two-hour daily time limits for under-18s, screen time caps, overnight usage blocks, and notifications muted during school hours.

Further signalling the protection to come for teens, New Zealand's PM proposed following Australia's lead with an under-16 social media ban this week. And while that's promising for anyone hoping to get the life back behind the eyes of their teenager, new research says that eight months after our ban, 26% of 13 to 15-year-olds are still using TikTok. It’s only one percentage point below where it sat before the ban came in. Anyone surprised?

It's an awkward time to be marketing to teenagers. Advertising to children has always (and should be) been poo-pooed, but teens have always been a key demo. But today, we’re a long way away from Dolly magazine. I have a feeling these social changes will force brands to get more creative, and probably sneakier, like perfume house Jo Malone turning up in Fortnite. Or maybe, as we found out with Ooshies, the answer isn't targeting kids or teens at all. It's putting more guilt on their parents. 64% of parents with children under 18 said they changed their shopping behaviour because of the Ooshies promotion. Wild.

02

It will come as no surprise to anyone with a finger on the pulse that the checkout is under attack from all angles. A squad of tech companies seems intent on killing the website checkout. The funny thing is that the ones looking to kill it are the ones making the most money from it. Shopify, Stripe and Google.

On a podcast this week, Stripe's President of Technology and Business Will Gaybrick said checkout pages will "just go away" as agentic commerce takes over. You won't necessarily buy it yourself, an agent will do it for you. And if you do buy it yourself, you buy it from any page or site. The moment you decide to buy is the moment you buy. No cycling through to a checkout page.

Which is an interesting concept in itself. If you can drop a checkout button anywhere, the customer can decide anywhere, and conversion friction should fall off a cliff. Google's retail MD had a similar theme talking about turning YouTube into a shoppable channel, saying consumers should be able to buy the moment inspiration strikes. Sophie Neary said Google is making more of YouTube's surfaces shoppable, adding to the billion conversions that came from connected TVs in the past year.

Nothing changes tomorrow. But the next phase of conversion experimentation is AI search, and right behind it is checkout everywhere.

03

Some proper nerdy SEO research out of New York agency Shero this week. They analysed 1,000 Shopify stores to determine whether LLMs actually cite the stores that write the copy. The TLDR was that brands were cited or recommended just 9.5% of the time, and across 1,851 sources cited, only 2.8% were the brand's own page. Even when an LLM was recommending your brand, it was more likely to push people to publishers, Reddit, aggregators, review sites, marketplaces and bigger brands. 59% of citations went to publishers. Getting an LLM to actually point traffic at your site is a bit of a miracle.

And it's only going to get harder now that Google is letting people nominate preferred sources that show up more often in Top Stories, AI Overviews and AI Mode.

So we shouldn't be surprised that AI traffic converts so well. The real question is how much of the rest of our traffic is influenced by AI without ever reaching us. Are we getting a heap of recommendations via a proxy? Anyway, it's a great study with practical tips, breaking down the anatomy of an LLM-readable product page and what to do about descriptions and copied content when setting up product pages.

While we're here, one tidbit. Digital Commerce 360 has started AI commerce rankings in the US to see which brands are getting the most help appearing in AI. The top five this month were Nixon, Online Labels, Everlane, Fashionphile and Custom Ink. On the surface it’s a random bunch, but as the Digital Commerce 360 team pointed out, the thing they share is custom products. Are custom products a mermaid siren that the LLMs can't resist?

04

"Just browsing" used to be a polite way to tell a retail team member you want to be left alone. Now it's borderline offensive. New research from Manhattan Associates found that three-quarters of Australians trust AI shopping recommendations as much as, or more than, a retail sales assistant. 47% trust AI more. As someone who was 15, colour-blind, and had never painted a wall in his life, while working the paint section at Big W, my recommendation is that people take the AI recommendation over the retail team member every single time. Just makes sense.

New US research from PYMNTS adds another layer. Not only do 42% of customers now use their phone while shopping in-store, up from 30% two years ago, but what they're doing on it has changed. Hunting for coupons is the one thing going backwards. Reading product reviews, comparing prices elsewhere, and checking which payment methods are all on the rise.

It's not that they don't trust store associates. It's that more is riding on every purchase as wallets tighten. Customers aren't necessarily chasing the cheapest, they want to know they're getting the right product, quality and value for what they spend. This aligns with Deloitte's new 2026 Retail Report, which says that 40% of shoppers are transactional and only 13% are experiential. Product quality was the most important consideration at checkout.

But wouldn't life be more fun if you painted your whole house the random colour the colour-blind 15-year-old at Big W recommended? Life is for living.

05

It's enough to make any HR manager collapse into their charity cupcakes. When All For Mimi wanted a new social media team member, they put the call out on TikTok with a gloriously loose brief of "no holding up signs", "you have to be aesthetic", "you need to be funny", and “start time tomorrow”. No word on whether they have to wear pink on Wednesdays.

I've watched the video five times and I still have no idea what job title they're hiring for. But it's racked up over 180k views and pages of very public applications are popping up on FYPs everywhere, some of them pulling 90,000 views on their own. Which is great news if you're another brand hiring social talent, like Rach Wilde of York Street Brands, who will happily snap up the leftovers.

ATC PLAYBOOK #657

It’s a false fight. Brand isn't competing with conversion. Brand is what makes conversion cheaper.

You know the pull of the conversion number. The moment leadership finds that one magic metric, it's all anyone wants to talk about, no matter how flimsy it gets when you pull it apart. Meanwhile, brand spend has to justify every dollar. Brand gets called fluffy and is often the first thing cut when budgets tighten.

But maybe we’ve got the argument the wrong way around. Leaders who stay strong with brand don't insist that brand is special or different. They treat it like an investment with an important job and hold it to account with real numbers, same as everything else in the business.

Today’s Playbook is anchored by Dave Dullens, Head of Brand and Consumer at Conquest Sports, which runs Converse in Australia and New Zealand. He’s spent 20 years keeping iconic brands relevant, and he regularly faces this brand vs. conversion battle.

Examples from this Playbook

👟 Dave from Conquest Sports carves out a specific brand budget and labels it non-conversion before it goes anywhere near the algorithm. Meta understands conversion but it does not understand your brand priorities. If you don't ring-fence it, it gets optimised away.

📊 Matt from Tracksuit gives a benchmark to start with. The Binet and Field 60/40 brand-to-activation split, adjusted for your category and size. Brands over-indexed on performance are already seeing CAC climb and growth slow, so bring the category evidence, not gut-feel.

🏦 James from Previously Unavailable explains why your CFO just prefers conversion. Brand sits in the accounts as an operating expense but behaves like a capital expenditure, being paid back over two to three years. Reframe it as a capital decision and the conversation changes.

🥩 Justin from Chief Nutrition is a useful counterpoint if you're early on the journey. No separate brand budget at all, just a brand lens over everything: on brand, off brand, or neutral. Neutral still runs. When you can't wait 18 months for a return, that's the system that survives.