How to Optimise Your Google Shopping Campaigns by Impact

The short answer: optimise in order of impact

To optimise your Google Shopping campaigns, work through them in this order. Check that your conversion tracking is accurate. Fix feed errors and disapprovals. Improve product titles and attributes. Clean up search terms with negative keywords. Set targets based on your real margins. Only then refine campaign structure and bids. Most accounts waste effort adjusting bids while the feed and the data underneath are broken.

The order matters because Shopping has no keywords to bid on. Google decides when to show your products by reading your product feed, and decides how much to bid by learning from your conversion data. If either of those is wrong, every change you make further down the list is built on sand.

This guide walks through each priority in turn, explains what to check and why, and finishes with a routine you can follow every day, week, month and quarter.

1. Conversion tracking: make sure Google is learning from real sales and real values.
2. Feed health: get products approved and eligible to show.
3. Feed quality: titles, categories and attributes that match how people search.
4. Search terms and negatives: stop paying for clicks that will never buy.
5. Targets: set ROAS goals from your margins, not gut feel.
6. Structure: group products by how they actually perform.
7. Product and pricing decisions: price competitiveness, stock and promotions.

Priority 1: make sure your conversion data is telling the truth

Almost every Shopping campaign now runs on Smart Bidding, either through Performance Max or a Standard Shopping campaign with a target ROAS. Smart Bidding is only as good as the data it learns from. Feed it inflated revenue and it will spend confidently on products that lose you money.

Before you change anything else, check these points:

• Purchases are counted once. Duplicate tags (for example, one from a plugin and one added manually) can double your reported sales. Compare a week of Google Ads conversions with the orders in your shop's back end. They will never match exactly, but they should be close.
• Values are consistent. Decide whether your conversion value includes VAT and shipping, then stick to it. Mixing approaches makes ROAS figures meaningless.
• Only purchases are primary conversions. If newsletter sign-ups, add-to-baskets or page views are set as primary actions, Google will optimise towards them. Move them to secondary so they are reported but not bid on.
• Enhanced conversions are switched on. This helps recover conversions that cookie restrictions would otherwise hide, which gives Smart Bidding more to work with.
• Refunds and cancellations are understood. If you have a high return rate on certain categories, the reported ROAS for those categories is flattering you. You may need to aim higher for them.

This is the least glamorous job on the list and the one most often skipped. It is also the one that makes the biggest difference, because every automated decision Google makes depends on it.

Priority 2: fix feed health in Google Merchant Centre

A product that is disapproved cannot show. A product with a "limited" status shows less than it could. So the second job is getting as much of your catalogue eligible as possible.

Open the products and diagnostics sections in Google Merchant Centre and work through the issues, starting with the ones affecting the most products or your best sellers.

The issues that cause most problems

• Price or availability mismatches. If your feed says £49.99 and in stock, but the landing page says £54.99 or out of stock, Google will flag it. Repeated mismatches can lead to account-level warnings. The fix is usually a more frequent feed update or correct structured data on product pages.
• Missing or invalid GTINs. For branded products, the barcode number helps Google match your item to searches and to its product catalogue. Products without valid identifiers often perform poorly or are limited.
• Missing shipping and returns information. Set these up properly at account level. Shoppers compare delivery costs and return terms, and Google uses this information in the listing.
• Image problems. Promotional overlays, watermarks, placeholder images or images that are too small can all cause disapprovals.
• Landing page problems. Broken links, pages that require a login, or pages that redirect somewhere unexpected.

Set up email notifications in Merchant Centre so you hear about new problems quickly. A feed that was healthy last month can break overnight when a developer changes a template or a plugin updates.

Priority 3: improve feed quality, especially titles

Once products are eligible, the quality of the feed decides which searches they appear for and how appealing they look next to competitors. This is where many shops have the most room to improve.

Product titles

The title is the single most important attribute. Google uses it heavily to decide relevance, and it is the first thing a shopper reads. Many shop platforms pull a short, branded product name into the feed, which tells Google very little.

Compare these two titles for the same table:

• "The Hartley"
• "Hartley Solid Oak Extending Dining Table, 6 Seater, 160cm"

The second one contains the words people actually type: oak, extending, dining table, 6 seater. Put the most important terms near the start, because titles are often cut short on smaller screens. For branded goods, the brand usually belongs at the front. For your own products, the product type usually matters more than your brand name.

We cover this in more detail in our guide on how to optimise Google Shopping feed titles for search intent, including title formulas for different product types.

Other attributes worth getting right

• Google product category and product type. These help Google understand what you sell. Product type is your own categorisation and is also useful for structuring campaigns later.
• Colour, size, material, gender and age group. These matter a lot in clothing, furniture and homeware, where shoppers filter and search by them.
• Descriptions. Plain, factual and useful. Include the details a buyer needs, not marketing slogans.
• Additional images. Lifestyle shots and alternative angles give Google more options and help shoppers decide.
• Sale price. Use the sale price attribute properly so the original and discounted prices both show.

Use a supplemental feed if your platform is awkward

If your shop platform makes it hard to edit feed titles without changing the product names on your website, use a supplemental feed in Merchant Centre. It lets you override specific attributes for specific products, which means you can rewrite titles, add custom labels or fix categories without touching the shop itself.

Custom labels

There are five custom label fields. They do not show to shoppers, but they let you tag products for campaign structure. Common uses are margin band, price band, seasonality, best sellers and new arrivals. Setting these up now makes Priority 6 far easier.

Priority 4: search terms and negative keywords

Because Shopping is driven by the feed, Google will match your products to searches you never intended. Some of those searches will be close enough to convert. Others will spend money and never produce a sale.

