Watch what happened to one of our clients’ SEO channel in August 2026: $84,123.17 in revenue, 89 orders, and a customer arriving via a ChatGPT answer instead of a Google search.
That last part illuminates something most SEO reporting still misses. AI referral isn’t a future trend to plan for. It’s already converting, at a $597 average order value.
Google did the heavy lifting too, generating $67,103.09 for the month, up 15.3% on the month before. The headline total looks like it dropped 8.8%, but that’s a reporting quirk, not a real decline. Once you strip out a one-off order that inflated the previous month, the underlying channel mix is stable and improving.
Below, we break down exactly where this revenue came from, why the dip isn’t what it looks like, and what the ChatGPT numbers tell you about where ecommerce SEO is headed next.
The Problem: Ranking on Google While Invisible to AI Search Engines
One of our ecommerce clients came to us with what looked like a healthy SEO setup. Their collection pages ranked for a range of competitive terms in their category. Google traffic was steady and growing.
On paper, there was nothing to fix.
But here’s the thing about SEO in 2026: ranking well on Google isn’t the whole game anymore. When we dug into how the store’s product pages were built, we found the descriptions, specs, and reviews only loaded after JavaScript ran in the browser. Fine for a human shopper. Their browser handles it without a hitch.
AI crawlers don’t work that way. Tools like GPTBot and Perplexity’s crawler often only read the raw HTML a server sends, before any JavaScript fires. That meant the store’s most detailed product content, the stuff that answers a shopper’s question, was invisible to the exact tools millions of people now use to research purchases. And this was happening even though the store’s robots.txt file said those crawlers were welcome in.
Then a new Shopify theme went live in August, which added a second wrinkle. Theme migrations can quietly break internal links or slow down re-crawling, and Bing tends to be slower than Google to catch back up after that kind of change.
So the store had a good foundation and two invisible cracks in it. Neither would show up if you only checked rankings.
The takeaway: A store can look healthy by every metric you’re used to checking and still be leaking revenue somewhere a standard SEO audit doesn’t reach. Being invisible to AI search is one of those places.
Why this matters: Before fixing anything, we needed a way to measure what was happening, not just on Google, but everywhere customers now search.
The Evaluation: Deciding What “SEO Revenue” Means
Before touching anything, we had to answer three questions. Skip this step and you’re just guessing.
- What counts as SEO traffic right now? Google organic is the obvious answer. But Bing, DuckDuckGo, Google Discover, and AI referrals from ChatGPT and Perplexity all send visitors who found the store through indexed content. If you lump everything that isn’t paid into one bucket labeled “organic,” you miss exactly the shift that matters most this year: AI tools starting to send paying customers.
- How do you tell a real trend from a fluke? Ecommerce revenue is noisy. One big order from an odd channel can make a month look great or terrible for reasons that have nothing to do with strategy. Any measurement approach that doesn’t catch outliers is going to lead you to the wrong conclusion, every time.
- Can the content behind your rankings be read by the systems now sending you customers? This one’s easy to skip. You can rank #1 for a term and still lose the AI-referral game if a crawler can’t see your page’s content.
So instead of one lump “organic traffic” number, we tracked revenue by exact channel and session source, using Shopify’s Last Non-Direct Click attribution model throughout so credit always lands on the last real marketing touchpoint before an order, not just whatever channel happened to be open at checkout. We compared each channel month over month. And when a single unusual order showed up, we flagged it and explained it instead of letting it quietly distort the total.
Why this is the right call, not just a preference: grouping every non-paid channel together hides the exact signal you’re trying to catch. If AI referral had been buried inside “other organic traffic,” we never would have noticed it was already converting.
The takeaway: clean measurement comes before good decisions. You can’t fix what your reporting doesn’t show you.
The Solution & Execution
With clean measurement in place, the work split into three fixes.
Fix one: give AI crawlers something to read. We moved product descriptions, specs, and reviews into the initial HTML response instead of loading them after the fact with JavaScript. Simple change. It’s the difference between a crawler seeing the content and a crawler seeing a blank page.
Fix two: make the content worth citing. Access alone doesn’t get you cited. The store’s highest-traffic product pages, starting with laundry cabinets, its top organic query, got expanded FAQ sections and cleaner spec formatting. AI answer engines like ChatGPT and Perplexity extract short, self-contained passages instead of reading a full page top to bottom, so a clean FAQ block or spec table has a far better shot at being the passage they quote than a paragraph of marketing copy.
Fix three: recover from the theme migration. After the new Shopify theme launched, we audited Bing Webmaster Tools for crawl errors and broken internal links, then resubmitted the key collection pages so search engines could catch up faster. This directly targets the re-indexing lag that tends to hit Bing hardest after that kind of change.
