Google shopping launch checklist

Run this in the week before a Shopping campaign goes live, and again after any restructure. It covers the conversion tracking the campaign will report against, how to split the feed so budget follows margin rather than popularity, the campaign settings that decide who sees your products, and what to watch in the first seven days of spend before you judge anything.

Tracking before any spend

Confirm the purchase conversion action records order value, not a flat number.Why: Shopping bidding optimises towards value, and a fixed conversion value makes every product look equally profitable.How: Open the conversion action, set it to use a transaction specific value, and place a test order to confirm the amount that lands matches the order total.
Mark purchase as the only primary conversion goal for the campaign.Why: Secondary actions such as newsletter signups pull smart bidding away from revenue when they are counted as primary.How: Set purchase as the primary action for the account, and move add to cart and signup actions to secondary so they report without steering bids.
Choose an attribution window you can leave alone for a quarter.Why: Changing the lookback window later makes every before and after comparison in the account meaningless.How: Pick a click through window that matches your typical consideration time, write it in the account notes, and leave it while the campaign learns.
Deduplicate purchase events between the store platform and the ads tag.Why: Double counted orders inflate reported revenue and push bidding into overpaying for the same sale twice.How: Use one source for the purchase event, either the native platform integration or the site tag, and remove the duplicate snippet from the theme.

Feed segmentation

Add a custom label that groups products by gross margin band.Why: Margin, not revenue, decides how much a click on a given product is genuinely worth to the store.How: Work out margin per product in a spreadsheet, bucket it into high, medium and low, and push the band into custom label zero on the feed.
Add a second custom label for stock depth or season.Why: Spending on a product with two units left wastes clicks that other products could have converted.How: Map units on hand or season into custom label one, and use it to pause or lower bids on thin stock before the campaign does it for you.
Exclude products you never want advertised with an excluded destination rule.Why: Loss leaders, staff items and clearance oddments drag down account level performance and confuse the bidding model.How: Tag those products in the store, map the tag to excluded destination for shopping ads in the feed, and review the list monthly.
Fill product type and Google product category for the whole catalogue.Why: These are the fields you subdivide campaigns by, and blank values collapse everything into one everything else bucket.How: Use the product type field for your own store taxonomy, and set the Google product category at the deepest level that fits each product.

Campaign structure and settings

Split campaigns by margin band rather than by brand alone.Why: Separate budgets stop low margin products absorbing spend that high margin products would have converted.How: Create one campaign per margin band, subdivide the product group by product type inside each, and give every campaign its own daily budget.
Set the network options deliberately rather than accepting the default.Why: Search partners and display expansion quietly change the mix of traffic a Shopping campaign buys for you.How: Open the network settings before launch, decide on partners as a choice, and record it so a later performance change can be traced back to it.
Restrict location targeting to the places you actually ship to.Why: Clicks from countries your shipping settings do not cover can never turn into an order.How: Set target locations to the regions in your shipping profile, and choose presence rather than presence or interest so you pay for people who are there.
Build a starting negative keyword list from queries that never buy.Why: Shopping matches on product data, so a single word in a title can pull in research and repair queries.How: Add negatives for free, repair, manual, hire and second hand terms, then keep them in a shared list you can apply to future campaigns.
Set a daily budget that can buy a meaningful number of clicks.Why: A budget that funds only a few clicks a day will never gather the data smart bidding needs to work.How: Divide the daily budget by your expected cost per click, and either raise the budget or narrow the product set until the daily click count is workable.

The first week after launch

Check impressions the morning after launch to confirm items are serving.Why: Zero impressions on day one is almost always a feed or targeting problem rather than a bidding problem.How: Open the products tab, filter to items with no impressions, and check each one is approved, in stock and inside the targeted product group.
Read the search terms report every day for the first week.Why: Early irrelevant queries are cheap to block now and expensive to leave running for a month.How: Sort search terms by cost, add negatives for anything that cannot lead to a sale, and note the queries that convert so titles can be tuned towards them.
Hold bids and structure steady for the first two weeks unless money is being wasted.Why: Constant edits restart the learning period and make the early data impossible to read at all.How: Limit changes to adding negatives and pausing clear waste, and diary the structural review for after the learning period has finished.
Reconcile reported revenue against orders in the store admin.Why: A gap between the two means tracking is wrong, and every optimisation decision after it will be wrong too.How: Export the week's orders, filter by source, and compare the totals before making any budget decision based on reported return.

Questions

Enough to buy a meaningful number of clicks daily at your own cost per click. If clicks cost a pound and the budget buys four of them, the campaign has nothing to learn from.

Either works, but smart bidding needs conversion history. If the account has almost no recorded purchases, start with manual bids or maximise clicks with a bid cap until purchases build up.

Shopping matches on product data rather than keywords, so you cannot split brand out with keywords alone. Use campaign priority and negative lists if you want brand traffic in its own budget.

Your plan is free.

A person reads it. You keep the plan. No logins, no call first.

Google shopping launch checklist

Tracking before any spend

Feed segmentation

Campaign structure and settings

The first week after launch

Questions

How much should a new Shopping campaign spend per day?

Should I launch with manual bids or smart bidding?

Do I need a separate campaign for brand searches?

Related

Your plan is free.