#BlinkTank: Shopify's AI is writing specs for your products
As a reminder, last week we covered Shopify making bot filtering the default in Analytics. If your conversion rate looks wildly different week on week, that is why.
This week's lead might seem a little niche - Shopify's AI is writing product specifications on merchants' behalf, and some of them are wrong. What it goes to show is that a product page is no longer something only a customer reads. It is now raw material for Shopify's catalogue, for Google, and for whatever AI agent is shopping on someone's behalf.
Also this week: I'm on a live panel about product discovery at 4pm UK time today, Shopify has brought category metafields into Analytics, Search Console now reports on image search traffic, and Google's fourth spam update of the year is rolling out.
On to the news!
Shopify's AI is writing specifications for your products
AI discovery is still very new, and much of it is being built in public. Shopify has moved quickly to get every store into its Global Catalog so that customers can purchase in ChatGPT, Copilot and the rest, but this does mean there are some rough edges. We keep running into them in client work, and this one is a good example.
Last week Bijan Vaez at Merchkit posted an example of the Global Catalog publishing a chair's seat height as "18.9 lbs" when the page said 18 inches. The catalogue adds its own AI-written specifications, features and selling points to every product, and Shopify's own documentation says they "might vary in accuracy". However, this isn't made clear to any shopper reading them.
We ran the same check across 33 of our clients' stores and found 14,036 specifications that are being generated automatically. Most are fine when the information on page is clear, but the failures are interesting - a supplement dose of 6826 mg published as 6.826 mg, a furniture dimension of Ø100 × 75 cm turned into ø100.75 cm.
In about a third of the cases we checked by hand, the page was unambiguous and Shopify still published something incorrect.
In all likelihood, this is a small bug that Shopify are going to fix. As a general hygiene rule though, it looks like it's going to be worth being explicit about measurements, particularly when things come in multiple sizes or servings. When building specification tables, this is something to take into account.
🔗 Inferred product fields in the Global Catalog (Shopify developer docs)
🔗 Our findings across 33 stores (Sam Wright on LinkedIn)
Join me live at 4pm today on product discovery
If you're reading this on Wednesday morning, I'm speaking on a live panel this afternoon: Product Discovery in the AI Age: What's the Playbook? It's today, 30th September, at 4pm London time (11am New York), hosted by Chloë Thomas as part of the eCommerce Explored series.
I'll be alongside Ashu Dubey from Alhena AI and James Wall from Native Commerce, talking about how customers find the right product on and off your site now that AI assistants are part of the journey. I suspect product data will come up more than once.
It's free, the sessions are pitch-free (promise), and a replay goes out to everyone who registers.
🔗 Book your free place for today's panel
Category metafields are now available in Shopify Analytics
On 28 September Shopify made category metafields available as dimensions and filters in Analytics. If you've already categorised your products using Shopify's taxonomy, attributes like colour, material and fit now show up in reports without any extra setup. A companion update lets you group and filter by metaobject fields that are referenced from product, variant, customer or order metafields.
For large-catalogue stores this is extremely useful. Being able to see sales and conversion by material or fit, rather than just by product or collection, can be genuinely insightful analysis.
It also gives one more commercial reason to get category assignment and category metafields right, since the same attributes feed Shopify Catalog and, by extension, the AI surfaces mentioned above.
🔗 Category metafields are now available in Analytics (Shopify Changelog)
🔗 Use metaobject fields in Analytics reports (Shopify Changelog)
Search Console now reports on image-led searches
Google has added a new search type filter to the Search Console performance reports, rolling out globally from 24 September.
This focuses on searches where an image was part of the query, including Google Lens, Circle to Search on Android, image uploads to Google Search and Chrome's "Search this image" option. The filter appears in both the standard Search results report and the Generative AI features report.
The main limitation is that there is no query data, so you can see which pages image-led searches find, but not what people were searching for.
For stores selling visual or physical products (furniture, fashion, homeware, parts) this might be particularly useful.
🔗 Announcing web multimodal Search performance reporting in Search Console (Google Search Central)
Google's September 2026 spam update is rolling out
Google began rolling out the September 2026 spam update on 24 September, applying globally and to all languages, with a rollout window of up to two weeks. It's the fourth confirmed spam update this year after March, June and August, and the longest rollout window of the four.
Google hasn't said what this one targets. Most stores we work with don't need to do anything, but if you see movement between now and early October it's worth annotating the date before anyone starts attributing it to recent site changes.
🔗 September 2026 spam update (Google Search Status Dashboard)
Jobs
Hiring:
GLOWiD are looking for a Head of E-commerce in Stockholm. This is an interesting brand known for bringing K-beauty to the Nordics, and the role covers the full digital commerce P&L including CRO, CRM and paid acquisition.
Studio MEM are hiring a Project Manager to run Shopify builds, ERP integrations and digital marketing projects for luxury and fashion brands. I know the team here well and can very much vouch for them.
If you are hiring and the role fits someone with large-catalogue Shopify experience, or you are looking and have that background, send me a message.
Partner shout-out: Agent Checker
I've mentioned Agent Checker here before, and given this week's lead, it is worth doing again. If you want to know how your store actually behaves when an agent rather than a person is doing the shopping, this is the most direct way I've found to see it.
It's built by Mersudin Forbes, who I have a huge amount of respect for. He has poured a serious amount of expertise into this, and in my experience it is a long way ahead of anything else in the space.
🔗 Agent Checker
As ever, thanks for reading. If you found this useful, forwarding it on to someone else who runs or works on a large-catalogue Shopify store helps us more than just about anything.
See you next week.
Sam
Who am I and what does Blink do?
I'm Sam Wright, Managing Director and co-founder of Blink SEO. We're a search marketing agency that works exclusively with large-catalogue Shopify stores - the kind where the challenge isn't just visibility, but helping customers find the right product in the first place.
We specialise in SEO and PPC for merchants managing hundreds or thousands of products. Most of our work starts with taxonomy: how a store is structured and categorised, and what that means for organic performance, paid efficiency, and on-site discovery.
If you found this useful and aren't already subscribed, you can sign up at blinkseo.co.uk/blinktank.
If you run a large-catalogue Shopify store and want to talk about organic or paid performance, reply to this email or find me at blinkseo.co.uk.