Google Ads for Shopify Shops: 2026 Setup Guide

Christian Beeking

Co-Founder, Cloudginny

July 23, 2026 18 min read

Shopify does more of the Google Ads heavy lifting for you than any other commerce platform. Product feed, conversion tracking and your first Performance Max campaign all come out of a single onboarding flow, without you touching a line of code. That convenience is why so many Shopify setups are only half finished. They look finished. The things that break sit in completely different places than they do on a self hosted shop, and most guides out there still describe the Shopify of the pre checkout upgrade era.

This guide is written for Shopify merchants running a monthly Google Ads budget between €500 and €20,000, whether you are starting fresh or auditing an account that has been running for a while. It reflects the state of play in July 2026 and tells you where each claim comes from. Where a widely quoted number has no solid source behind it, we say so instead of passing it on.

One thing up front, because it overrides everything else. If your shop is not on Shopify Plus and your conversion tracking still lives in Additional Scripts on the order status page, you have until August 26, 2026. After that Shopify upgrades those pages automatically and those scripts are gone. No error message, no warning inside your Google Ads account. Which is why the section on it sits near the front.

Contents

What Shopify handles for you and what you still have to build

The Google & YouTube app sits in the Shopify App Store, published by Google rather than by Shopify. It is free to install and it lives in your admin under Sales channels. Shopify describes the channel as managed by Google, and the wording is worth taking literally. If your Merchant Center account gets suspended or an ad gets disapproved, Shopify support is not an escalation path; they will point you at Google.

The work it does cover is worth having.

• Product sync into Merchant Center. Google expires product data after 30 days. The channel refreshes it inside that window automatically and syncs again whenever something changes. On self hosted shops, the cron job nobody watches is the most common cause of a sudden loss of reach, and the channel removes that failure mode entirely.
• Conversion measurement for Google Ads. According to Shopify this works even without a Merchant Center account, so it works whether or not you run shopping ads at all.
• Verifying and claiming your domain. Both happen automatically during onboarding, as long as you only run one domain.
• One Performance Max campaign. You start it in the app, it is created in Merchant Center, and you manage it in Google Ads. No other campaign type can be created from the app.
• Product Studio. Google's image tooling sits directly inside the app for Shopify merchants, with background removal, generated scenes and upscaling. Video generation is only available in selected countries, as of July 2026 including Australia, Canada, India, Japan, the UK and the US. Google changes that list regularly, though the image features work everywhere.

Where the channel stops

Google documents the limits itself, and they account for most of the setups that stall halfway.

The order to do this in

Two of the steps below are miserable to undo after the fact, so the order is not arbitrary.

1. Set your primary domain. Under Settings, Domains your real domain has to be the primary one before you connect the app. Otherwise it writes myshopify.com links into the feed and the site cannot be claimed cleanly.
2. Turn off password protection. Under Online Store, Preferences. A password protected shop cannot be crawled, and the errors that follow tend to send people hunting through the feed for a fault that was never there.
3. Create and link your policy pages. Settings, Policies holds your refund policy, terms of service, privacy policy and, for German shops, the Impressum in its own field. Careful here, because Shopify only links saved policies automatically in the footer of the checkout pages, while Google inspects your public storefront. Those links also belong in the footer menu under Online Store, Navigation.
4. Make contact details visible. You need at least a postal address, phone number, email address or contact form on a publicly reachable page.
5. Prepare shipping and returns. Only shipping rates from your general shipping profile sync to Merchant Center. Shopify says that with custom shipping profiles, rates sync incorrectly and cause errors. It is the profile that decides this, whatever rate type you use. Carrier calculated rates are supported by the channel only for certain carriers in Australia, Germany and the US.
6. Connect the app. Sales channels, Google & YouTube, connect your Google account, pick your Merchant Center account, confirm ownership with a six digit code. Shopify says the first product sync takes up to 24 hours, after which Google reviews new products for three to five business days.
7. Enter your return policy in two places. This is the one almost everyone misses. The channel syncs your store domain, product data and shipping and tax settings, but leaves your return rules alone. The simplest route is to enter them directly in Merchant Center under Shipping and returns. Alternatively, Google accepts account level policies from Search Console, structured data on your product pages, or the Merchant API. Once submitted, review takes up to ten business days.
8. Link Merchant Center and Google Ads. In Merchant Center under Settings, Access and services, on the Apps and services tab. You need admin rights in both accounts.

