Making sense of spend, search and structure
Welcome back to #BlinkTank - a weekly look at what is happening across Shopify, search, and eCommerce for large catalogue brands.
This week starts with something very practical - a meaningful change in GA4 that could bring brands back towards a shared source of truth - and then looks at how browsing, merchandising and Shopping quality are evolving around that.
GA4 ad cost imports - and why this really matters
One update that almost slipped through the net is GA4’s ability to import ad costs from Meta, TikTok and Pinterest directly into GA4.
Edward Upton wrote a very good breakdown of this, and it is worth paying attention to.
We talk a lot about attribution and incrementality. One of the big problems over the last few years has been a loss of trust in GA4 as the central view of performance. Many brands have drifted towards third party dashboards, or they are not really looking at all channels together in one place.
As our focus is on organic search and GA4 is really the only place that this data can be viewed, this creates a pretty big challenge for us.
Being able to pull paid social and other ad costs into GA4 does not magically fix attribution, but it does close part of the gap that pushed people away in the first place. It becomes much easier to:
- Line up spend and performance across channels
- Compare campaigns side by side in one environment
- Apply GA4’s attribution model consistently
There are still tracking issues, and GA4 is far from perfect, but this change could be a significant step back towards having a single, shared source of truth instead of a patchwork of disconnected reports.
🔗 View the post here: https://www.linkedin.com/posts/edwardupton_hot-take-triple-whales-biggest-feature-activity-7393635317523832832-3XQm
OpenAI Atlas - what we have actually seen so far
Over the past couple of weeks, OpenAI’s Atlas browser has prompted a lot of “what does this mean for SEO and eCommerce?” conversations.
Right now, the honest answer is: not much in day-to-day terms. Usage is tiny. We are hardly seeing it in GA4 at all - and even if we were, GA4 would only ever show part of the picture. It is Mac only, very early, and most of the traffic so far looks like technically minded people experimenting.
Where it becomes interesting is not today’s numbers, but the direction it points to.
Shopify has spent years improving how product data is structured and pushing that into OpenAI’s ecosystem. That work will matter as assistants and browsers get better at answering product questions and comparing options. Atlas could end up as a bridge between the current “chat shopping” experience - which largely sidesteps merchandising - and something closer to search, but driven by richer attribute data underneath.
The hard part is behaviour. Changing how people search is difficult, and Google has had decades to train those habits. So the sensible move for most brands is not to “optimise for Atlas” yet, but to make sure their catalogue, attributes and taxonomy are in good shape for LLM driven discovery whenever adoption catches up.
Smarter control over smart collections
Shopify has released a genuinely useful improvement to smart collections – you can now exclude products by tag using an “is not equal to” operator when you set collection conditions.
For anyone managing a large catalogue, that unlocks a few things that previously needed awkward workarounds:
- Keeping specific tagged products out of broad, automated collections
- Excluding recalled, retired or limited items cleanly
- Tightening up themed or promotional collections without resorting to manual curation
It makes merchandising logic less brittle, especially for teams that lean heavily on tags to keep collections in line.
🔗 Read the Shopify changelog entry: https://changelog.shopify.com/posts/exclude-products-by-tag-in-smart-collections
Making product pages and listings less interchangeable
A theme that keeps coming up is how similar most eCommerce experiences feel. Two posts this week made that point well and are worth a look if you own PDPs or PLPs.
Jurjen Jongejan described many product pages as “copy paste clones”. When everything looks and sounds the same, the customer’s brain has no easy reason to choose. His suggestion is straightforward - find one clear point of contrast and say it plainly. It might be as mundane as “battery included” or “free lifetime repairs”, but if nobody else is saying it, it becomes a genuine reason to pick you.
🔗 View the post here: https://www.linkedin.com/posts/jurjenjongejan_ecommerce-abtesting-activity-7393546720611237888-449c
Simon Wesierski shared a nice example from lighting - a product listing page where you can toggle the “lights on / lights off” state directly in the grid. The whole page dims slightly and you immediately see how each lamp behaves in different conditions. It is a neat interaction that makes the products feel more tangible without forcing a click into the PDP.
🔗 View the post here: https://www.linkedin.com/posts/simonwesierski_ecommerce-design-ugcPost-7393578200733249536-uDac
If structural patterns are going to stay familiar, the details need to work harder. These are the kinds of tweaks that often move the needle without requiring a complete redesign.
Benchmarking Shopping experience
There is a free tool called Store Ratings that builds a Google Shopping Experience Scorecard for your site and lets you compare it to other domains in your market.
It pulls together datasets that Google uses as inputs for organic Shopping rankings and wraps them in a simple score out of 100. It is particularly useful if you already have - or are close to - the Top Quality Store badge, because it shows more clearly where you might be falling short.
For teams who like something more concrete than “improve UX”, this is a practical way to frame Shopping quality as something you can measure and improve over time.
🔗 Explore the tool here: https://storeratings.co/
Retail as a “wicked problem”
If you are interested in going deeper into catalogue structure, data and why retail so often resists neat solutions, I recently joined the Retail is Detail Podcast to talk through some of that in more detail.
We covered how taxonomy, product data and measurement fit together for larger catalogues, and why getting the underlying structure right tends to make everything else easier.
🎧 Listen on Spotify: https://open.spotify.com/episode/5FswWh3DEmAUmUwpzjID5N
That is it for this week’s #BlinkTank.
From GA4 inching back towards being a useful source of truth, to Atlas hinting at what future browsing could look like, to Shopify tightening up how collections behave, it feels like we are adjusting both the plumbing and the decision making layer at the same time. The brands that tend to win are the ones that keep an eye on both.
All the best,
Sam