51,123 founders and marketers are getting this newsletter today.
Welcome to the 221 new operators who joined us this week! 🤯

Hey Fastlane Insiders! 👋

Something's been coming up in my inbox a lot lately.

Founders telling me their Direct traffic is climbing. No brand campaign, no podcast hits, flat ad spend. And they're not sure whether to feel good about it.

I get why. On paper it looks like free growth. But I came across a study this week that explains what is actually going on, and it is not good news dressed as a mystery. An AI traffic detection company called Loamly ran its own tracking across 446,405 visits and found that roughly 70 percent of AI-referred traffic arrives with no referrer header at all. Google Analytics has nowhere to file it, so it lands in Direct.

That lines up with everything else happening right now. ChatGPT referrals are up about 5.5x since January. Shopify says AI-attributed orders climbed 11x through 2025. All of it flowing into the one bucket you have been trained to ignore.

Here's the part that actually matters though. It is not that the traffic is mislabeled. It is what that traffic does once it hits your site. More on that below, because it changes the math on your Q4 budget.

This week's edition also covers whether your team is big enough to need an AI coworker yet, which most teams under eight people are not, plus a conversation with OptiMonk on why fixing conversion beats buying more traffic. Whether you are auditing your GA4 setup, your app stack, or the traffic you are already paying for, the fixes below are worth your Thursday afternoon.

Here's what's inside:

🎧 Podcast: How OptiMonk uses AI to turn paid traffic you already bought into more sales.

💡 Knowledge Drops: How to tell if your team is big enough to need an AI coworker, plus where your best-converting traffic is hiding.

🔥 Tool of the Week: Katana gives you real-time stock visibility across every channel.

📡 Industry Pulse: Shopify's Q2 numbers, AI agents that can now check out on your storefront, and more

Let's dive in. 👇

Every merchant running Meta or TikTok ads knows the story: you pay for the tap, the shopper pokes around the in-app browser, and vanishes — no email, no phone number, no way back.

Klaviyo and Postscript can't fix this. Email and SMS can only win back the few who reach checkout and hand over contact info. That leaves the other 90%+ — the shoppers who browse and bounce — completely out of reach.

StorePush is built for them. On iPhone, your ad opens your store as an Apple App Clip — nothing to download, no signup, shoppers just tap and they're in. If they leave without buying, a push notification on their lock screen brings them back to where they left off. And you see the whole path, tap to purchase.

Here is the math most Shopify founders quietly avoid: you pay Meta, Google, and TikTok more every quarter to send people to your store, and roughly 97 to 98 out of every 100 still leave without buying. The default reaction is to spend more on acquisition. The cheaper move is fixing what happens after the click.

This week, I sat down with Krisztian Kiraly, International Partnership Manager at OptiMonk, a bootstrapped Hungarian company founded in 2014 that now powers more than 30,000 websites across 150+ countries. Before OptiMonk, Krisztian ran his own CRO agency for five years and once drove a 70% revenue lift for a large Hungarian ecommerce store without adding a dollar to the ad budget.

Krisztian closes with one line about buckets and leaks that reframes the whole acquisition conversation. It is the last two minutes of the episode.

A founder doing $2M on Shopify sent me a pricing page for an AI coworker platform last month with three words attached: is this real? He'd already sat through two demos. The product connected to Shopify, Klaviyo, Meta, and Stripe, drafted purchase orders, and answered questions in Slack.

My first question was not about the product. It was about his team: six people, two of them part-time, one Slack workspace. That answer decided the purchase, and it decided against it. Harnesses, the system prompt, tools, memory, and permissions wrapped around a model, are real and increasingly powerful. But their value scales with how many people need the same context and cannot get it from each other, and at six people in one Slack workspace, that number is close to zero.

If you are between $500K and $2M, the highest leverage move this quarter is not a purchase at all.

