I've been managing Google Shopping campaigns long enough to have watched the conventional wisdom shift underneath my feet more times than I'd like to admit, and if there's one thing I keep coming back to, it's that the brands who struggle most with Shopping aren't struggling because they don't care. They're struggling because they set things up, felt good about it, and then moved on to the next thing. Which I understand completely, because setup feels like an accomplishment, and in some ways it is. But it's also just the beginning, and I think confusing those two things is responsible for a lot of underperforming campaigns that nobody quite understands.

So let me try to lay out what I actually think matters in 2026, with the caveat that this space keeps moving and some of what I say here will probably need updating again in another year.

Also, importantly, it's worth noting that these are a big random and not in any particular order... other than the order that they came into my brain!

‍

Setup Is Not the Finish Line

The single most common misunderstanding I see about Shopping feed optimization is the assumption that once the feed and campaigns are live, the heavy lifting is done. It isn't. The setup matters, a lot, and I don't want to minimize that, but there is real, important work that must come after as well.

Your feed needs regular attention.

Your campaigns need ongoing refinement.

The algorithm rewards engagement with the system over time, and walking away after launch is roughly equivalent to planting a garden and then never going back outside.

PMax specifically tends to reward patience and consistency in a way that can feel unrewarding in the short term, which is part of why I think so many brands get frustrated with it and start making changes too quickly. There are times when the right move is to make an adjustment and then genuinely give the system time to respond before you do anything else. That's harder than it sounds when you're staring at a dashboard that isn't doing what you hoped.

‍

On Found by Google automated site scraping (And Why We Still Don't Rely on It)

There's a version of feed management advice that basically says "let Google handle it," and I want to address that directly because it's become more tempting as Google has expanded its automated feed capabilities, including something called "Found by Google" in Merchant Center Next that can pull product data directly from your site without a traditional feed file.

I'll acknowledge that Google's crawling and automation have gotten better. That's true and worth saying. But from what we've observed at ZATO across the accounts we manage, relying on that automated system as your primary feed source tends to introduce errors that are frustrating to diagnose and correct, especially by introducing outdated (or even dead) products on the site that you wouldn't actually want to be sent to GMC. The quality of your feed data is ultimately your responsibility, and I think you're better served by owning that rather than outsourcing it to a crawler and hoping for the best.

Clean, accurate, manually managed data still outperforms automated sourcing in our experience, even if the gap has narrowed somewhat. Will it be like this forever? No, but you're managing now.

‍

Your Feed Is the Foundation of Everything

I think of the Shopping feed as something like the foundation of a building. You can do a lot of interesting things above ground, but if the foundation has problems, everything built on top of it is compromised in ways that aren't always visible until something goes wrong. The feed is where Google gets the information it uses to match your products to searchers, and the richer and more accurate that information is, the better chance you have of showing up in the right places.

What I mean practically is filling out every field you can, including the optional ones. Titles, descriptions, product types, images, prices, Google product categories, and then the attributes that often get skipped like color, size, and material. These feel minor, and individually they might be, but collectively they give the algorithm more to work with when it's deciding whether your product is a good match for a particular search. The brands I've seen do this well treat every attribute as an opportunity rather than a checkbox, which is a mindset shift that sounds small and actually matters quite a bit.

With the exception of the all important titles, Product types in particular are worth calling out specifically because they're one of the attributes that Google has noted has an actual impact on ranking when optimized well, and they're also one of the most commonly left blank or filled in carelessly. If you're not sure where to start with feed optimization, product types are a reasonable place to put your energy first.

‍

Metrics Worth Actually Watching

In Merchant Center, the metrics I keep coming back to are conversion rate and click-through rate, specifically watching for sudden drops in either one because those tend to be early signals that something has shifted, whether that's a competitive change, a product issue, or something in the feed itself. This is because you actually get to analyze qualitative changes, not simpley

Auction insights are also worth checking regularly, particularly if you're in a category where Amazon competes heavily. Understanding whether big players are ramping up or pulling back gives you context for your own bidding decisions that you can't get from your account data alone. I find this information underused relative to how useful it actually is.

For more on how we think about Google Merchant Center diagnostics and what to watch for on an ongoing basis, there's quite a bit on the blog that goes deeper on specific metrics and what they tend to signal.