August 26, 2026 and what quietly happens that day

Shopify rebuilt its checkout in stages, and two dates from that rebuild matter for Google Ads. The additional scripts field in your checkout settings has been read only since August 28, 2025, so nothing new goes in there. Shopify Plus shops had to migrate their thank you and order status pages by that date, and since January 2026 Shopify has been upgrading the remaining Plus shops automatically with 30 days' notice. For every other plan the deadline lands on August 26, 2026, after which the field disappears there too.

What happens next is spelled out plainly in Shopify's own documentation. Shops that have not upgraded get upgraded anyway, and every customisation on the legacy pages disappears in the process. Analytics scripts are removed. Shopify attempts a best effort migration to custom pixels, but it guarantees nothing. Google's help centre names the same two dates and describes the consequences as critical limitations for shops that have not migrated. In plainer terms, Google expects a noticeable impact on customer reach and on the effectiveness of your campaigns and bid strategies.

Work through the migration in this order.

1. Take inventory. Shopify generates a personalised upgrade report for each shop. You will find it under Settings, Checkout in the Configurations section. It lists every script it found, sorted by purpose. The tracking and analytics section is the one you care about.
2. Migrate the tags first, upgrade the pages second. Never the reverse, because once you upgrade, the scripts you needed as a reference are already gone.
3. Move Google tracking into the Google & YouTube app. The app detects existing tags and offers to migrate them, and it sets up enhanced conversions and Customer Match in the same pass.
4. Audit your apps. Every app that relies on a script tag needs a current version with an app pixel. Vendors without an app pixel have to be replaced.
5. Place a test order today, not on the deadline. Run a real order through and confirm in Google Ads that the purchase conversion arrives with the right value.

One detail regularly gets confused with the tracking deadline. Shopify Scripts, the old system for discount and shipping logic, ran alongside the new framework until June 30, 2026 according to Shopify's documentation. That concerns pricing logic, not conversion tracking. Confuse the two and you will spend your time migrating the wrong thing.

Conversion tracking: three paths, two of them officially unsupported

On a self hosted shop the purchase tag lives in the template and nowhere else. Shopify gives it three possible homes, and none of them switches off the other two. Duplicate conversions are more common here than clean ones.

Unless you have a compelling reason to do otherwise, take path 1. That has nothing to do with convenience. Too much demonstrably fails inside the custom pixel sandbox.

Why the sandbox breaks so much

Custom pixels run inside their own iframe with no access to the storefront document. A dataLayer.push() from inside the pixel lands in a pixel local data layer, never in window.dataLayer on the page. Lift a snippet from a guide written for a self hosted platform and it will do nothing at all, and the same goes for anything that reads values out of the DOM, captures clicks or needs the real page URL. The page URL detected automatically inside a custom pixel belongs to the sandbox, not to your shop.

Google's own list of what fails in that environment runs long. It covers URL based conversions, enhanced conversions, call conversions, cross domain measurement and Consent Mode URL passthrough, plus automatically collected events and enhanced measurement in GA4, and in Tag Manager the preview mode, many trigger types and all built in variables.

What the reported conversion value actually contains

Google's reference for Shopify event parameters defines the purchase value as subtotal minus discounts, explicitly excluding shipping and tax. Tax and shipping travel as separate parameters. That is the reason the revenue in Google Ads will never match your gross Shopify revenue, and the reason a target ROAS calculated against gross revenue can never be hit.

Compare against net merchandise value after discounts, and work out your own conversion factor from five real orders, once. Two more things shape what you see in the reports.

