Google Ads Campaign Types for Ecommerce: Search, Shopping and PMax

Google Ads gives an online store three campaign types for its products: Search for keyword-led text ads, Standard Shopping for product listings you bid on product by product, and Performance Max (PMax), which spreads one budget across all of Google’s channels.

Which one fits depends on the product behind the click. A bestseller at a ROAS of 5 with a 15% margin loses €25 per €100 of ads, while an accessory at a ROAS of 3 with a 45% margin earns €35.

The benchmarks below come from Shopiator’s 23-account dataset (September 2026); every target is calculated from margin, with the formula shown.

What does each Google Ads campaign type control?

Each Google Ads campaign type trades control for reach: Search lets you pick the query, Standard Shopping the product and its bid, and Performance Max only the inputs, while Google picks the placement.

Shopping formats take about 85% of the clicks on Google retail campaigns, a WebFX and DemandSage figure compiled in Koongo’s Google Shopping statistics for 2026.

It is an industry estimate rather than a Google disclosure, but the direction holds: for a store, the product feed does most of the advertising.

For costs, metrics and the wider profit-first strategy, see ecommerce PPC by product profit.

When should you use Search?

Use Search when you know the query: your brand name, a model number or a precise non-brand phrase.

You control keywords, match types, negatives and ad text, but Search is keyword-led, so a text ad pointing at a product page still has no bid per product.

Keep brand and non-brand in separate campaigns. Someone typing your store’s name already knows you, and mixing their sales with new-customer searches makes acquisition look cheaper than it is.

When should you use Standard Shopping?

Use Standard Shopping when merchandising decisions need to become advertising decisions.

It matches products from your Google Merchant Center feed to searches and lets you split them into product groups, isolate item IDs and set a bid per group, manually or with an automated strategy.

A fragile item with expensive delivery shouldn’t share a bid with a compact, high-margin accessory, and clear product categories in your feed make that split easier.

Campaign priority routes products between overlapping Standard Shopping campaigns, but it doesn’t override Performance Max.

When should you use Performance Max?

Use Performance Max for reach once your feed, prices and purchase tracking are dependable.

It serves ads across Search, Shopping, YouTube, Display, Discover, Gmail and Maps from the assets, products, budget and target you give it, and Google Performance Max explained covers how it decides.

Audience signals guide Performance Max rather than restrict it, and asset groups organise creative without getting a budget of their own. So set boundaries with listing groups and product exclusions before you add budget.

One test before you scale it: trace a recent sale from the advertised product to the basket that was bought. If you can’t see what the click sold, you can’t steer growth product by product yet.

What do ecommerce PPC benchmarks show?

Google Shopping averaged a CPC of $1.32, a 2.24% conversion rate, a CPA of $59.09 and a ROAS of 2.81 across 23 ecommerce accounts from January 2025 to Q3 2026, according to Shopiator’s ecommerce Google Ads benchmarks (updated 13 September 2026).

Show the dataHide the data

Google Shopping ROAS by country across 23 ecommerce accounts

Country	ROAS
United Kingdom	3.77
United States	2.82
Australia	2.06
Canada	1.97

Why don’t cheap clicks mean profitable orders?

Cheap clicks don’t mean profitable orders because ROAS depends on conversion rate and order value as much as on CPC.

Canada’s clicks cost $0.86, close to the UK’s $0.81, yet Canada converts at 1.57% against the UK’s 3.85% and returns the lowest ROAS.

Australia shows the reverse: the most expensive clicks, $1.91, and the highest conversion rate, 5.54%, still end at a ROAS of 2.06.

These are different account mixes, not one set of products tested in four countries, so category, order value and conversion settings explain part of each gap.

How do you test a ROAS against your margin?

Test a ROAS against the product’s own break-even ROAS, 1 ÷ contribution margin, never against a market average. A 15% margin breaks even at a ROAS of 6.67; a 45% margin at 2.22.

That is why the bestseller in the chart below loses money on a ROAS of 5 while the accessory earns on 3.

In profit terms, POAS = ROAS × contribution margin and break-even POAS is 1.0; POAS vs ROAS explains when to bid on profit instead.

Show the dataHide the data

Profit after ads per €100 of ad spend: higher ROAS, lower profit

Product (ROAS, contribution margin)	Profit after ads
Bestseller (ROAS 5, 15% margin)	-€25
Accessory (ROAS 3, 45% margin)	€35

To run the test across your products, build one comparison per product group with three sets of columns. The break-even ROAS calculator gives the floor for each group.

Benchmarks tell you whether your market looks unusual. They don’t tell you which product deserves the next euro.

Which products belong in which campaign type?

Products belong in one campaign when their prices, margins and conversion volumes are alike; when those diverge, split them so each group gets a target it can reach.

One campaign works like one thermostat for a whole house: fine until the rooms need different temperatures.

