Skip to content
All playbooks
PlaybookNovember 3, 202612 min read

Tracking first, then seven levers, ranked by how much they move sales. Most checklists start with bids, which sit fifth.

Google Shopping Optimization: What Moves Sales

Google Shopping optimization for stores: product titles, feed fixes, bids, search terms and product splits, ranked by how much each one moved real sales.

Written and reviewed by

Published Last checked

Google Shopping optimization means getting more sales from the same budget. Do it in this order. First make sure sales are counted right and fix the feed errors that stop products from showing. Then rewrite the titles of your most-seen products and split winners from losers with custom labels. Block searches that never buy. Only then touch bids.

I'm Christopher Krassnig. I run ZenoX Media, a Google Ads agency that has run the ads for 400+ store accounts, and I built Scaley. Below is how we work those accounts. The order is ours. Google does not publish one.

Why the Order Matters

Shopping ads have no keywords. Google reads your product data and decides which searches your products can show for (Google Ads Help).

So the feed decides which searches you can show for, and the bid decides how much you pay to show there. A bigger bid on a bad title buys more of the wrong clicks.

That is why most 30-point Shopping checklists give every item the same weight. They put "adjust bids" next to "fix GTINs" as if both matter the same. They don't.

New to Shopping? Read what Google Shopping ads are and what a product feed is first.

Google Shopping Campaign Optimization, Step by Step

This is the order we work in: tracking first, then seven levers, from the biggest effect on sales to the smallest.

#LeverWhat it fixesImpact (what we see)Effort
0Sales trackingEvery number you judge byEverything sits on itLow, once
1Feed errorsProducts that can't show at allHighLow
2Product titlesWhich searches you show forHighMedium
3Product splitWhere the budget goesHighMedium
4Search terms and negativesClicks that never buyMediumLow, weekly
5Bids and targetsHow hard you pushMediumLow
6ImagesWho clicksMediumMedium
7PriceWhether they buy after the clickHigh if you are overpricedHard

The impact column is what we see in our own accounts, not a measured percentage.

Lever 0 is not optional. If Google counts the same order twice, or misses half your orders, every other fix is judged on fake numbers. Check tracking before anything else. The optimization checklist covers how.

Lever 1: Fix Feed Errors First

Open Merchant Center. Go to the product diagnostics. Look for products that are disapproved or limited.

A disapproved product shows nowhere. Zero impressions. No bid fixes that. And it is often your best seller, because best sellers get edited most and break most.

What we check, in this order:

  1. Disapproved products. Sort by how much each one sold last month. Fix the top ones first.
  2. Missing GTINs. Google says that if a product has a GTIN, "including this in your product data can increase the performance of the product" (Merchant Center Help). If you sell branded goods, send the GTIN.
  3. Price and stock mismatches. The price in the feed must match the price on the page. Shopify sales and discount apps break this a lot. On Shopify? Our page on Google Shopping for Shopify covers the feed setup.
  4. Wrong categories. A good Google product category helps Google know what the item is.

Check for new errors every few days. One bad feed sync can switch off a winner on a Friday and you find out on Monday.

For the tools that help fix a feed, ZenoX Media has a guide to Google Shopping feed optimization tools. Their post below shows why a weak feed gets blamed on Performance Max. Our own take on what to automate in the feed is in Google Merchant Center automation.

Lever 2: Google Shopping Title Optimization

The title is the closest thing Shopping has to a keyword. It is also what the shopper reads.

Google's rules, from the title spec (Merchant Center Help):

  • You get up to 150 characters.
  • "Users will usually notice only the first 70 or fewer characters of your title, depending on screen size."
  • "Put the most important details first."
  • Add the details that set each variant apart, like color or size.

Put the words people search in the first 70 characters. Use the rest for detail.

A formula that works for most stores:

Store typeTitle formula
FashionBrand + product type + gender + material + color + size
Home decorProduct type + style + material + size + color
ElectronicsBrand + model + product type + key spec
BeautyBrand + product + key benefit + size
Unbranded or private labelProduct type + main feature + material + size + color

Two examples of what a rewrite looks like:

  • Before: "Aurora Dress". After: "Women's Linen Midi Dress, Sleeveless, Sage Green, Size M".
  • Before: "The Nordic". After: "Oak Wood Floor Lamp, Scandinavian Style, 150 cm, Linen Shade".

Nobody searches "Aurora Dress". People search "linen midi dress". The old title could only win searches for its own name.

Where to start: don't rewrite 2,000 titles. Pull the 50 products with the most impressions and the lowest click-through rate. Those are shown a lot and skipped a lot. Rewrite those first. Then check the search terms they show for two weeks later.

If you change titles with a supplemental feed or feed rules in Merchant Center, your store titles stay the same. Only the ad changes.

Lever 3: Split Products so Winners Get the Money

One Shopping campaign with "All products" in one group is the most common setup we see. It is also where the most money goes to waste.

In that setup, your best seller and 400 products that never sold share one budget. Google spreads the money. The dead products eat clicks every day.

The fix is custom labels. Merchant Center gives you five, custom label 0 to 4, and Google says to "use custom labels to group products for reporting and bidding" (Merchant Center Help). Shoppers never see them.