• Google books conversions to the moment of the ad interaction, Shopify to the order date. With the default 30 day click window, an order placed on July 20 following a click on July 3 shows up in the July 3 row in Google Ads. Comparing the two systems day by day therefore tells you nothing. Compare closed periods of at least 30 days, with a few days of lag.
• On April 30, 2025 Shopify changed how the value is calculated. Shopify announced it as a change rather than a fix. Until then the subtotal in the purchase event deducted product level discounts but left order level ones out. Since then those are deducted too. Wherever a pixel passes that subtotal as the conversion value and order discounts are involved, reported values have been lower ever since. Google recommends adjusting ROAS targets to the corrected numbers.

Where old Google tags hide

When purchases get counted twice, the culprit is rarely the new setup. Usually it is a leftover nobody remembers installing. Google lists four hiding places with the path to each:

• Custom pixel→Settings, Customer eventsopen the pixel, disconnect or delete
• Additional scripts→Settings, Checkoutread only now, disable the script
• checkout.liquid→Shopify Plus onlyduplicate the theme first, then the code editor
• theme.liquid→Online Store, Themes, Edit codeusually in the head, remove whole script blocks

Search the code editor for gtag, UA-, AW- and GTM- and delete whole script blocks. Leave Liquid objects such as {{ content_for_header }} alone or you will break the page.

Transaction ID and counting method do different jobs

Google Ads discards the second conversion of the same conversion action when both carry the same transaction ID. The app sets that ID automatically. Deduplication only fires when the two values match character for character, so the moment one setup sends the order number as 1234 and another as #1234, Google sees two different orders.

If you build your own, use event.data.checkout.order.id instead of the checkout token. The token identifies the checkout. It also sits further forward in the payload, which is why people grab it by mistake.

The most common wrong reaction to inflated numbers is to switch the counting method from Every to One. That treats the symptom instead of the cause. For purchases Google explicitly recommends Every, because every sale represents real value. On One you lose genuine repeat purchases inside the conversion window, and those are the signal bidding needs.

Post purchase upsells and what they do to your value

If you show an offer after the sale, Shopify fires the purchase event on the first upsell page rather than on the thank you page, and it does not fire again afterwards. The value sent to Google therefore comes from the moment before the upsell. Your Shopify order value keeps growing after that, while the figure in Google Ads stays where it was.

If that gap matters to you, send only the incremental amount, as a conversion adjustment carrying the same transaction ID. Never resend the whole order. The old Liquid trick people used on the order status page relied on Additional scripts, so that route is gone.

Consent Mode v2: why the Shopify banner is not enough on its own

Shopify ships its own cookie banner. You find it under Settings, Customer privacy, preconfigured for visitors from the EEA and the UK, and it looks like a finished solution. Then you read the decisive sentence in Shopify's own help centre. The banner governs Shopify specific tools, meaning Shopify cookies and Shopify pixels. Anything running outside that infrastructure keeps running.

So whether that matters depends on your setup. If your Google tracking runs through the Google & YouTube app, the automatic path applies, and Google states that consent mode is automatically enabled once the banner is active. If your tracking runs through your own custom pixel, say with your own Tag Manager container, Shopify does not pass the consent mode signals through automatically. What Shopify documents there is just its own gating, meaning a pixel does not load at all until the permissions it declares are present, and how to read the consent state. Deriving the gtag commands from that is on you. It only happens automatically for app pixels, that is, through the Google & YouTube app or a certified consent app. That gap is what breaks setups in practice. Your banner is up, your Shopify pixels behave, and your own container sits there waiting for a signal that never arrives.

Four Shopify categories, four Google signals

Shopify's Customer Privacy API defines four categories: analytics, marketing, preferences and sale_of_data. On Google's side there are four signals. ad_storage covers advertising cookies and device identifiers, ad_user_data covers sending user data to Google for advertising purposes, ad_personalization covers personalised ads and remarketing, and analytics_storage covers analytics storage. Version 2 added the middle two, and those are the ones missing most often.

There is no official mapping table between the two systems, from Shopify or from Google. The obvious reading, where the Shopify marketing category feeds the three advertising signals and analytics feeds the fourth, circulates in third party documentation but appears in no official documentation. If you are going to rely on it, put it through Tag Assistant first.