Run this check before you read anything else. Is your Direct traffic growing faster than your branded search, your paid spend, and your email list? If it is, something is sending you visitors your analytics cannot name, and in 2026 the most likely candidate is an AI assistant.

Here is the mechanism. Roughly 70% of AI-referred sessions arrive with no referrer header, so GA4 has no choice but to file them as Direct. Three things cause it: shoppers copy and paste a URL out of a ChatGPT answer instead of clicking, mobile AI apps strip referrer data in their internal browsers, and every platform handles referrers differently. None of it is fixable on your end. All of it is measurable.

Here's what's working:

Sell on Shopify, Amazon, wholesale, and retail with cross-channel, real-time stock visibility. With Katana, you get the right numbers every time, across sales, orders, and production. All the features you need without the ERP price tag or complexity. Start today and go live in as little as a week.

AI agents can now drive your storefront from inside the shopper's browser — On August 5, Shopify switched on WebMCP tools across every Liquid storefront and the Hydrogen developer preview, letting agents search your catalog, update the cart, and start checkout in the shopper's own tab, with nothing to install or configure. Unlike Agentic Storefronts, where the agent transacts through an API and never loads your pages, this runs on the live session, so your theme, your cart drawer, and your speed work all still apply, which makes it the first agent channel your existing CRO investment actually reaches.

Shopify posted its fifth straight 30% quarter — Q2 results released August 5 showed GMV up 32% to $115.6 billion, revenue up 34% to $3.58 billion, and free cash flow margin widening to 18% from 16% a year ago, with Q3 guidance calling for revenue growth in the low thirties. If you are building toward an exit or just betting your stack on this platform for another five years, the operating leverage matters more than the growth rate, because it is the part that says the company can fund the roadmap without needing you to absorb price increases.

Campaign Autopilot can now buy Microsoft ads for you — On August 4, Shopify added Microsoft Advertising to Campaign Autopilot, which will create or connect a Microsoft ads account, run Performance Max campaigns alongside your other channels, and reallocate budget across all of them automatically based on performance. Microsoft's inventory spans Bing, Edge, Outlook, and Copilot, so a Shopify-native tool is now placing your budget in front of an AI assistant's users, and it is in early access through the Growth tab rather than a general rollout.

Shopify quietly lowered the bar for fulfillment app badges — On August 1, Shopify relaxed three Built for Shopify requirements for fulfillment services apps, extending the response windows for fulfillment and cancellation requests and reducing the required completion rate on assigned fulfillment orders. Merchants treat that badge as a shortcut for vetting a 3PL app, so a badge that appears in your app store search next month may reflect a standard that moved rather than an operator that improved.

Everything in this week's edition keeps landing on the same idea from different angles. What you already have is usually worth more than what you are being pitched.

I know what time of year it is. The BFCM emails have started, the countdown widgets are back, and the pressure to add one more thing to your stack before Q4 is about to get loud. If you are feeling behind already, I want to say clearly that you probably are not. Most of what gets sold in August is a solution to a problem you have not confirmed you have.

So here is my permission slip for the month. Spend August on the boring stuff. Look at what your Direct bucket is actually made of. Try the AI tools already sitting inside your Shopify plan before you pay for another one. Fix the leaks before you pour in more water. None of that is exciting, and all of it will make your Q4 easier.

If you only have time for one thing, the GA4 diagnostics from the second knowledge drop are the highest return ninety minutes you will spend this month. Do it now, and you will have thirty days of real data before your first Q4 budget call. That is the whole reason this is landing in your inbox in early August instead of October.

And then tell me what you find. Hit reply, even if it is one line, even if the answer is "nothing weird here." I read every single one, and honestly they are the best part of my week. They shape what shows up here more than anything else does.

Summer is short up here in BC, and I am trying to be outside for as much of it as I can. Hope you are getting some of that too.

P.S. Missed a previous edition? Browse the archive. Several past issues walk through stack audits like this one.

Cheers!
Steve