The split we use most:

LabelWhich productsWhat they get
WinnerSold above break-even ROAS in the last 30 daysMost of the budget, own campaign or own group
TestNew, or not enough data yetSmall capped budget
LoserSpent enough to judge, sold under break-evenLow bid or excluded
SeasonalOnly sells part of the yearSwitched on for its season

In Standard Shopping, campaign priority makes this sharper. The highest priority campaign bids first for a shared product, and "If the campaign with the highest priority runs out of budget, the next lower priority campaign will place the bid" (Google Ads Help).

If you run Performance Max next to Standard Shopping, give each one its own label. The full split is in Performance Max vs Standard Shopping.

Re-label every month. A winner from August can be a loser by November. For the full method, see how to tier products by margin with custom labels.

Lever 4: Search Terms and Negatives

You can't pick keywords in Shopping. You can block them.

Open the search terms report every week. Sort by cost. Look for searches that spent and never sold. The common ones in stores:

  • "free", "DIY", "how to make", and "cheap" unless you sell on price
  • A competitor's brand name you can't beat on price
  • The wrong product type. A lamp shop showing for "lamp oil".
  • Repair, parts, used, rental

Add them as negative keywords. Use phrase or exact match so you don't block good searches by accident. Long lists sorted by type are in negative keyword examples for online stores, and how to read the report is in the search terms report guide.

Don't block a search after 3 clicks. Give it enough clicks to judge against your conversion rate. If you sell 1 in 50 clicks, 10 clicks with no sale proves nothing.

Lever 5: Bids and Targets

Bids come this late on purpose. Before you touch them, know your break-even ROAS. It is 1 divided by your margin after product cost, shipping and fees. A 40% margin means break-even at 2.5x. Under that, every sale loses money.

Then:

  • Test products with no sales history: manual CPC or a tight budget cap. Smart Bidding has nothing to learn from yet.
  • Products with steady sales: target ROAS, set near your real goal. Set it far above what the products can do and spend drops off a cliff.
  • Move targets in small steps. Big jumps reset the learning. We move a target a little, then wait at least a week.

For a rough sense of where stores land, here is the median ROAS per account across ZenoX Media managed accounts, by niche, from 30 August to 2 November 2026 (65 days).

NicheAccounts (N)Median ROASMiddle half (p25 to p75)
Fashion2922.19x1.72x to 2.50x
Home decor662.42x2.06x to 2.74x
General store292.43x2.04x to 2.63x

Source: ZenoX Media managed accounts, 30 August to 2 November 2026 (65 days), median ROAS per account, rows with 10+ accounts only. Your break-even matters more than any benchmark. A 2.2x store with 60% margins makes money. A 2.6x store with 30% margins does not.

Lever 6: Images

The image is what gets the click. On a Shopping results page, a row of products sits side by side.

Google's rules (Merchant Center Help):

  • Show a "clean view of the item for sale".
  • No watermarks, no promo stamps, no borders.
  • The product should fill "no less than 75%, but not more than 90%" of the image.
  • Google recommends images of 1500 x 1500 pixels or more. It has also announced a minimum of 500 x 500 pixels for all products, starting 31 January 2027.

What we test after the rules are met: a plain white image against a lifestyle shot as the main image. Same product, same price. Then compare click-through rate over two weeks.

Lever 7: Price

Shopping puts your price next to everyone else's. If you sell the same product as five other stores and you are 30% more expensive, no title saves it. You get the click from people who did not look at the price, and they leave.

Ask two questions. Is this the same product other stores sell? Is my price far above theirs? If both are yes, either compete on price, bundle it so it is no longer the same offer, or stop paying for its clicks.

What Rarely Moves Sales

Things we see stores spend hours on that almost never change the result:

  • Changing bids every day by a few cents
  • Rewriting the product description field before the title is fixed
  • Restructuring the whole account every two weeks, so nothing gets time to learn
  • Adding every optional feed attribute before the disapprovals are fixed

How Scaley Helps with Shopping

Most of this starts with finding where the money goes to waste.

Scaley chat answer listing products under the break-even ROAS and search terms that spent and never converted, with account details blurred

That is a real Scaley answer, blurred. Scaley has three levels:

  • Scaley Consultant answers from screenshots and CSV exports you paste in. It connects to nothing.
  • Scaley Media Buyer reads your live Google Ads account read-only. Ask "which Shopping products are under break-even this month?" or "which searches spent and never sold?" and you see the waste right away. It never changes anything, and it only looks when you ask.
  • Scaley Suite finds the fixes, like a new label split or a list of negatives, and makes each change only after you say yes to that one change.

If you want to find the waste in your Shopping campaigns without giving anyone write access, start with Media Buyer. You can start the free trial and pick Media Buyer when the app asks for a plan, or compare the three levels on pricing.

If you would rather a team ran Shopping for you, that is what ZenoX Media does.

Where These Facts Come From

Checked on 2026-11-03:

Frequently Asked Questions

Written by

I founded Scaley AI and run ZenoX Media, the Google Ads agency behind it. We have $200M+ in client revenue tracked through our campaigns, across 400+ store accounts. Scaley Media Buyer puts those answers in a chat.

Try it on your account

See where Scaley would cut waste on your Google Ads

Connect read-only in under five minutes and ask where the money went. Every level starts with a 3-day free trial. Card required. Cancel anytime.

Start the 3-Day Free Trial

Ask the Live Account Before You Raise the Budget

3-day free trial. Card required. Cancel anytime