Basic or advanced mode

Basic mode holds the Google tags back until the visitor has decided. Nothing leaves the page beforehand, not even the default status, and Google falls back on a generic model. Advanced mode loads the tags immediately with a default of denied and sends cookieless signals even on refusal, which lets Google build a model tuned to your account. Google has never put a number on how much data that modelling needs.

And nobody at Shopify or Google says which of the two the automatic path implements. A German legal assessment would turn on that single point. If the answer matters to you, your own account is the only place to measure it.

What Google actually enforces

Google's EU user consent policy applies to end users in the EEA, the UK and Switzerland, regardless of where you are based. The policy itself states the consequence in general terms, namely that Google may limit or suspend your use of the product and terminate your agreement. The accompanying help centre gets more specific. There Google describes the process as a manual review by a Google team, allows a reasonable amount of time to make changes, and names the account level measure as suspension of ad personalisation, including remarketing, and of conversion measurement. Separate from all that is the standalone Consent Mode v2 requirement for EEA traffic, in force since March 6, 2024.

What reviewers write up: a banner that is missing or inadequate, no affirmative consent action such as a real confirm button, disclosure of data sharing with third parties that does not go far enough, consent signals implemented badly, and cookies that are already set before anyone has agreed to them.

How to work through it

1. Take stock, about 30 minutes. Open Settings, Customer privacy and check whether the banner is active and for which regions. Then Settings, Customer events: which pixels are running at all, and which of them belong to Google?
2. Check each banner surface separately. Shopify lets you control display individually for storefront, cart, checkout and customer account pages. Checkout display arrived in February 2025 and is off by default. If you built your setup before that, sessions that start straight in the checkout run with no prompt at all.
3. Check your customer account domain. Banners and pixels only work on customer account and order status pages if the domain used for them is a subdomain of your storefront domain, for example account.yourshop.com under yourshop.com. If it sits elsewhere, consent given on the storefront does not apply there.
4. Decide on a mode and document it. Basic or advanced, together with your data protection officer. With advanced mode, the cookieless signals belong in your privacy policy.
5. With your own container, set the snippet yourself. Default denied on all four signals as the very first command, before the container loader. Then react to the Shopify event that fires when consent is given and send an update. Use the wait parameter so tags do not fire before an asynchronous consent tool has answered.
6. Check the withdrawal path. Since February 2025 the Shopify banner can be reopened, so visitors can change their decision without clearing cookies. That was not possible before. If your shop still runs an older configuration, this is something you have to add yourself.

Testing, and why the usual test fails here

The standard route is Google Tag Assistant. You open your shop in it, interact with the banner, and check two states in the consent tab. On the earliest consent event all four signals must appear with a default of Denied. After clicking accept all, the most recent event must show Granted. An empty tab means consent mode is not implemented on the page at all.

On Shopify that test can lie to you. Shopify states explicitly that Tag Assistant's troubleshooting feature is not compatible with custom pixels and does not detect Google tags inside one. Track through a custom pixel and the standard tool reports a failure where there is none. That is what pushes people into a second tracking implementation, and from there straight into the double counting described further up.

You will get further with Shopify's own test feature for custom pixels under Settings, Customer events, with the network tab in your browser developer tools, and with the conversion action diagnostics in Google Ads, where the timestamp of the last recorded conversion tells you whether anything is arriving at all.

Product data: where Google pulls every field from

On a self hosted shop you maintain an export format and decide which column comes from where. On Shopify the app decides that, and your only control is the fields on the product. Convenient enough, as long as you know which field feeds which Google attribute. Neither Shopify nor Google publishes that mapping in full, so the overview below marks what is documented and what is established convention.

