Google Shopping campaigns in 2026: the setup guide that skips the basics

Google Shopping in 2026 is not what it was in 2022. The introduction of Performance Max, the deprecation of Smart Shopping, and three years of machine learning improvements have changed the decision structure significantly. The fundamental question is no longer “how do I set up Shopping ads?” but “which campaign type, and why?”

This guide answers that question first, then works through feed requirements, the setup sequence, ROAS benchmarks, and the mistakes we fix most often when inheriting Shopping accounts. See our guide on what Google Ads costs if budget planning is the prior question.

Standard Shopping vs. Performance Max

Standard Shopping campaigns give you control. You set bids at the product or product-group level. You see the search terms that triggered your ads. You can exclude queries that waste budget and build ad group structures that reflect your catalogue logic.

Performance Max gives Google control. It serves Shopping ads, but also Display, YouTube, Gmail and Maps — all from one campaign, optimised by machine learning toward the conversion you specify. You cannot see individual search term data. You cannot bid by product. You can influence it through asset quality and audience signals, but you cannot direct it.

The practical question: when does giving up control improve results?

Start with Standard Shopping if: your account has fewer than 50 conversions per month (PMax needs data to optimise, and without it will behave erratically); your margins differ significantly by product and you need granular bid control to stay profitable; or your feed has quality issues that you haven’t fully resolved yet — PMax amplifies feed signals, good and bad.

Test Performance Max if: your account consistently generates 50+ monthly conversions and has clean purchase data with real revenue values; you have a broad catalogue where managing bids by product group in Standard Shopping becomes operationally unsustainable; or you are expanding into new markets where you don’t yet know which queries convert.

The transition from Standard to PMax is not automatic. Feed quality and conversion signal quality are what make it work. Without both, consolidating into Performance Max typically produces worse results, not better.

Merchant Center feed requirements

The feed is the creative in Shopping. A well-structured feed in a mediocre campaign outperforms a brilliant campaign structure built on a weak feed. Every field below must be present and accurate before a campaign launches:

Optional but high-impact fields: brand, color, size, material, product_type. These feed into Shopping filters and significantly improve Google’s ability to match your products to relevant queries. Add them.

The most common feed error that triggers Merchant Center suspensions: mismatched prices between the feed and the landing page. The price in the feed must exactly match what a user sees on arrival, including VAT where applicable. This is the single most frequent reason accounts get suspended — and the most avoidable.

The 6-step campaign setup

The sequence below avoids the most common launch failures. Steps 1 and 2 are prerequisites — skipping either produces a campaign that spends before it can learn.

1. Audit and clean the Merchant Center feed. Every required field present, prices matching landing pages, GTIN populated wherever available. Run the feed through Merchant Center Diagnostics and resolve all errors and warnings before creating a campaign. A feed with active errors will restrict impression eligibility from day one.
2. Set up conversion tracking with transaction values. Purchase events with real revenue values are essential for ROAS optimisation. If you use GA4, link it to Google Ads and import the purchase event with the value parameter. Verify that the import is recording actual revenue figures, not a fixed placeholder — the algorithm optimises against whatever value it receives.
3. Choose Standard Shopping or Performance Max. Apply the criteria from the section above. For accounts under 50 conversions per month, Standard Shopping first. For established accounts, run both in parallel with a clear budget split and a 6-week evaluation period before committing to one.
4. Structure Standard Shopping campaigns by product category. One campaign per major category (Apparel, Electronics, Home & Garden), with ad groups for sub-categories. This gives you budget and bid control at the level that matters — critical when margins differ significantly across your catalogue.
5. Set your initial ROAS target conservatively. A target set too aggressively starves the campaign of traffic during the learning phase. Start at 20–30% below your actual target ROAS and tighten after four weeks of stable conversion data.
6. Apply negative keywords and brand exclusions. In Standard Shopping, upload a negative keyword list to exclude irrelevant and wasteful query patterns. In Performance Max, add brand exclusions to prevent the campaign from cannibalising your Brand Search campaigns — without this, you pay twice to capture demand you already owned.

ROAS benchmarks by category

These are pmax internal benchmarks from managed accounts in Germany and Spain, 2024–2026. Use as orientation rather than targets — your gross margin, average order value, return rate and competitive density will determine what’s achievable in your specific account.

The Electronics range looks low. That’s accurate. Electronics has the thinnest margins and the most price-sensitive buyers in European e-commerce. A 2.5× ROAS on a product with a 12% gross margin is unprofitable. Calculate your break-even ROAS before setting a target: divide 1 by your gross margin percentage. At 35% margin, break-even is 2.86×. Set your target above that number, not below it.

A Mallorca case: Mallorca Artesanía

Mallorca Artesanía is an artisan products e-commerce business based in Palma — handmade ceramics, textiles and leather goods sold to European customers online. They came to us running Standard Shopping with a 2.2× account ROAS. Technically profitable, but well below what the catalogue and the margins supported.

The diagnosis was structural. Their Standard Shopping account had 38 ad groups, most of them severely under-funded. Google was not collecting enough conversion data from any individual ad group to optimise effectively. The campaigns were fragmented to the point of paralysis — each group had a budget too small to exit the learning phase.

We ran a feed quality pass first: fixed 23 missing GTINs, rewrote 140 product titles to include key attributes (material, colour, origin), and added lifestyle images alongside the existing plain product shots. Only then did we consolidate to Performance Max.

The PMax transition was not the intervention. The feed quality pass and the conversion signal cleanup were the intervention. Without both of those, consolidating 38 under-funded ad groups into Performance Max would have produced an equally under-performing PMax campaign with fewer levers to diagnose it. See our client cases for the full breakdown methodology.

Common mistakes

Running Performance Max on a new account. PMax without conversion history optimises for nothing useful. You will spend budget in the learning phase before Google has any signal to learn from. Standard Shopping first — let it collect 50 conversions, then evaluate PMax.

Ignoring feed quality. The feed title is the primary relevance signal in Shopping. “Blue Dress 123” loses to “Women’s Floral Maxi Dress — Navy Blue — Sizes 8–18” on every query that matters. Title and description rewrites consistently produce the highest ROI of any feed optimisation task — higher than bid changes, higher than audience layers.

Setting ROAS targets without knowing break-even. Target ROAS without knowing your gross margin is guessing. Calculate break-even first (1 ÷ gross margin). Set your target above it with a buffer for return rates. Agencies that skip this step and set targets by feel are optimising your account for their dashboard, not your P&L.

Broken GA4 linkage. Shopping campaigns need clean purchase data with revenue values. If your GA4 import fires without the value parameter, Google is optimising for conversion volume — not profit. Check the import in Google Ads under Conversions: if the reported values are zero or identical across all transactions, the linkage is broken.

Missing brand exclusions in Performance Max. Without brand exclusions, PMax will compete with your own Brand Search campaigns for branded queries. You pay twice to capture demand you already owned. Add brand exclusions in PMax campaign settings from launch day.

Our Google Ads service covers Shopping, Performance Max and Search. For businesses based in Germany or running campaigns targeting the German market, see our Google Ads Agentur Spanien page.

Need help with this?

If any of the above feels like a problem you have, tell us a bit about your situation and we will come back within a working day. First conversation is 30 minutes, on us.