Pooling still has a use. Products that sell rarely share conversion data in a pooled campaign, which gives automated bidding more to learn from.

Heavy goods, sizes that come back often, clearance stock and high-margin accessories shouldn’t all inherit the same target, though.

To decide which products go where, segment them by ad clicks and by return against their own target.

Product Segmentation in Product Metrics does this with machine learning and writes each product’s segment to a Merchant Center custom label, so a listing group can route it. Two or three campaigns are usually enough; Google Performance Max explained maps each segment to a campaign.

A higher target restricts delivery, but it doesn’t guarantee profit or cap spend. Use exclusions or budgets when you need a hard limit, and never build a campaign per product.

Reviewing segments by hand works for 50 products. Across thousands, after every price, margin or stock change, it doesn’t scale.

How do targets work in each campaign type?

Each Google Ads campaign type gives products a different target at a different level, and in Performance Max only the campaign split decides which products share one.

Work out each group’s target as 1 ÷ (contribution margin − profit kept), with both as decimals of revenue excl. VAT: a 25% margin that keeps 5% needs a target ROAS of 500%.

Standard Shopping is the only type that can give products different targets inside one campaign.

In Performance Max, products with a 15% and a 45% margin in one campaign get the same target, which overpays for one or starves the other.

How to set a target ROAS from your margin has the full margin grid and how to change a target without shocking the bidder.

Is a higher target ROAS always better?

No, a higher target ROAS isn’t always better: it can lift reported efficiency while cutting profitable orders. Judge a target change by total contribution after ads, not by whether the percentage rose.

smec’s State of Performance Max 2025 found the median ROAS target across 4,000+ campaigns rose from about 470% to about 600% as accounts matured, as MoonSauce’s ecommerce PPC guide summarises.

A stricter target shows the account got pickier, not that it kept more profit.

Google Ads has no POAS bid strategy, but Maximise conversion value with a target ROAS steers on whatever value you send.

Send profit as the conversion value, and it bids on profit; just don’t reuse a revenue-based target unchanged.

Then sort each margin band into one of three actions: scale products clearing their target, constrain those selling without enough margin, and pause or repair those below break-even.

Before scaling, check that the extra demand won’t need a discount or shipping subsidy that changes the sum.

When is conversion data good enough to scale a product?

Conversion data is good enough to scale on when mature attributed results, backend orders and product margin support the same decision; if they disagree, close the gap before buying more traffic.

A product can look weak because its purchases haven’t arrived yet.

Google Ads records a conversion against the click date, so recent days fill in later; leave out the incomplete period, or compare click cohorts that have had the same time to convert.

Consent shapes what Google Ads can see. Google’s consent mode requires ad_user_data consent for enhanced conversions and tag-based conversion tracking, and denied consent limits the click-linked data available for bidding, which consent mode partly models back.

Modelled conversions aren’t inflated by default; the mistake is treating them as observed product orders.

Where Google doesn’t say which conversions were modelled, record the uncertainty at campaign level and check the result against your shop’s orders. If a product has too few orders, judge it with a coherent product group instead.

How do you get more conversions back without bypassing consent?

Send conversions from a first-party endpoint, and check what your tags actually deliver. Enhanced conversions vs server-side tracking separates the implementation choices from the consent requirements, which every setup still has to honour.

Full Signal Tracking sends conversions from an endpoint on your own domain, so it gets through ad blockers, preserves click IDs and improves match rates.

To see what blockers do to a standard setup, run the ad blocker conversion tracking test in your own browser.

Which mistakes waste ecommerce PPC budget?

Most wasted ecommerce PPC budget comes from acting on a number without checking what produced it: a blended ROAS, an immature week or a margin that has since changed.

The clicked product also isn’t always the one bought. Check basket contents before excluding a product that brings in shoppers who buy profitable complements, and don’t credit a low-margin product with a whole basket’s profit.

How do you run a break-even check before a budget change?

Run the check per margin band, and change one band at a time:

1. List each band: its products, revenue excl. VAT, contribution before ads and stock.
2. Calculate its floor: break-even ROAS = 1 ÷ contribution margin.
3. Mark the shortfall: flag products below the floor, and separate real losses from small, inconclusive tests.
4. Validate the period: mature conversions, stable consent and reconciled order values.
5. Change one band: start where the avoidable loss is biggest, and note the change and its review date.

The Google Ads budget calculator turns the result into a budget. Don’t change targets and budgets across the whole account at once, or you can’t tell a useful constraint from lost profitable demand.

Final thoughts

Ecommerce PPC isn’t a choice between Search, Shopping and Performance Max.

It’s a choice about which products get the next click, and a campaign’s ROAS can’t make it for you, because the product with the best ROAS can be the one losing money.

Start with one column: contribution margin next to ROAS in your product report. The products that deserve more priority, and the ones quietly costing you, usually show up in the first afternoon.