• Title→titlerequired, 150 characters max, must match the landing page
• Description→descriptionrequired, 5,000 characters max, no links
• First media item→image_linkrequired, no watermarks, borders or promotional text
• Variant price→pricerequired, including VAT in Germany and most of Europe
• Inventory→availabilityrequired, values in_stock, out_of_stock, preorder, backorder
• Variant barcode field→gtinnothing else populates it, see below
• Category field→google_product_categoryvia Shopify's standard product taxonomy
• Product type field→product_typeconvention, free text field, not the same as Category
• Vendor field→brandconvention, only documented by feed vendors, verify in your account
• Product ID plus variant ID→id and item_group_idevery variant is its own offer, the product ID groups them

The English options requirement

Shopify's channel requirements contain one sentence that costs non English shops real money. Product options, including variants, need to be in English to sync. That covers option names and values, so Color rather than Farbe, Size rather than Größe, with English values to match.

A German shop naturally sets its options in German and then wonders why variants are missing or whole products never arrive. Switch the options to English in the backend and translate the storefront display through Shopify Markets or the translation tooling, or use a feed app that rewrites them on export. The first route calls for caution. Renaming options can touch variant URLs and theme logic, so test before you commit.

GTIN lives in the barcode field, and only there

On the product page each variant has a barcode field. That value, and nothing else, is submitted as gtin. Not the SKU, not a metafield. Only real manufacturer numbers with 8, 12, 13 or 14 digits and a valid check digit are accepted.

The classic slip is the internal item number landing in the barcode field because it happens to be empty, and Google rejects such offers. Google explicitly bars you from making up, guessing or borrowing values from similar products. The other constant failure is a spreadsheet import that strips the leading zero from an EAN and turns a 13 digit number into a 12 digit one with an invalid check digit. Format the barcode column as text when importing a CSV.

Own brand without a GTIN, the route Google itself recommends

The reflex is to set identifier_exists to no and watch the warnings disappear. Google sees it differently. If you manufacture the product yourself or have no official brand, the recommended route is to use your own shop name as the brand and supply a self assigned mpn. Do that and you do not need identifier_exists at all. The boundary matters. That freedom only applies if you really are the manufacturer. If you resell someone else's goods, you may not invent a part number and have to use the manufacturer's.

The attribute is meant for real exceptions such as handmade one offs, vintage and antique goods, books published before 1970 and certain spare parts. One rule always holds. Never send a GTIN and identifier_exists set to no at the same time.

Applying the attribute to every GTIN warning costs you reach. Products that do have an identifier you fail to submit cannot be matched to Google's catalogue and get limited exposure. Nothing in the interface calls that an error. The status reads as limited performance, and shops live with it for months.

Category, product type and collections do different jobs

The Category field belongs to Shopify's standard product taxonomy and feeds Google's product category. Set it as deep as you can. Google asks for the most specific category available and at least two levels of hierarchy. As soon as you set it, Shopify unlocks matching category metafields such as target gender, age group, colour and size, which happen to be the fields Google requires for apparel and accessories.

The Product type field, by contrast, is a free text field feeding Google's product_type. That is your own categorisation, and a useful segmentation dimension in Google Ads, because you can cut it however your campaign logic needs.

Collections, finally, are not a category tree. A product can sit in any number of them and has no single membership. People arriving from a self hosted shop tend to bring the assumption that category equals feed structure equals campaign structure. On Shopify that assumption does not hold. You build campaign structure from product type, tags and custom labels.

Price, tax and shipping

Google requires prices including VAT for every target country except the US and Canada, where tax must stay out of the price attribute. Europe is the default case here, not the exception. In Shopify this hangs on a single switch under Settings, Taxes and duties, labelled include sales tax in product price and shipping rate. Set it wrong for a European shop and you produce systematically wrong feed prices. If you sell into both Europe and the US, the country specific presentation is also controlled through your Markets settings.

On shipping, the limit already mentioned applies. Only rates in the general profile sync. Google's own advice for borderline cases rarely gets quoted. When in doubt, state slightly higher shipping costs in Merchant Center than on your site. Underreported costs trigger an error, overreported ones do not.

With discounts, what matters is where they take effect. Anything that only applies at checkout, so discount codes and automatic discounts, does not show up as a reduced price in the feed. Those belong in a separate promotions data source in Merchant Center, which Google reviews separately and serves as a special offer link. The product feed will never show you a sale price that lives in the cart. Promotional prices you set through the compare at price, where you lower the regular price and keep the old one as a reference, do travel normally as price and sale price.

