Wednesday, September 16 · Issue #1 · 5 min read.
Good morning.
Three weeks from today, Amazon and Target both flip the switch on their October deal events, which means the holiday season is functionally starting in 21 days. If your inventory plan is still a Google Doc titled "holiday ideas," this is your nudge.
This week: AI became the first thing your customers talk to, ecommerce posted its best quarter in five years, FedEx started selling insurance against its own industry's chaos, and TikTok Shop got proof that it plays by completely different rules than Amazon.
Let's get into it.
Mastercard surveyed 26,000 parents and teenagers. Thirty-one percent of teens trust an AI product recommendation more than a friend's. Twenty-three percent trust it more than their parents. More than six in ten use AI to research products at least once a month, and 27% say they'd hand over the entire purchase to an autonomous shopping agent.
That's the next decade of customers. Adults aren't far behind. Bain found that 24% of holiday shoppers plan to start browsing using an AI tool this year, up 17 percentage points from last season.
Retailers noticed. In one week, Instacart launched an assistant called Clementine that builds a cart from a photo of your fridge or a conversation about dinner, and Shipt launched Ask Shipt, which does roughly the same thing and already uses ChatGPT and Claude to answer. Chewy, Lowe's, Target, and even a rug retailer named Rugs Direct all shipped assistants of their own.
The shopper who used to find your product page through a search result is increasingly seeing a summary of your product page, written by a machine that has never seen your photography, your founder story, or your beautifully art-directed hero video. What it did see was your title, specs, materials list, sizing table, and structured data.
Product data is a merchandising job now. The brands that get recommended are the ones whose product pages are boring, complete, and unambiguous. Sexy copy sells to humans. Clean data sells to the robot that talks to humans.
Practical Ecommerce made the other half of this argument the same week. Your own site is the one place where the relationship isn't mediated by somebody else's algorithm. Feed the machines, but don't move in with them.
U.S. online sales grew 12.4% in Q2. That's the first double-digit quarter in five years, and it's more than twice the pace of retail overall.
We spent two years being told ecommerce growth had "normalized." Apparently not.
The holiday outlook is calmer. Deloitte expects online holiday sales to grow between 7.5% and 8.4%, ahead of total retail's 4% to 4.8%. The channel keeps taking share, just not at Q2's pace.
Deloitte's footnote: shoppers are still hunting for value, and they're switching between brands and retailers to find it. A growing channel full of switchers is a very different business than a growing channel full of loyalists. Your traffic can go up, and your month can still feel harder because you're renting customers instead of keeping them.
If you only look at one number between now and January, make it the repeat purchase rate. Everything else this season will be noisy.
FedEx launched a Shopify app that shows international shoppers a guaranteed landed cost at checkout, including duties and taxes. If the actual bill comes in higher than the amount the customer saw, FedEx covers the difference. It's $99 a month, Shopify only for now, and it landed the same week as a broader set of cross-border trade tools.
FedEx's own research says 68% of small and mid-sized businesses have customers who are hit with surprise duties or fees at the doorstep at least some of the time. Every one of those is a refund request, a support ticket, a one-star review, or all three.
Two things before you sign up. "Guaranteed landed cost" is a conversion product, not a shipping product. The friction it removes is at checkout, where an unknown number scares people off, and at delivery, where a known number shows up three weeks late and ruins the experience. Price it against your international abandonment rate, not your shipping spend.
The other is the trade. To use it, your cross-border shipping runs through FedEx. For a brand doing real international volume, $99 is nothing. For a brand doing eleven orders a month to Canada, it's a subscription to a problem you could solve by publishing a clear duties note on your shipping page.
Marketplace Pulse ran the seller lists for both platforms side by side and found that almost nobody wins on both. Of the top 10,000 sellers on each, roughly 5% overlap.
The overlap isn't symmetrical. One in five of TikTok Shop's hundred biggest U.S. sellers is also a top-10,000 seller on Amazon. Going the other direction, only one in twenty-five of Amazon's hundred biggest cracks TikTok Shop's top 10,000.
Winning on TikTok travels. Winning on Amazon doesn't.
TikTok manufactures desire. Amazon captures it. A product that goes viral on TikTok creates demand that people then satisfy wherever they already shop, usually Amazon. A product that dominates Amazon search has no equivalent way to generate demand elsewhere.
These are two different skill sets, not two storefronts for the same skill set. TikTok rewards products that demo well on camera and founders willing to build creator relationships. Amazon rewards keyword discipline, review volume, and fulfillment you don't have to think about. If your product's greatest strength is that it lasts 11 years, that's an Amazon asset. It is not a fifteen-second video.
And even among the 498 sellers who rank on both, how well you do on one tells you basically nothing about how you do on the other.
The holiday calendar just got set.
Amazon's Big Deal Days and Target's Circle Deal Days both land on October 6 and 7, with Target Circle 360 members getting early access on the 5th. Whether you participate or not, your customers' price expectations are reset that week.
Phone lines are having a moment.
Sproos, Made In, and Trade Coffee are all staffing their phone lines and treating them as a selling point. A SurveyMonkey study found 79% of Americans would rather talk to a person, and 41% say service has gotten worse since the bots showed up.
Somebody finally wrote down the AI creative policy.
Boll & Branch published where it will and won't use generative AI in advertising, landing between Coca-Cola's full embrace and REI's public retreat. Steal their framework if you haven't had this argument internally yet.
Reformation's first public earnings were good.
Revenue up 24%, active customers up 23%, and a plan to double its 70-store fleet in five years. Notable for a brand born online: wholesale grew roughly twice as fast as DTC.
Retail media keeps getting fatter.
Kroger's ad business posted its best profit growth since 2021, up 24%, and executives talked about it before groceries. If you sell to retailers, more of your trade budget is about to be pulled toward their ad platforms.
Stablecoins crept into B2B payments.
Circle is acquiring Tazapay to push USDC settlement into cross-border business payments. Not something to act on this week, but the plumbing for paying overseas suppliers is being rebuilt.
A good pre-Q4 read.
Practical Ecommerce on using sales velocity to spot stockouts and markdowns before they happen. Twenty minutes before you lock your Q4 buy.
That's the week.
If you do only one thing after reading this, go look at your ten best-selling product pages and ask whether an AI could accurately describe your product using only what's written on those pages. Most brands fail that test in terms of sizing and materials.
And if you need help finding a better or new 3PL, we’d love to help you at FulfillYN, the #1 3PL matchmaking service.
See you next Wednesday.