The order matters: account, then product data, then campaign
Most stores set this up backwards. They connect everything, launch a campaign the same afternoon, and then spend three weeks wondering why half the catalogue never shows. Do it in this order instead.
First, the Merchant Center account and its requirements. This is where the store's identity, website, shipping and returns live. Nothing shows anywhere until this is complete and the site is claimed.
Second, the product data. Shopify sends what your product records say, so the quality of your titles, images, prices, availability and identifiers decides what your ads can show for. Fixing this before spending is the difference between Shopping working and Shopping being expensive.
Third, the campaign. Link a Google Ads account and start with a single, readable structure. You cannot judge a feed problem and a bidding problem at the same time, so give yourself one variable.
Each step takes hours rather than days, but the review and approval waits in between mean the whole setup usually spans a week or two. Plan for that rather than promising a launch date you will miss.
Complete the Merchant Center requirements before anything else
Install the Google channel app in Shopify and connect it to a Google account, then either link an existing Merchant Center account or create one. The app handles the product sync, but the account level requirements are yours to complete.
You need your website claimed and verified against the same domain your store runs on. If you have recently changed domain or moved from a myshopify address to a custom one, check this specifically, because a stale claim silently blocks everything.
You need business information that matches what is on the site, including a contact method. Merchant Center expects a store to be reachable.
You need shipping settings that reflect what you actually charge, and they must match the rates your checkout applies. A mismatch between the shipping shown in an ad and the shipping charged at checkout is a common cause of suspension.
You need a returns policy that is on the site and accurately described in the account.
And your site needs the basics Google looks for: a working checkout, visible prices, clear contact details and secure pages. Fix these before you spend a day debugging individual product errors.
The sync only carries what your product records already say
There is no clever mapping happening. If your Shopify title is a bare product name, that is the title Google gets, and it will match far fewer searches than a title carrying brand, product type and the attribute people search on.
Work through the fields that matter. Titles should read like a search rather than like an internal label. Descriptions should carry the specifics, since Google uses them to understand the product. Images should be clean product shots without added text, badges or watermarks. Prices and availability must match the storefront exactly, including any currency or market differences.
Identifiers are the field most stores get wrong. If the product has a manufacturer barcode, put it in the barcode field on the variant so it maps to GTIN. If you make the product yourself and there is no barcode, set the brand and use the MPN or SKU field rather than inventing a number.
Google product category and product type help the system understand where you sit. Set them for your main categories rather than leaving them blank across the whole catalogue.
Then read the diagnostics in Merchant Center and clear every error before spending. Warnings can wait. Errors mean products are not showing.
Link Google Ads and start with one campaign you can read
Link the Ads account to Merchant Center from within Merchant Center, and accept the link on the Ads side. Then decide what to launch.
A first campaign should cover a limited set of products rather than the whole catalogue. Pick the twenty to fifty products with the healthiest margin and the best product pages. A campaign with three thousand products and a small budget spreads itself so thin that no product gets enough data to learn from, and you end up paying for browsing on your worst sellers.
Start on manual or a cost led bid strategy while you are gathering conversion data, and move to a return based strategy once conversions are being recorded reliably. Setting a return target on day one with no history gives the system nothing to work from.
Make sure conversion tracking is working before the first click. Check that purchases are recorded with the correct value, that you are not counting the same order twice through both the Shopify channel and a separate tag, and that the currency is right.
Set the campaign to the countries you can actually ship to, at the rates you configured.
What to check in the first two weeks
Look at the search terms report before you look at the return on spend. It tells you what people typed to reach your products, and in a new account it is usually full of queries you would never have chosen. Broad, generic terms burning budget on your cheapest products is the standard first problem, and it is fixed with negative keywords and by tightening which products are eligible.
Check the product level report. If a handful of products are taking most of the spend, decide whether they deserve it. Often they are the cheapest items with the thinnest margin, because low prices attract clicks.
Check impressions against approvals. If a product is approved but getting nothing, the title and the category are usually the reason.
Watch the account for new disapprovals weekly, especially after a bulk price change, a theme update or an app that rewrites titles. Feeds break quietly.
And give it enough time. Two weeks tells you whether the plumbing works. Judging profitability needs a longer window and enough conversions that the numbers are not being driven by a handful of orders.
Worked example, not a result
A first two weeks on a limited product set
- Products synced
- 1,200
- Products approved
- 1,140
- Products in the first campaign
- 40
- Daily budget
- 50
- Clicks in two weeks
- 900
- Search terms report entries reviewed
- top 200
- Negative keywords added
- 35
Worked example, not a result