That split produces the most common price conflict. A discount app lowers the amount shown on the product page while the feed still carries the full one. If a promotion is meant to run permanently, build it into the product price rather than into an app that only does the maths in the cart.

Dates worth putting in your calendar

The disapprovals that hit Shopify shops most often

Diagnostics live in Merchant Center under Products in the Needs attention tab. The prioritised fixes toggle is on by default and hides issues with low impact, which is where missing identifiers, missing size and colour attributes and weak titles pile up. None of that gets a product disapproved. It just gets the product shown less often. The shop looks green and loses impressions anyway, so make a point of turning the filter off at every routine check.

Six patterns come up again and again on Shopify.

Two sets of structured data on one page

Shopify themes output product markup. SEO, review and rich snippet apps often output a second set. If one block says in stock and the other says out of stock, or the theme lists every variant in a single block, Google reads the wrong value and automatic item updates go wrong along with it.

Run your product page through the rich results test and count how many product objects it emits. There should be one, carrying the price of the preselected variant. Recheck after every theme update, because updates overwrite manual edits without warning.

What SEO guides get wrong about robots.txt

Shopify's default robots.txt contains a dedicated block for the Google Ads crawler, introduced by Shopify's own comment line noting that this crawler ignores the file unless it is named specifically. Google's crawler documentation explains why. This bot ignores the global wildcard user agent and therefore has to be addressed by name. It does evaluate wildcards inside the path rules perfectly normally. Plenty of SEO tutorials suggest replacing the template with hand-written lines. Do that and the block disappears, and Shopify can no longer ship its regularly updated default rules.

The other common piece of advice is to pre-emptively block the variant parameter in the URL. Shopify demonstrably does not block that parameter as of July 2026, and blocking it makes exactly the variant landing pages your feed links to unreachable. Extend robots.txt.liquid additively, use the provided Liquid objects, and fetch the live file after every change.

What happens if it does escalate

For account level problems Merchant Center usually sends a warning email first, with a deadline of either 7 or 28 calendar days depending on the violation rather than a blanket 28 days. During that window products keep running, sometimes with reduced visibility. Google then checks again, and only if something is still unresolved does suspension follow. Every item is pulled from shopping ads and free listings. For severe violations the warning is skipped and suspension is immediate. A review request takes three to seven business days.

Campaign structure at €1,500 and at €20,000

The Google & YouTube app builds exactly one campaign type, Performance Max. The app is not being stingy. It is following where the channel has gone. According to Tinuiti's Digital Ads Benchmark Report for the fourth quarter of 2025, among merchants running both Performance Max and standard shopping campaigns, Performance Max accounted for 62% of Google shopping spend and 61% of shopping revenue. The revenue share sits slightly below the spend share. Tinuiti reports on the accounts it manages, so the sample leans heavily American. The trend still points one way.

Everything beyond that single campaign gets built in Google Ads, not in Shopify. So it pays to know which rules Google has written down and which ones the industry has invented for itself.

What Google really says

• Performance Max has no minimum budget. Every euro figure you read on the subject is a rule of thumb. Google ties its recommendation to your own costs instead. The average daily budget should be at least three times the cost per conversion of the selected conversion actions. In practice, campaigns may spend up to twice the average daily budget on any given day, capped at 30.4 times that daily budget per month. Compare a single day against your budget and you learn nothing. Compare the month and you do.
• Six weeks of leaving it alone. In its best practices Google recommends running new campaigns for at least six weeks while avoiding frequent changes to budget, bid strategy and targeting, because those reset the learning period. The learning period itself, according to the setup documentation, typically runs one to two weeks and in some cases up to six. After significant changes, expect one to two weeks or at least one full conversion cycle before things stabilise.
• Target ROAS needs volume. For search and shopping campaigns Google specifies at least 15 conversions in the past 30 days. The same page sets no separate threshold for Performance Max, so carrying that number across is an inference, though a sensible one. An efficiency target only means something once there is stable conversion value underneath it.
• Do not split without a reason. Google's own wording asks you to avoid creating separate campaigns unless they require different budgets, CPA or ROAS targets, or settings for business reasons. No documented budget threshold marks the point where segmentation starts to pay.
• Assets come with a countable minimum. Each asset group requires at least 3 headlines, 1 long headline, 2 descriptions and one image each in landscape and square format. The ceilings sit at 15 headlines, 5 long headlines, 5 descriptions and 20 images per format. Portrait images and video are optional. Google recommends filling the slots but does not require it. If you do not supply your own video, Google builds one out of whatever the asset group already contains.

