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

I always love getting replies to these emails. They range from "you’ve got too much time" to "I’m leaving ecommerce to become a farmer”.

Last week, a Marketing Manager at a large retailer messaged me to say she’s just overwhelmed. Everything is moving so fast and she doesn’t know how she’ll keep up.

My intention with this newsletter isn’t to overwhelm you. And I hope it doesn’t feel that way.

I want to remind you that just by reading this and staying across these changes, you're already ahead of most. You don’t have to be trialling, using or mastering everything in here. We’re all making it up as we go.

CHEERS, BUSHY

01

Shopify’s Q2 results landed with GMV up 32%, revenue up 34%, and gross profit and free cash flow both north of 30%. Harley called it "a monster quarter" and the market agreed, with the share price up 17% on the day.

BigCommerce wasn’t going to let them have all the light, so they announced that checkout now loads 41% faster than it did in January. Their largest contentful paint went from 2.7 seconds in December, to 1.7 in January, to 1 second in July. Goooooood foooor yooooou. If I’m a BigCommerce client, that stat scares the hell out of me. It means 41% of speed was sitting there the whole time and nobody went looking for it until now.

They also launched Commerce Connect with WP Engine, which means WordPress is available in BigCommerce. Or BigCommerce is available in WordPress. I couldn’t work it out. WordPress keeps the front end, BigCommerce runs the commerce behind it. I was going to make a joke about local cake makers and model train stores finally getting into ecom, then remembered I still run the Add To Cart site on WordPress. I DON’T CARE. I STILL LIKE IT.

And yes, I’m a bit of a Shopify fanboy. And yes, they’ve backed Add To Cart since day one. And yes, they’re very good at tooting their own horn, especially when the numbers get announced. But we need competition. And when the pace of change is this fast, a checkout speed and WordPress announcement is not going to cut it. Expectations are moving fast and Commerce isn’t keeping up.

02

I used to beat myself up because I could never build or manage a pivot table in Excel. Constant source of shame. But just like tying knots and changing the oil in your car, you can now file pivot tables under skills only smug people need. This week ShipBob launched its MCP, an Anthropic-verified connector with 70+ read and write actions across your warehouses. Real-time visibility into your 3PL, just like the robots inside it have. You can reorder a SKU, reroute a shipment, fix stuck orders in bulk or upgrade a customer to a faster carrier. Australia is included from day one. The power is in your Claude.

This is prompt-based ecommerce management arriving quickly, and a small detail in Shopify’s quarter emphasises it. Daily active merchants using Sidekick grew 3.6 times year-on-year. There was no comparison stat for those ignoring our Sidekick friend. Pretty safe to say 3.6x growth isn’t normal. Sidekick might finally be getting employee of the month.

03

It’s a confusing time to identify as a theme. New data from Shop Theme Detector has Horizon running on 8.6% of stores scanned, up 1.2 points on Q1 and the biggest riser on the platform by a distance. Second among the climbers is Minimog OS 2.0 at 2.9%, then Refresh at 1.7%. Horizon is close to three times its nearest challenger and it has been out for barely a year. Dawn is still the most-installed theme overall at 15.2%, but it’s drifting down while Horizon runs up the middle. Horizon is the theme king in waiting.

It might also be the last of the big JSON themes. Shopify’s July Liquid preview introduces block and partial tags, so a template is composed in one readable Liquid file instead of scattering itself across JSON config. Shopify’s own words are that it lets "developers and coding agents read and edit everything in one place." Christopher Dodd brought this to my attention and explains it well: the move away from sections and into blocks, in a world where you won’t build the change; you’ll prompt it.

04

The battle for AI slop is on. Amazon has started limiting some shoppers to reading eight reviews per product because it suspects their accounts are bots. Suspect users need to email Amazon to get access to more reviews. Google killed its Earth AI feature one day after launch, having apparently only then considered that letting people paste generated imagery onto satellite maps might go badly. LinkedIn added a button to report AI slop. Just burn it to the ground. All legit cases.

