Your Weekly DTC Industry Roundup
Your ROAS Is Probably Misleading You
Your measurement system is probably misleading you.
And the distance between what your dashboards report and what's actually driving revenue is where most brands waste their next dollar.
We're getting into some really interesting stuff this week.
Here's what everyone is buzzing about this week:
- We just our full incrementality methodology with 146 geo holdout tests, 80 stores, and channel benchmarks that should change how you allocate spend
- Flexport launches AI agents for tariff refund prep after the Supreme Court IEEPA ruling
- ChatGPT traffic converts 31% higher than non-branded organic across 94 ecommerce sites
- OpenClaw and the AI agent wave are reshaping how software goes from idea to production
Let's get into it …
Limited Time Opportunity
Get Your Brand in Front of 40,000+ DTC Operators
Each week, this newsletter reaches 40,000+ inboxes, including operators behind brands like Liquid Death, Solo Stove, Barstool Sports, APL, and Native.
Already subscribed?
Login to unlock.
Stay in The Thread
The Thread is a weekly DTC newsletter from Common Thread Collective. Subscribe to read the full post and unlock all future issues.
Your Weekly DTC Industry Roundup
Your ROAS Is Probably Misleading You
Your measurement system is probably misleading you.
And the distance between what your dashboards report and what's actually driving revenue is where most brands waste their next dollar.
We're getting into some really interesting stuff this week.
Here's what everyone is buzzing about this week:
- We just our full incrementality methodology with 146 geo holdout tests, 80 stores, and channel benchmarks that should change how you allocate spend
- Flexport launches AI agents for tariff refund prep after the Supreme Court IEEPA ruling
- ChatGPT traffic converts 31% higher than non-branded organic across 94 ecommerce sites
- OpenClaw and the AI agent wave are reshaping how software goes from idea to production
Let's get into it …
Limited Time Opportunity
Get Your Brand in Front of 40,000+ DTC Operators
Each week, this newsletter reaches 40,000+ inboxes, including operators behind brands like Liquid Death, Solo Stove, Barstool Sports, APL, and Native.
The audience skews senior, so many readers are the people who actually approve tools and vendors (CEOs, CMOs, CFOs, and cofounders).
That reach extends beyond the newsletter, too. We reach 187,000+ active audience members across the Ecommerce Playbook Podcast, YouTube, and Taylor Holiday's channels.
If you sell to scaling ecommerce brands, this is a direct way to get in front of decision-makers. Sponsors like TaxCloud, Omnisend, AppLovin, and Fermat have already done it, and we only feature one sponsor per edition.
To get started, schedule time with Tayla below and she'll walk you through the details.
Book a call with Tayla
Measurement & Methodology
Google Says Your Branded Search ROAS Is High but Our Data Says It's Actually Low
We just published our full measurement methodology.
The system we run across 80+ brands to figure out what's actually incremental versus what the ad platforms want you to believe.
146 geo holdout tests of 80 stores.
The channel benchmarks are rough reading if you've been trusting your dashboards.
Facebook Acquisition has a median incrementality factor of 1.14 across 45 stores, meaning the platform actually underreports by about 15%. Google Branded Search? 0.27 across 39 stores.
That's a 4x overstatement. TikTok landed at 0.50 across 15 stores. Pinterest and Snapchat both at 0.30.
We call this approach "progressive truth." No single test gives you the answer, but each one reduces the error rate of the whole model.
We operationalize the factors through Statlas and calibrate cost controls in real-time.
If you haven't run a geo holdout test, you have no idea what you're missing out on.
Read the full methodology
Supply Chain & Tariffs
How Flexport Is Using AI Agents to Prepare for Tariff Refunds Before Anyone Else
U.S. tariff policy changed every week and a half in 2025.
The customs industry is calling 2026 the "Year of the Audit." Filing errors are at an all-time high and enforcement is ramping.
Then the Supreme Court ruled IEEPA can't be used to impose tariffs. Refunds on duties already paid entered a gray zone with no clear pathway.
Flexport just shipped a fleet of AI agents for exactly this chaos. One audits previous customs entries for compliance errors. Another handles container optimization.
A third does real-time language translation for cross-border logistics.
They say they've cut their own filing error rate to 0.2% with the audit agent. 10x better than any other broker, by their estimate.
Whether refunds materialize or not, the brands with clean auditable records will be in a fundamentally different position than those trying to reconstruct transactions that were never built to be reversed.
Get your filings audited now. Especially if you import product.
Learn more here
AI & Ecommerce
94 DTC Brands Exposed a 31% Conversion Gap Between ChatGPT Traffic and Organic Search
Visibility Labs pulled 12 months of GA4 data across 94 seven and eight-figure ecommerce sites.
ChatGPT referral traffic converted at 1.81% versus 1.39% for non-branded organic. 31% higher.
Volume is still small. $474K in ChatGPT-attributed revenue against $32.1M from organic. Roughly 1.5%.
But the growth rate is the thing to watch. ChatGPT visits went from 1,544 in January to 18,202 in December.
That's 1,079% growth.
Organic grew 17%.
The conversion edge probably comes down to intent compression.
People refine what they want inside ChatGPT before they click. By the time they hit a product page, the comparison shopping already happened.
Two caveats. AOV was 14.3% lower on ChatGPT traffic ($204 vs $238). And GA4 is almost certainly undercounting. People get a recommendation from ChatGPT, search the brand on Google, and buy there.
That sale gets credited to branded organic.
Post-purchase surveys are the only reliable way to measure AI-influenced revenue right now.
Read the full details
AI Operations
OpenClaw Got Acquired by OpenAI in 2 Months and Exposed the Biggest Gap in AI Software Development
OpenClaw went from open-source project to 140K+ GitHub stars to OpenAI acquisition in about two months.
An AI agent that runs locally, connects through your messaging apps, and actually executes tasks without waiting for a prompt.
That's the fun part.
The less fun part is what everyone discovered after the first wave of AI coding tools.
You could describe an app and get something functional fast. Then you'd find the missing error handling, the security holes, the placeholder functions nobody filled in. Starting got cheap. Finishing stayed expensive.
Human in the Loop, from Jon Butler, is going after that gap. One human paired with 19 specialized AI agents across the full build lifecycle.
It builds scored across security, performance, completeness, and design.
If it doesn't pass? It loops until it does.
Every DTC operator knows the distance between a tool that demos well and a tool that actually works at scale.
Read the full story
TLDR
What We Covered
Read the measurement piece twice.
Most brands are allocating budget based on numbers their ad platforms are incentivized to inflate.
Google Branded Search overreports by roughly 4x. Meta Acquisition underreports by about 15%.
TikTok's median incrementality is half what the platform claims.
If you haven't run a geo holdout test yet, our aggregate benchmarks are at least a better starting point than whatever your dashboard says.
Tariff situation is moving fast. Brands with clean customs records will have options others won't.
ChatGPT as a commerce channel is small and real and growing at a rate you can't wave off.
Set up post-purchase surveys now so you're not guessing in six months.
And the AI agent wave hitting software is going to reach DTC operations sooner than most people think.
Same pattern across every story this week.
Automation gets cheap, reliability stays expensive, and the operators who build for finishing are the ones who compound.