Your brand belongs in its own campaign

This is the one structural point even small accounts should not skip. Google documents a four step hierarchy here. At the top sits an exact match keyword identical to the query in a search campaign, where spelling corrections count but plurals and synonyms do not. Below that come phrase and broad keywords and search themes identical to the query, then a relevance based prioritisation, and only between equally ranked candidates does ad rank decide. A brand search campaign on exact match brand terms is therefore the most reliable way to keep brand queries out of Performance Max. Two things limit it. The hierarchy does not apply to shopping formats, and Performance Max can still serve if your search campaign is budget limited or more narrowly targeted.

Brand exclusions belong in the Performance Max campaign alongside it, under Campaigns, Settings, Additional settings. Google explains why they do more than negative keywords. They automatically block the brand name, common misspellings and related subsidiary brands, and brand lists match across languages and writing systems. The exclusion applies to search, shopping and YouTube search inventory. If you want brand shopping ads to keep running, there is a dedicated option for that in the same setting.

Three budget tiers for orientation

The split below is how we work, and Google prescribes none of it. It follows the logic that every additional campaign needs conversion volume of its own.

Two Shopify specifics around audiences

Shopify Audiences is not available to European shops. The myth that it is available has survived a remarkable number of debunkings. The feature requires Shopify Plus, Shopify Payments and a merchant operating in and selling to the United States or Canada. A great many guides omit that condition because they are written for the US market.

The route that does work runs through customer lists, which come from Shopify into Google Ads through a direct connection. The path runs through Tools, Data Manager, Connect product, Shopify. It requires the Google & YouTube app to be installed and a linked account that is not a manager account. Plan around a limit of 25 lists per import run. Those lists are what feed audience signals and the new customer goal in Performance Max.

That new customer goal comes in two modes. One bids higher for new customers without excluding existing ones. The other excludes existing customers entirely. For most shops the first is the right call, because at low volume the second cuts your conversion volume in half.

The most expensive default settings in Google Ads

Campaigns created from the Shopify interface or from Merchant Center never show you these switches. They live in Google Ads only, and they sit at factory settings.

• Location option→set to presence or interestCampaigns, Settings, Locations, Location options. For shops usually presence
• Search partners→included by default in search campaignsfull placement reporting since August 2025, so measure before you decide
• Auto applied recommendations→no master switch, one checkbox eachCampaigns, Recommendations, Auto apply
• Counting method→website actions default to Everykeep purchase on Every, move newsletter and contact to One
• Attribution model→data driven or last clickthe four older models are gone, do not switch mid season
• Conversion window→30 days after click, 1 day after impressionup to 90 days depending on the conversion source, Google suggests at least 7, never retroactive
• Placement exclusions→belong at account levelsince March 2024 they also cover search partners and Performance Max
• Negative keywords→up to 10,000 per Performance Max campaignthey only affect search and shopping inventory, account level caps at 1,000

A word of caution on the auto applied recommendations. Nowhere does Google document that they are on by default, and the widespread claim that Google switches them on behind your back has no primary source behind it. What is documented is what can be enabled, and that is the list worth reading. Two of those items do real damage on a small budget. Automatically adding broad match keywords is one, expanding reach to search partners the other. Both widen your reach without touching your budget, so the same money spreads across more traffic of weaker quality.

What is left once the setup is done

Shopify takes the feed mechanics off your hands. What it does not take is the data quality behind them, so that is what you are left maintaining. You stop watching an export and start looking after products instead. The engineering shrinks, the workload does not.

