Google Performance Max vs. Standard Shopping: 12 months of data.

PMax promises automation. Standard Shopping offers control. We ran both side-by-side for twelve months across a portfolio of ecommerce accounts spanning North America, Europe, and APAC. The result wasn't a winner — it was a clearer answer to a question most brands are asking wrong.

Google's framing is that Performance Max is the future. Standard Shopping is being deprecated for some advertiser cohorts, has fewer new features shipped against it, and is increasingly positioned as a legacy product. That framing has caused a lot of brands to migrate spend before understanding what they were giving up.

Across the twelve months of testing, we ran matched-budget A/B splits between PMax and Standard Shopping on identical SKU sets, at identical bid targets, with consistent feed quality. Same products, same audiences, same goals. Different vehicle.

The pitch vs. the reality

The PMax pitch is automation: feed it a feed, set a target, and let machine learning find conversions across Search, Shopping, YouTube, Discover, Gmail, and Display in one campaign. The reality is that PMax does in fact deliver conversions — but the breakdown of where those conversions come from, and what would have happened without PMax, is harder to see than Google would have you believe.

Standard Shopping is the older model: bid by product group, see exactly what queries triggered impressions, control exactly which SKUs compete for which budget. Less automation. More transparency.

Our test setup

For each of the accounts in the test we created two parallel campaign structures: a PMax campaign with feed-only assets (no asset groups beyond the feed, to isolate Shopping inventory) and a Standard Shopping campaign with matched product groups, matched bid targets, matched budget caps. We ran a four-week burn-in for both, then measured for the next 48 weeks.

Where PMax won

• Newer or smaller catalogs. When the account had fewer than ~500 SKUs and limited historical conversion data, PMax outperformed Standard Shopping by a meaningful margin — typically 15–25% more conversions at parity ROAS. The machine learning had less to work with, but it also had less to overfit to.
• Long-tail discovery. PMax was consistently better at finding incremental queries Standard Shopping missed. The breadth of placements gave it surface area to discover demand.
• Lower management overhead. Genuinely. Once stable, PMax required substantially fewer hours of manual optimisation. That's a real cost saving.

Where Standard Shopping won

• Large, margin-aware catalogs. When the account had thousands of SKUs with material margin variance between them, Standard Shopping outperformed by 20–35%. The granularity of bid control let us push high-margin SKUs hard and back off low-margin SKUs aggressively. PMax's blended optimisation flattened that out.
• Brand-term protection. PMax frequently cannibalised brand search queries with Shopping inventory at the top of the SERP, paying for clicks the brand was going to get for free. Standard Shopping with negatives gave us full control. This was the single biggest source of waste we found in PMax accounts.
• Diagnosability when things go wrong. When CPA drifted upward, Standard Shopping told us exactly which queries were responsible. PMax gave us a black box and a chatbot.

The hybrid we landed on

The conclusion most of the portfolio converged toward wasn't "pick one." It was a hybrid:

• Standard Shopping carries the core catalog — the SKUs that drive 80% of revenue and where margin control matters.
• PMax carries tail discovery — the long tail of SKUs where machine learning's broader matching beats manual bid management.
• Brand terms get an explicit Search campaign with PMax brand exclusions applied. This single change recovered meaningful spend in every PMax account we audited.
• Budgets are reviewed monthly and rebalanced based on incremental performance — not on PMax's own conversion attribution, which still over-credits itself.

What we'd recommend by brand type

• Smaller D2C with <500 SKUs: PMax-first, Standard Shopping as a control comparison once per quarter.
• Mid-market retailer with >1,000 SKUs and margin variance: Standard Shopping primary, PMax for tail discovery only.
• Any brand spending meaningful money on Google: brand exclusions in PMax, always. The single highest-leverage change in the entire test.

The takeaway

PMax is not a scam. Standard Shopping is not dead. The actual question isn't which one to use — it's where each one belongs in your account structure. Twelve months of data says the answer is almost always both, structured intentionally, with brand-term protection baked in from day one.

The brands losing money on Google right now mostly aren't losing it because PMax is bad. They're losing it because they migrated everything into PMax, including brand-term clicks they were already winning for free.