SMART SHOPPING VS PERFORMANCE MAX: WHAT CHANGED

I managed Smart Shopping campaigns right up until Google pulled the plug in 2022, so the whole smart shopping vs performance max debate isn't theory for me. I lived the migration across a dozen accounts, watched the numbers move, and spent the following months figuring out what I'd actually gained and what had quietly slipped out of my hands. This article is that reckoning: not a feature table, but the practitioner's version of what changed and what it cost in control.

The short version is that Performance Max is genuinely more powerful and, on most of my accounts, more profitable. The longer version is that "more powerful" and "more visible" are not the same thing, and the upgrade traded a narrow, opaque box for a much wider, still-opaque one. If you're weighing the two, or just trying to understand what your account inherited when Google flipped the switch, this is the honest angle.

I'll walk through how Smart Shopping actually operated, what the upgrade added and removed, the specific levers I lost, whether the trade was worth it, and how I structure Shopping now to get some of that control back. Everything here is inference from managing real spend, not a rewrite of the release notes.

How Smart Shopping campaigns actually worked

To understand what changed, you have to understand what you were starting from. So, how do smart shopping campaigns work? You gave Google three things: your Merchant Center product feed, a budget, and a single goal, usually Target ROAS or Maximize conversion value. From there the automation did everything else. It decided bids, chose products to push, and spread ads across Shopping, Display, Gmail and YouTube, with no keywords and almost no manual dials to turn.

That was the appeal and the trap in one. Smart Shopping was already a black box: no Search terms report, no way to split Shopping spend from Display, no placement exclusions. Yet it felt safe because it stayed inside Shopping surfaces. I roughly knew the shape of what it was doing even when I couldn't see the detail. Understanding that a Shopping campaign runs on feed signals rather than keywords starts with grasping how keywords behave differently in Shopping, since there was never a positive keyword field to lean on in the first place.

In practice I steered Smart Shopping the only ways I could: feed quality, product titles, budget, and the ROAS target. Google's own documentation frames Shopping ads as running on product data instead of keywords, which is exactly why the feed did so much of the heavy lifting. You can read the mechanics in Google's reference on how Shopping ads and Performance Max campaigns work. That feed-first logic carried straight over into Performance Max, which is why the transition felt continuous rather than a clean break.

What actually changed in the upgrade

Google didn't rebuild Shopping from scratch; it wrapped Smart Shopping in a bigger campaign. Performance Max kept the same core: product feed in, conversion goal set, automation runs. What it added was reach and inputs. On the reach side, ads became eligible on Search text ads, Discover and Maps on top of the Shopping, Display, Gmail and YouTube surfaces Smart Shopping already used. On the input side, Performance Max introduced asset groups, so you now feed it your own headlines, images and videos rather than relying on the product feed alone.

The results backed the change. Google's upgrade testing reported an average 12% increase in conversion value at the same or better ROAS, and while I treat vendor numbers with caution, my accounts broadly agreed. The extra inventory found buyers Smart Shopping simply couldn't reach, particularly on YouTube and Discover. When the campaigns auto-upgraded, learnings and settings carried across, so I didn't restart the algorithm from zero, which mattered heading into a holiday season.

Two genuinely useful additions came with it. Audience signals let me tell the automation who I thought would convert, nudging its exploration instead of leaving it blind. And a channel report eventually arrived, giving at least some view into where spend was going across surfaces, something Smart Shopping never offered. If you're choosing between this and a lighter-touch build, my comparison of standard Shopping vs Performance Max covers where each one earns its place.

The control I actually lost

Here's the honest part, and it's more nuanced than the usual "Performance Max is a black box" complaint. I didn't lose granular bidding, because Smart Shopping never gave me that either. What I lost was boundary. Smart Shopping stayed inside Shopping surfaces, so its spend was at least contained. Performance Max can push budget into Display, video and Search text ads I never explicitly asked for, and the ratio between those surfaces isn't mine to set.

The bigger practical loss was cannibalization. Because Performance Max can serve Search text ads, it started competing with my existing brand and generic Search campaigns in a way Smart Shopping never did. Spend that used to be cleanly attributed to Shopping now overlapped with Search, and untangling it took real work. Optimising a campaign you can only partially see means leaning harder on the levers you still hold, which is why my approach to optimising Shopping campaigns shifted toward feed and exclusion work rather than bid tinkering.

I also lost the comforting illusion of understanding. With Smart Shopping I couldn't see much, but the box was small enough to reason about. Performance Max is a bigger box, so even with the newer channel report, the gap between what it's doing and what I can observe is wider. The controls I have left are real but blunt: feed quality, budget, audience signals, brand exclusions and account-level negatives. Everything downstream of those, I trust the machine with.

Was the trade worth it?

For most e-commerce accounts I run, yes, measured on results. The extra reach was real revenue, the learnings carried over, and the audience signals gave me a way to steer that Smart Shopping lacked. If your only question is whether Performance Max makes more money than Smart Shopping did, my experience says it usually does, and Google's 12% figure roughly matched what I saw.

But "worth it" depends on what you were optimising for. If you valued reach and were comfortable steering with feed, budget and exclusions, the upgrade was a clear win. If you needed to see and control every placement, if visibility mattered as much as outcome, then you lost something the day Smart Shopping disappeared, and no amount of conversion lift fully replaces it. The uncomfortable truth is that Google decided that trade for you; the auto-upgrade wasn't optional. The fundamentals of how Google Shopping works as a channel didn't change, but who holds the controls did.

How I run Shopping now to claw control back

The workaround most experienced managers landed on is simple: don't rely on Performance Max alone. On my larger accounts I run Performance Max for scale and reach, then keep a lean Standard Shopping campaign alongside it purely for the visibility Performance Max took away. Standard Shopping is the only current type that still gives me the full Search terms report and manual bidding, so it doubles as a testing ground and a window.

Beyond that, I invest heavily in the levers Performance Max still respects. Feed quality and product titles do more work than ever, because they're the clearest signal I can send. Audience signals shape exploration. Brand exclusions and account-level negatives stop the campaign wandering into searches I don't want to pay for, which also blunts the Search cannibalization problem. None of this restores the granular control Smart Shopping never had, but it restores the boundary and the visibility the upgrade quietly removed.

If I compress the whole smart shopping vs performance max question into one line, it's this: Performance Max is the more capable campaign, but it asks you to trust more surfaces with less sight. The job now isn't to fight that, it's to steer it well with the inputs you're still given, and to keep one campaign running where you can actually see what's happening. That's the control worth protecting.