50,779 founders and marketers are getting this newsletter today.
Welcome to the 1627 new operators who joined us this week! 🤯
Hey Fastlane Insiders! 👋
If you're building your Black Friday promo off last year's landed cost, I've got some bad news. That number moved. Five different times, actually, over the course of this summer, and not one of them sent you an email about it.
Meanwhile, Walmart just got a $2.9 billion tariff refund and is spending it on shelf prices. Target and TJX got theirs too. So your costs went up while the price your customer thinks is normal went down. That's a rough combination to walk into December with.
Here's the thing I keep noticing, and it isn't really about tariffs or AI. It's stale numbers. Your landed cost expired in July. Your open rate is counting machines instead of people. The sort order on your best collection page was set the day you launched and hasn't been touched since. None of them look broken. That's exactly why they're expensive.
Doesn't matter if you're doing $50K months or $500K months. Same move: go check one number before you lock your Q4 calendar.
Here's what's inside:
🎧 The Podcast: The two rows of your collection page that decide whether your ad spend converts, and the triple cost of leaving them wrong.
💡 Knowledge Drops: What the MIT 95% AI failure stat actually counted (it isn't what either side quoting it thinks), and the five Q4 costs that moved on you since spring.
🔥 Tool of the Week: Redo Recover, reaching the 70% of abandoned carts your SMS flows never touch.
📡 Industry Pulse: Agentic buying, a Google Ads opt-out deadline that closes September 4, email open bloat, and the Shop app's holiday push.
Let's dive in. 👇
Your product's not the problem. Your listing photos are.
Upload your product once, choose a scene, and build the listing images around it. A clean hero shot, alternate angles, a lifestyle scene, and the rotating video for the gallery, all from the same photo and all consistent with your brand. Then do the same for the next SKU without starting over. The workflows are built by Kittl's own designers rather than generated on the fly, so you start from a concept that holds up instead of a blank prompt box, and you can edit any of it in Kittl's editor. 11 million people design in Kittl, rated 4.7 on Trustpilot.
25% off any monthly plan, up to 44% off annual with code FASTLANE.
🎧 New Podcast Episode! 🎧
The Hidden Revenue Leak Sitting on Your Shopify Collection Pages
You spend thousands driving high-intent traffic to a collection page. TikTok ads, SEO, email flows, all engineered to land someone in front of your products. Then that page greets them with a wall of sold-out badges and a top row sorted by a rule you set at launch and never touched again.
I sat down with Julia Wawrzynek, marketing manager at Mezereon, the team behind SortWise. She spends her days inside a problem most operators never think to audit: which products earn your prime digital shelf space, and which ones are quietly burning your conversion rate and your ad spend. SortWise ranked number one in the collection merchandising listicle I wrote earlier this year, which is why I wanted her on the show.
Here's what we unpacked:
95% of shoppers never scroll past the first page. Your top 10 to 20 product slots are the only real estate that matters, and Julia puts a cost on every one of them you leave wasted.
A sold-out best seller is a triple loss, not a missed sale. Two more costs stack up behind every grayed-out badge, and one of them keeps charging you long after the shopper is gone.
Sorting by popularity creates a doom loop. Your best seller stays on top because it's already on top, while your higher-margin products sit on page two where nobody sees them.
Revenue per view is the metric Julia calls her North Star. Traffic up and RPV down tells you exactly which page to fix first, and it's a number most merchants have never pulled.
The three-layer system: AI, business rules, and manual control. Most tools force you to pick one. Julia explains why leaning on a single layer leaves real money on the table, and the one situation where you have to override the algorithm yourself.
What changes between $50K and $500K a month. Same store, two genuinely different merchandising problems, and a stage-specific move for each one.
Julia also walks through an audit you can run tonight from your phone in under an hour, no dashboard required.
💡 Knowledge Drops of the Week 💡
The MIT 95% AI Failure Stat Is Real. It Doesn't Say What You Think.
In the last twelve months, I've had the same statistic quoted at me from two opposite directions. An agency founder used it to argue AI is mostly theater. Three weeks later, a vendor used it in a pitch deck to argue that merchants need a done-for-you agent build because 95% of those who try it themselves fail.
Both were citing the same document. Neither had read it.
The number is usually quoted accurately. What almost nobody carries across is what was being counted, and at $500K to $2M, that gap is the difference between two completely different budget decisions.
Here's what's working:
Read what the study actually measured. It counted one narrow category of AI tool against one narrow definition of success inside roughly six months. General-purpose tools like ChatGPT sat in a different column of the same exhibit, and that column tells a very different story.