But we all know AI isn’t the problem. People are, as demonstrated by the influencers who copped it for attending OpenAI’s first luxury retreat, a weekend of farm-to-table dining and beekeeping in upstate New York. I’d sit home and talk to Elon’s Grok before hanging with these gronks.

The one thing AI still can’t fake is being funny, which is why Snapchat has taken it upon itself to teach us millennials how to laugh and win with the next generation. It’s not all fluff: 86% of 13-28 year olds like it when brands attempt humour or self-deprecation, and 74% are more likely to buy from a brand that makes them laugh. Here’s your task for today. The very last page is a Zoomer Humour Playbook with four steps on how to be funny. Screenshot it and send it to the unfunniest person on your team with absolutely no context.

05

I enjoyed seeing this one. Sydney’s Hello Tailr has been doing alterations and repairs online for a while, and this week they launched a Shopify app that lets fashion merchants add a tick box for the perfect fit. If a customer opts in, they transact as normal and get their order delivered to their door, then hear from Hello Tailr straight after delivery asking for measurements and where to send the garment. It was only when I dug a little deeper that I realised Martin Cox from Native App Co, who we had on the show a few weeks ago, helped with the build. Nice to see him branching out from the menagerie of fish, storks, owls and pelicans.

It makes a lot of sense. Fit is the single biggest driver of fashion returns, and returns are the killer. Anything that converts a "close enough" size into a kept garment is worth a look, even at this step count. No doubt the end goal is routing goods straight to tailoring and then on to the customer, cutting that middle step out entirely.

And in sustainability, there are apparently no limits, which Oxfam proved by launching a rental service for donated second-hand fashion. Fourteen days from £22, with a Balenciaga dress and a Stella McCartney jacket in the opening range. These garments have more lives than my cat. And he’s an idiot.

ATC PLAYBOOK #651

A replatform is building a whole new house while the old one keeps trading, hitting its numbers and paying the bills. The platform is usually the easy bit.

Every few years, the same job lands on someone’s desk: move the whole business onto a new platform without pausing trade. You’re still running the sales, still shipping the product, still holding the numbers up, all while shifting the entire thing across to a new site. No project plan ever makes it look as hard as it is.

I’ve lived this one. Years ago, we replatformed off Salesforce across multiple brands, and we got through it by taking the whole team, agency partners, and all, out to an unused warehouse shed in the back of Brisbane and just getting through it together in one room. You might not need the shed anymore, but the principle hasn’t changed. A transformation this size needs its own space and its own people. You can’t squeeze it around the day job of trading.

That’s why I loved this chat with Craig Mildenhall. His team moved Kathmandu off Magento onto Shopify Plus in four months, then went live in the middle of a flash sale, something their agency reckons they’d never seen done before. The lesson underneath it is that a replatform is a people problem more than a technology one, and the teams that get through cleanly are the ones who plan for that from day one.

Examples from this Playbook

⛰️ Craig from Kathmandu split his team in two: a Shopify crew heads down on the rebuild, and a trade crew keeps the business selling. He held a hard go-live date, stayed close to the MVP, and sat his agency in the office for weeks so problems got solved in the room, not in a ticket queue. Four months, live mid-flash-sale.

👜 Lou from Louenhide didn’t split her team; she resourced up. She brought in a dedicated external project manager she’d worked with before, and the payoff was two in-house people who came out the other side knowing exactly how the thing was built. Don’t be shy about paying for help when the go-live window is this critical.

🧴 Kira from The Body Shop replatformed the whole Australian business in about 90 days, which forced her to be brutal about what makes launch day and what waits. Pick the pain you’re willing to accept, hold your date, and let the rest be a little bumpy. It has to be better than today, not perfect.

🧵 Silvan from Ghanda makes the case for a partner who does more than build what you ask for. The good ones tell you the thing you didn’t know to ask, because they build sites every day and you don’t. Let the ticket be the record of the decision, not the way you have the conversation.