The jobs that keep coming back:

• A quick Merchant Center scan every day or two. The product errors can wait; the account level issues cannot. A suspension realistically costs one to two weeks of revenue, because the fix, the review and the campaign ramp-up stack on top of each other.
• The full list once a month. Switch the prioritised fixes filter off, otherwise the quiet data gaps stay invisible.
• Search terms and exclusions. Since 2025 Performance Max has a real search terms report at the same granularity as search and shopping campaigns. Google advises weighting conversions higher than clicks there, because click through rate and CPM are misleading in Performance Max.
• Add assets before you swap them. Google explicitly advises against removing asset groups based on cost per conversion or ROAS. Fill every slot first, then start replacing.
• Ration your changes. Google gives no maximum step size for target ROAS changes; Smart Bidding reacts to large and small adjustments alike, immediately. For promotional days Google gives 30% increments as an example, with one to two conversion cycles of waiting in between. So after a target change, wait one to two conversion cycles before you judge, and do not touch the target twice within one cycle. Handle promotional periods through seasonality adjustments of one to seven days rather than budget jumps. Google advises against seasonality adjustments longer than 14 days, because Smart Bidding already accounts for seasonality on its own.

There are more built in tools than most people use. Attribute rules in Merchant Center transform feed data without you touching products, though they sit behind an add on and are no longer called feed rules since the rebuild. Google Ads scripts are the most powerful free option for budget alerts and feed watchdogs, but somebody has to maintain the JavaScript. Product Studio generates image assets right inside the Shopify app.

What no tool will take off you: policy and suspension cases, allocating budget by contribution margin, decisions about range and promotions, and knowing when to leave a learning campaign alone.

Where Cloudginny fits in

That routine is what we built Ginny for: an AI agent for Google Ads aimed at ecommerce shops in the €500 to €20,000 monthly budget range. Ginny reads your account and tells you in plain language where budget is leaking. Nothing changes until you approve it. The analysis is not deeper than yours, it just happens more often. Your manual audit lands once a week; the monitoring never stops.

You can ask Ginny anything about your account, have campaigns analysed, optimised or built from scratch. The questions worth asking are collected in the prompt library at cloudginny.com/ginny-prompts. The strategic calls stay with you. On the operational eighty percent, automation buys back hours you currently spend clicking.

Questions Shopify merchants keep asking

The essentials in five points

• Shopify handles the engineering. The data quality is on you, since everything Google knows about your products comes from the product fields in your shop.
• Duplicate tracking costs more than missing tracking, and it is the mistake we see most often on Shopify. There are three ways to install a tag, and switching one on does not switch the others off.
• If you are not on Shopify Plus, your conversion tracking has a deadline. August 26, 2026.
• Products published only to the Online Store, and product options that are not in English. Both quietly cost you reach, and neither shows up as an error anywhere.
• You set the account up once, then you keep optimising it for as long as it runs. If you would rather not take that on yourself, hand it off to an agency or let an AI agent like Ginny run the day-to-day.

This guide is updated regularly. Last reviewed: July 2026.

Sources

• Official Shopify sources: help centre articles on the Google & YouTube channel requirements, product syncing, custom pixels and how to test them, customer privacy, and the thank you and order status page upgrade. Plus the developer documentation for the Web Pixels API, the checkout completed event, the Customer Privacy API and the standard product taxonomy, along with the Shopify changelog.
• Official Google sources: Google Ads help on tag migration for Shopify, transaction IDs and counting methods, Performance Max optimisation, brand exclusions, auto applied recommendations and AI Max. Merchant Center help on the product data specification, data sources, disapproval reasons and policy enforcement. Developer documentation on Consent Mode, the Shopify event parameters and the Merchant API.
• Legal framework: the official text of section 25 of the German TDDDG, plus Google's EU user consent policy and the audit documentation that goes with it in the Google Ads help centre.
• Market data: Shopify's first quarter 2026 results release, Tinuiti Digital Ads Benchmark Report Q1 2026, Contentsquare Digital Experience Benchmark 2026, IRP Commerce Market Data.
• Cloudginny: internal account data from ongoing ecommerce account management.