Check who benefits from the conclusion. The report's recommended fix is agentic infrastructure, and the group that published it builds agentic infrastructure. That's printed in the document, not hidden.
Steal the budget finding instead of the headline. Roughly half of AI spend went to sales and marketing, while the documented savings clustered in back-office functions nobody was funding. I watched that exact pattern stall Shopify brands between $500K and $2M for years.
Run the 45-minute test. You're in the 95% if you can't state right now the specific number each AI tool in your stack was bought to move.
A larger university survey three months later reached an almost opposite finding. Two credible groups, adjacent questions, answers 70 points apart. Which one you believe changes what you fund next quarter.
Five Q4 Costs Moved. Your Model Probably Didn't.
Every input that sets your Q4 margin has changed since most merchants placed their holiday orders. That's not a forecasting problem. It's a stale spreadsheet problem, and it resolves in the most expensive month of the year.
Ocean freight is the loudest. Asia to US East Coast spot rates hit a new high of roughly $9,400 per forty-foot container in the Freightos update for August 18, West Coast climbed 9% in a week to around $7,400, and carriers have announced Panama Canal surcharges of $200 to $1,000 per container starting mid-September.
Here's what's working:
Reprice your shipping thresholds, not your products. On July 12 the USPS dimensional weight divisor dropped from 166 to 139 and fractional inches now round up instead of down, which raises billable weight on every dim-rated package by roughly 19%. That adds to the 8% temporary rate increase already in effect through January 17, 2027. If you ship anything bulky and light, your per-parcel math from last Q4 is fiction.
Find out whose name is on your customs entries. Only the importer of record can claim a tariff refund. If your forwarder or 3PL is listed on your CBP Form 7501, the refund on duties you economically paid is legally theirs. This catches smaller importers far more often than large ones.
Plan less discount depth, not more. Circana data reported August 17 showed July retail revenue down 1.0% with unit sales down 2.0%, and discretionary general merchandise dollars down 4.3% in the four weeks to August 1. Circana's read: broad discounts alone are generating less incremental demand.
Know that the anchor price moved without you. Walmart is spending a $2.9 billion tariff refund on shelf price and ran more than 11,000 rollbacks last quarter, up from 7,200. Target and TJX booked refunds too. Your cost base didn't change because of their refunds. Your customers' expectations did.
Bundles, thresholds, and gift-with-purchase protect contribution margin in a way a sitewide 30% never will.
🔥 Tool of the Week 🔥
Your SMS flows reach 30% of carts. Recover gets the other 70%.
Your abandoned cart flows only text shoppers who subscribed to SMS marketing. That's about 30% of your carts. The other 70% never opted in, so your flows never reach them.
Recover does. Trained agents text those shoppers one-to-one in real conversations. They answer sizing questions, handle shipping objections, send a discount code, and win the sale back. No automation, no templates, no blasts.
It runs alongside the flows you already have. Nothing to rebuild, nothing to rewrite, and five minutes to install on Shopify.
You only pay for results. 10x ROI guarantee.
⚡ This Week’s Industry Pulse ⚡
A few updates actually move the needle. Here's what made the cut…
Consumers are warming up to agentic AI purchases — 42% of American millennials will let an AI agent spend up to $250 on their behalf, but only when one specific condition is met. Take it away and support collapses by a quarter.
Google is about to reframe ads you already approved — Generative AI will auto-generate missing aspect ratios from video assets already sitting in your Performance Max campaigns, so creative you signed off and shipped gets recut without a second look. Logos, burned-in text and legal disclaimers near the frame edge are first at risk. The opt-out window closes September 4, right as Q4 creative goes live.
Your open rate is counting machines — A large share of reported opens aren't people, which quietly keeps dead subscribers inside your engaged segment and drags your sending reputation down with them. There's a one-segment fix you can build in Klaviyo this afternoon.
Shopify's Shop app goes front and center — Shop app GMV grew more than 70% in Q2, and Shopify is running its longest, most aggressive holiday campaign for the app ever. If you have never touched your Shop storefront, this is the quarter that starts costing you.
Until Next Thursday
Late August is my favorite stretch of the year in this business. Nothing is on fire yet. Q4 is still a spreadsheet instead of a scramble, and you can still change your mind about almost anything.
That window shuts faster than it feels like it will. By the second week of October, you're not really making decisions anymore; you're just running the ones you made back in August.
So if this edition nudges you toward anything, make it small. Pick one number you're planning Q4 on and go check whether it's still true. Half an hour, tops.
Then hit reply and tell me what you found. I read every one, and those replies are where most of what I write comes from.
P.S. If you're picking, start with the customs entries. That's the one where the money is already sitting somewhere and just isn't yours yet.
Cheers!