‍

When to Exclude a Product (And When Not To)

Not every product belongs in your feed, and it's worth being deliberate about exclusions rather than either keeping everything or pulling things out reactively. The easy exclusions are gift cards, free samples, and out-of-stock items. Google may manage stock statuses automatically in some cases, but I'd typically rather manage this myself and know it's clean than trust automation on something that directly affects whether your products are eligible to serve.

For low-stock products (fewer than two or three units), removing them can prevent the awkward situation of advertising something you can't actually fulfill across multiple channels simultaneously... and there is also a case to be made for removing products that will sell anyway, so you're not wasting ad dollars on them! With that being said, be careful not to over-exclude, and that's not a good rule for everyone but it might be something to consider.

Another thing to consider as you're thinking about "feed exclusions" is that there is a difference between "not advertising something actively" and "actually excluding it from the product feed". The difference became big when Google dreated Free Product Listings. With these, you can not advertise a product, but still send it to GMC so it can potentially get free organic shopping clicks. This is a great option to consider for any product instead of FULLY excluding from the feed, and in this case you can utilize the destination_excluded attribute in Google or simply point your individual Google Shopping campaigns to a certain custom label that only ever targets your "active" products.

For genuinely underperforming products, I'd also resist the impulse to exclude them immediately. A product with poor performance in one campaign structure might behave differently in another context: with a tighter ROAS target and a smaller dedicated budget. Testing that before excluding gives you better information than just assuming the product is the problem.

The one situation I'd really encourage you to investigate carefully before taking action is a product with high clicks and low or no sales. Before you exclude it from advertising, or assume the product is wrong for Shopping, check whether those clicks are influencing sales through other channels or through view-through attribution. A drop in clicks on that product sometimes correlates with a broader dip that tells you the product was doing more work than the direct conversion data suggested.

‍

The Opportunities Most Brands Leave on the Table

1) Free product listings in Merchant Center are typically underutilized when it comes to thinking about them, and I think it's because they don't feel like a lever worth pulling when you're focused on paid performance. But it's free exposure, and in a category where visibility matters, that's not nothing. If you're not submitting all of your products to free listings, it's worth doing. Like I mentioned above, if you have 500,000 SKUs but only a $20,000/mo budget, then you should ONLY focus on advertising your top selling SKUs (let's say, your top 1000)... but still send all 500K SKUs to free product listings!

2) Using search term data to inform your product titles is another one that I think gets more lip service than actual implementation. The terms people are actually using to find products in your category are some of the most valuable data you have access to, and incorporating that language into your titles directly is one of the more reliable ways to improve your relevance scores and visibility over time. Our thinking on Google Shopping query sculpting and title optimization goes deeper on this if you want to get into the specifics.

‍

Campaign Segmentation: Some Principles Worth Keeping

There's no universal right answer for how to segment Shopping campaigns, and I want to be upfront about that rather than pretend there's a formula that works for everyone. But there are some principles I keep coming back to.

For standard campaigns, starting with segmentation based on major product categories and then drilling down by margin or search behavior tends to give you more control over where your budget goes than a single catch-all structure. For PMax, the asset group is where most of the meaningful segmentation work happens, and the question worth asking is which products have genuinely different audiences or require meaningfully different creative, as well as "which product groups generally should receive the more aggressive bidding or budgets". When the answer to that question is "yes, these are distinct groups," building separate campaigns, and then asset groups for them tends to produce better results than lumping them together and hoping the algorithm figures out the difference.

‍

The Bottom Line

Shopping feed optimization in 2026 is still, at its core, about the same things it has always been about: clean data, thoughtful structure, ongoing attention, and genuine curiosity about what the performance signals are telling you. The tools have changed, PMax has added complexity, Merchant Center Next is still finding its footing, and Google keeps moving things around in ways that occasionally require everyone to relearn where everything lives.

But the brands that consistently do well in Google Shopping tend not to be the ones who found a clever hack. They're the ones who treated the feed as something worth caring about on an ongoing basis, stayed engaged with the data, and resisted the temptation to set it and walk away. Which, from what I've seen, is available to anyone willing to actually do the work.

‍