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

I've had so many conversations this past week with people saying it feels like the end of the year is already here. I blame Spring… the big tease. But don’t get too thirsty yet. You've still got Black Friday to get to. It’ll be party season soon enough.

CHEERS, BUSHY

P.S. Don't forget to register for our webinar with StudioHawk, where we'll unpack the data behind a hundred Australian ecommerce brands and what they're seeing in SEO and AI search. Perfect timing in the lead-up to peak.

01

It's amazing when the two loves of your life come together. No, Sarah didn't become a Canberra Raiders member. Anthropic released Claude commerce agents this week and Shopify had its own agent builds running on top of them within the day. It's an open framework for two things: a shopping agent for the front end, so customers can research, compare, get service and have the agent remember them and work with their account inside the chat, and a merchant agent for the ecommerce managers out the back. Handy if you want a sidepiece to Sidekick.

Anthropic are claiming carts up to 35% larger and shoppers 60% more likely to complete a purchase. An interesting callout was that the agents "feature guardrails designed to constrain prices and products to actual catalog data, and avoids manipulative upsell patterns". Very disappointing.

The full GitHub repository is here, and Shopify's examples are here, running the storefront agent over UCP and Sign in with Shop. If it's anything like Shopify's example video, I'll have mine live tonight. The race is on.

While we're talking about two of our favourites coming together, two past Add To Cart guests of the show, PYRA and Birchal, have come together for PYRA's crowdraise. Sam Moore has built it to $11 million in lifetime revenue and is raising to push into North America and Europe. If you want a stake in a growing technical streetwear brand, now's your time. Here's the episode if you missed it.

02

Just like the nephew who tears apart your car looking for coins, Amazon always finds a new way to make a buck. This week, it emailed its US-based FBA sellers inviting them to bid a per-unit price to get their products into same-day delivery. That's delivery in two to five hours. You only pay for the units that actually ship that way. Amazon claims products delivered sub same day sell 12% more than standard FBA in the areas where it runs, across 2,300 US metro areas.

The same-day delivery bidding is a floodgate to watch. It will penalise categories with thin margins and anything that spikes at one time of year, because you're bidding against everyone else who wants the same window. And with free delivery already baked in for millions of Prime members, is bidding really optional if your business leans on Amazon? Unit economics are going to get out of whack.

As an aside, great news for the team at Mateship, who have signed a deal with Australia Post to have their social shopping app deliveries go to parcel lockers around the country. A great example of patience, continually showing up… and perfect goddamn smiles.

03

Big news in the US on the war on TikTok advancement of social shopping. Amazon has joined the YouTube Shopping affiliate program, so eligible creators can tag Amazon products in Shorts, long-form and Live, and earn on them. Of course, it was announced on Shorts, because we know influencers don't watch horizontal video.

A bit of the devil in the details. Amazon supplies a curated catalogue, so you can't just tag your favourite back massager. And creators only get daily totals for revenue, clicks and sales, with no per-video or per-product breakdown. Don't be silly, Amazon's not giving you that data. It's US only for now, which is the frustrating part, because it blows my mind that social affiliate still isn't a core shopping channel in Australia. It's one of the biggest differences in ecommerce between the two countries.

While we're here, a new AiMCO report says 84% of organisations are investing in creator marketing but only a quarter have a dedicated budget for it. Sounds right. And 80% said they use creator marketing for brand building with only 51% for performance and sales. That bit sounds like a fib.

04

Shop Pay Installments has been running overseas since 2021, and last week we finally got it here. It's underwritten by Affirm, who exited the Australian market back in 2023. Shoppers split a purchase into fortnightly or monthly payments, with Affirm underwriting every transaction individually rather than approving everyone.

But this isn’t interest-free as we have been trained with BNPL. Affirm offers interest-free or interest-bearing plans depending on the approval, shown up front before the customer commits, with no late fees. So your customers may be taking on a small loan at your checkout, not a pay-in-four. I reckon that'll catch a few out.

05

I've got one story slot left and two half stories that don't relate to each other at all, so I'm not even going to try.

Firstly, Amazon has launched "Update me when", which lets you ask Alexa to watch Amazon and the wider web and tell you when something meets a condition. Their examples lean towards new releases and drops, a new Kindle, a new season landing, rather than price and stock, which still sit under separate alerts. I think it's pretty cool, and I wonder whether "update me when" becomes new ecommerce language.

And in terms of language, National Tiles' Frank Walker has moved from radio to TV with its annoyingly good bad ad. The platform is literally called Annoyingly Good, and the ads have Frank crashing through other brands' commercials to remind you you're watching a National Tiles ad. Update me when his voice can no longer be heard.

ATC PLAYBOOK #659

There is no formula for ordering the right amount of stock. It’s always one of the biggest risks for an ecommerce business - especially for a first run.

Ordering the right amount of product is one of the most underrated skills in ecommerce, and it can make or break a business. There is history you can look at and a forecast you can make, but nobody tells you how much stock you will be stuck with or how quickly you will sell out.

What costs you is not only the write-down at the end. It is the cash tied up in stock that is going nowhere, instead of going into stock that could be moving. And if you are running ecommerce, that tends to become your problem, because you are the one who has to shift it. There are really only three levers: how much you commit up front, how fast you find out whether it is working, and what you do with the answer.

Today's Playbook is anchored by Bronte Campbell, Co-founder and CEO at Earthletica, who built the brand while she was still swimming for Australia. Four Olympics, three gold medals, which is not the standard route into activewear. What got my attention is that half her range showing as sold out is not a supply problem, it was all part of the plan.

Examples from this Playbook

🏊 Bronte from Earthletica treats the first run as the price of finding out, not a commitment to the product. Limited drops, a wait list, inventory kept deliberately small. A 50% repeat rate over six months told her the product was right before she scaled the buy.

🌱 Bec from Seed & Sprout went back to long-standing suppliers and asked for a smaller first run, putting products up on pre-order to read the SKU mix before committing to volume. The minimum order quantity is more negotiable than it looks.

⚡ Jason from LSKD runs hype drops as small as 100 units. Read the speed of the sell-out, not just whether it sells out. Two drops of the same size, one clearing in nine days and one in eleven weeks, are telling you completely different things.

🧴 Adam from Who Is Elijah is a reminder to go in wanting the data to surprise you. Four years of it showed the Discovery Set customer had the worst lifetime value, on the product the whole business treated as the hero. They stopped pushing it and gross profit lifted.