One more thing worth calling out: Google Search Console showed 398,744 impressions and only 4,380 clicks that month, a 1.1% CTR at an average position of 15.9. A chunk of the opportunity wasn’t about ranking higher. It was about earning more clicks from the rankings the store already had, by sharpening titles and meta descriptions on high-impression pages.
This isn’t a fix you do once and walk away from. It’s measure, find the gap, fix it, check if the number moved, repeat.
The Results (Proof)
Here’s what that work produced in August: $84,123.17 in SEO-attributed revenue across 89 orders.
Google Search did most of the heavy lifting, generating $67,103.09. That’s up 15.3% from the month before, and now 80% of all SEO-claimed revenue.
Now, if you just looked at the total dollar figure month over month, you’d think something went wrong. It’s down 8.8%. But that number is misleading, and here’s why: the previous month included a one-off $12,277.63 Yahoo! order that didn’t repeat. Pull that outlier out, and the real story is Google growing steadily while everything else holds roughly steady too.
Bing revenue did genuinely decline, down 26.9% to $12,839.28, in the same window the new theme went live. But Bing sessions rose over that window. Orders fell anyway. That’s not a visibility problem, that’s a conversion problem: the theme migration didn’t stop Bing from sending traffic, it broke something in what that traffic did once it landed. Theme migrations often change checkout flow, page load speed, or button placement, and any one of those can quietly tank conversion while sessions keep arriving normally.
But the number we want to draw your attention to is this one: ChatGPT sent 514 sessions and generated $4,180.80 in revenue across 7 orders, at a $597.26 average order value. That’s not a hypothetical about where search is heading. That’s a real customer, finding a real product through an AI-generated answer, and buying it. And it wasn’t a one-off, either: ChatGPT generated $4,256.95 the month before, essentially flat against August’s number.
Put it all together: $84,123.17 in revenue. 89 orders. Google up 15.3%. AI referral already converting at a solid order value. A short-term Bing dip with a clear, explainable cause.
The Lesson: What This Means for Your Store
Don’t judge a month by its headline number alone. One unusual order can make a strong month look like a decline on paper. Always check what’s driving a change before you react to it.
Check whether your content is visible to AI crawlers, not just to Google. If your product pages load key content through JavaScript, tools like ChatGPT and Perplexity may never see it, even if your robots.txt says they’re welcome. Worth auditing regardless of how well you currently rank.
Track AI referral as its own channel. Bury it inside “other” or ignore it, and you’ll miss one of the fastest-growing sources of qualified buyers right now. Even a handful of AI-referred orders can carry a strong average order value.
Expect a dip after a theme change, and don’t assume it’s just a visibility problem. A site migration can slow down re-indexing on search engines other than Google, but it can just as easily break something in the experience itself, so traffic keeps arriving while conversion quietly drops. Audit both: crawl errors and re-indexing, and whether the new theme changed anything in the path from click to checkout.
High impressions with low clicks is usually a quick fix. If you’re showing up in search but not getting clicked, the answer is often sharper titles and meta descriptions on your existing pages, not new content from scratch.
What This Proves: SEO Now Pays Off in Places Google Doesn’t Reach
The headline number speaks for itself: $84,123.17 in SEO-attributed revenue in a single month, with Google Search growing 15.3% and carrying most of the load. That alone is a solid month for any ecommerce store.
But the real story here isn’t that SEO still works. It’s where it’s starting to work. A customer found this store through ChatGPT, not a search bar, and bought something. That’s not a prediction about the future of search. It already happened, in a normal month of trading.
The dip in the headline total wasn’t a real decline. It was a reporting quirk caused by an outlier order the month before. And the short-term Bing drop had a clear, explainable cause tied to a theme migration, not a breakdown in strategy. Once you understand those two things, what’s left is steady, broad-based growth across every channel that matters.
It’s also a reminder of what each channel returns. Paid spend runs at roughly a 1:4 ROI while it’s live, and stops the moment the budget does. Organic search compounds instead, typically clearing 1:10+ over 12 months, and it doesn’t switch off when you do, because indexed content keeps ranking and getting cited long after the work that built it is finished.
If you’re still measuring SEO success by Google rankings alone, this is the shift worth paying attention to. The stores that start measuring and adapting to AI-referred traffic now are the ones that’ll be positioned to capture more of it as it grows.
If your product pages aren’t set up to be read and cited by AI search engines, see how ecommerce AEO and GEO works. Or, if you’d rather talk through where your own store stands, get in touch with our team.
