Google Ads for Giant Catalogs: What to Do When Heroes, Zombies, and Villains Aren't Enough

If you've got a massive product catalog and you've already segmented your Google Shopping campaigns into performance tiers (Heroes, Zombies, Villains, and so on), congratulations. You're doing more than most. You've read the right things, you've done the work, and you're asking the right questions. Which is probably why you're still a little frustrated, because you've done all this and you're not seeing the explosion in growth you were hoping for!

Here's what I want to say right up front, and I mean this sincerely: I don't think there's a smoking gun waiting for you. I think incremental improvement, applied consistently over time, is actually the path. And I know that's not the exciting answer, but from everything I've seen, it tends to be the real one.

‍

First, a Word About Zombie Campaigns

I want to spend a moment on Zombie campaigns specifically, because I see a lot of catalog-heavy brands treating them like buried treasure. The idea is appealing. You've got all these products just sitting there, not selling, and surely with the right campaign structure they'd start moving, right?

Maybe. But there are two things worth sitting with before you go deep on Zombie campaigns.

The first is cannibalization. When you launch a new campaign, even a well-structured one, Google sees a fresh opportunity and gets excited. It starts pushing traffic toward those "new" products to test them. The problem is that this traffic often isn't new traffic at all. It's traffic that would have gone to your already-converting products. So you haven't unlocked new growth, you've just reshuffled the deck, and possibly made things worse. (We've written about things like campaign cannibalization, and a lot of other shopping/pmax ponderings, in more depth over on our Google Shopping resource page, if you want to go deeper on that particular rabbit hole.)

The second thing is a bit humbling, and I'll admit I've come around to believing it more than I expected to: Google's system, particularly with Performance Max, has gotten genuinely decent at identifying which products will actually hit your targets. When a product is a Zombie, it's worth asking whether Google is hiding it from you, or whether Google has already figured out what you haven't yet acknowledged. Sometimes Zombie campaigns are an expensive way to confirm that those products just aren't going to move.

None of this means Zombie campaigns are never worth running. But if you do run them, the key is making sure they're targeting genuinely different audiences, search terms, and product clusters than your main campaigns. Partial differentiation just creates partial cannibalization.

‍

Start From the Top: What Are Your Main Campaigns Actually Capturing?

The more useful question, I think, is this: what search terms are actually driving your successful PMax campaigns right now?

If the answer is that your main campaigns are winning on specific, intent-rich searches (something like "Mercedes C43 taillights" rather than just "car parts"), then here's something worth considering. For a catalog-heavy site with a good on-site experience, getting the click might matter more than getting the exact right product into the ad. If someone searches for a specific part, clicks through, and your site does a solid job of routing them to exactly what they need... then obsessing over which specific SKU appeared in the ad might be less important than making sure you're showing up for the right categories and makes in the first place.

Which leads me to what I think is actually the more valuable question for a big-catalog business: what brands and product categories are you NOT showing up for? That's where I'd be putting my energy, and it's something we touch on fairly regularly in our ecommerce PPC writing if you want more context around how that kind of thinking applies across different catalog types.

‍

A Campaign Structure Worth Testing

Here's a framework worth experimenting with, and I want to be clear that this is a test worth running carefully, not a wholesale restructuring of everything.

Think about splitting your approach into two lanes.

The first lane is your existing top-performing products on more generic, higher-level keywords. Leave those alone as much as possible. Let your core PMax campaign keep doing what it's doing. Don't mess with what's working.

The second lane is where it gets interesting. Pick a brand, vehicle make, or product category you want to grow but aren't currently winning. Set up a targeted Shopping campaign using query sculpting, which is a method of using negative keywords across campaign priority tiers to make sure your bids land where you actually want them to. The idea is that a high-priority campaign captures generic searches at a low bid (because you don't want to waste money there), while a second campaign with negative keywords funneling out those generic terms lets you bid more aggressively on the specific, high-intent searches where you actually want to compete.

So let's say you want to grow your Mercedes AMG auto parts business, you build one Shopping campaign that catches generic traffic cheaply, and a second one where you're really going after something like "C43 AMG headlights" with a real budget behind it. You're not just hoping Google figures out that you want to sell Mercedes parts. You're telling it, clearly, and putting money behind that signal.

This won't work overnight. It might not work at all for some categories, and that's actually useful information. But it's a smarter test than launching another PMax campaign and hoping for the best. (For a detailed walkthrough of query sculpting mechanics, this piece is worth bookmarking: Google Shopping Query Sculpting.)

‍

One More Idea to Sit With (Carefully)

This last thought comes with a genuine warning label: Please don't make major feed changes based on a blog post, including this one. But I think it's worth considering, so here it is.

For a catalog with hundreds of thousands of SKUs, there's a real question about what you're actually sending into your main PMax campaigns. All those variants, every color, size, and configuration, give Google a lot of signals to sort through. And for generic, top-of-funnel searches, that noise might actually be hurting your bidding efficiency.

The idea worth testing (on a smaller, non-critical campaign) is whether limiting your main PMax to your core, highest-converting variants gives the algorithm cleaner data to work with. Your site, if it's good, can do the job of routing someone from a core product to the exact variant they need. You're not losing the specific variants, they'd still be in other campaigns. You're just asking whether a more focused signal produces better bidding decisions on those broad searches. This connects to some broader thinking about campaign strategy that I think is worth exploring if you find yourself wrestling with these kinds of structural questions regularly.

I genuinely don't know if this applies to your specific situation. It might not. But it's the kind of question worth asking when you have a catalog this size.

‍

The Bottom Line

None of this is a silver bullet. I know you were maybe hoping for one, and I'd be doing you a disservice if I pretended otherwise. What I do think is that smart, systematic testing of these ideas, running a query sculpting campaign here, refining your PMax product mix there, getting more intentional about which categories you're actually trying to win, is the thing that compounds over time into real growth.

The brands that win in Shopping tend not to be the ones who found one perfect campaign structure. They're the ones who kept showing up, kept testing, and kept refining. Which, it seems to me, sounds a lot like most things worth doing.

‍