Review the search terms report every week. Look for patterns rather than individual oddities. Typical wasted traffic includes:

• Searches containing "free", "jobs", "how to", "manual", "repair" or "second hand" when you only sell new goods.
• Competitor brands you do not stock.
• Product types you do not sell that share a word with ones you do.
• Searches far too broad to convert, such as single generic words, if the data shows they spend without selling.

Add these as negative keywords. Account-level negative lists are useful for terms you never want to appear for in any campaign, such as "jobs" or "free".

Search terms in Performance Max

Performance Max historically gave much less visibility over search terms than Standard Shopping. Google has been adding more search term reporting and negative keyword controls to it, so check what your account currently shows. Brand exclusions are also available, which is useful if you want to stop Performance Max taking credit for people who were already searching for your shop by name.

Do not over-negate

The opposite mistake is just as common. Adding a negative for every search that has not converted yet can choke a campaign, especially on lower-traffic accounts. A search term with a handful of clicks and no sale is not proof of anything. Wait until there is enough spend to judge, and be careful with broad negatives that might block good searches too.

Priority 5: set targets from your margins, not gut feel

A common question is "what ROAS should I aim for?" There is no universal good number. The right target depends on your margins, and many businesses set targets without ever working this out.

Work out your break-even ROAS

Break-even ROAS is the return at which your ad spend exactly cancels out the gross profit on the sale. The formula is simple: divide 1 by your gross margin.

A worked example. You sell a lamp for £50. After the cost of the product, your gross profit is £20, which is a 40% margin. Your break-even ROAS is 1 divided by 0.4, which is 2.5, or 250%. If you spend £20 on ads to sell that lamp, you have made no profit at all on it.

Gross margin	Break-even ROAS	What that means
20%	500%	£5 of sales needed for every £1 of ad spend to break even
25%	400%	£4 of sales per £1 spent
33%	300%	£3 of sales per £1 spent
40%	250%	£2.50 of sales per £1 spent
50%	200%	£2 of sales per £1 spent

Your target should sit above break-even if you want profit from the first sale, or around break-even if you are deliberately buying new customers because you know they come back and buy again. That is a business decision, not a Google Ads setting, and it is worth making consciously.

Change targets gradually

Smart Bidding needs time and data to adjust. Large, sudden jumps in target ROAS often cause spend to collapse or behave erratically. Move in modest steps, then give the campaign a couple of weeks before judging the result. Weekly target changes based on a few days of data are one of the most common ways Shopping accounts get stuck.

Low conversion volume

If a campaign has only a small number of sales each month, a strict ROAS target can starve it. In that case, starting with "maximise conversion value" without a target, then adding one once there is enough history, is often more sensible. Combining products into fewer campaigns also helps, because the bidding has more data to learn from.

Priority 6: structure campaigns around how products actually perform

Structure is where many people start, and it is genuinely important. But it only works once the feed and data are sound, which is why it sits sixth here.

The aim is simple. Group products that should be treated differently so they can have different targets and budgets. A single campaign containing everything forces one target onto products with very different margins and demand.

Performance Max or Standard Shopping?

Factor	Performance Max	Standard Shopping
Where ads show	Shopping, Search, YouTube, Display, Discover, Gmail and Maps	Shopping placements only
Control	Less granular, more automated	More control over products, bids and search terms
Reporting	Improving, but still less transparent	Full search terms and product reporting
Data needs	Works best with reasonable conversion volume	Can be run manually with low volume
Good for	Shops with solid tracking and a healthy feed	Testing, tight budgets, or when you need visibility

Many accounts use both. For example, Performance Max for the main catalogue, with Standard Shopping used for specific products where you want closer control. Be aware that when the same products are in both, Google no longer automatically favours Performance Max. The two campaigns compete on Ad Rank, so whichever ad ranks higher for a given search will show. Plan the split deliberately, and make sure the targets and budgets in each campaign reflect what you actually want to happen.

Segment by performance and margin

Once a campaign has a few months of data, look at product-level results and sort them into groups. Custom labels make this easy to apply.

Refresh these groups monthly. Products move between them as seasons, stock and prices change. For a deeper look at splitting campaigns and setting bids, see our guide on optimising Google Shopping bids and structure.

Keep structure as simple as the data allows

Splitting too finely is a real risk. Every extra campaign divides your conversion data, and Smart Bidding works worse on thin data. A small shop might only need two or three campaigns. Only add more when each one will have enough sales to learn from.

Priority 7: product and pricing decisions

Shopping ads put your price right next to your competitors' prices. No amount of campaign work makes an uncompetitive price attractive.

• Check price competitiveness. Merchant Centre has price insight reports that show how your prices compare with others selling the same product. If you are consistently more expensive on a key item, that will show up in click-through and conversion rates.
• Keep stock accurate. Promoting products that are out of stock, or about to be, wastes money and annoys shoppers. Make sure availability updates reach the feed quickly.
• Use promotions. Merchant promotions let you show offers such as a discount code or free delivery directly on your listings, which can make your ad stand out.
• Collect product reviews. Product ratings can appear on Shopping ads once you have enough reviews and submit them through an approved route. Genuine reviews from real customers only.
• Compete on delivery and returns. Free or fast delivery and clear returns terms are shown to shoppers and often influence the click.
• Do not forget free listings. Your Merchant Centre products can also appear in unpaid Shopping listings. A healthy feed benefits both.

Your routine: what to check and when

Optimisation is not a one-off project. Here is a routine that works for most small and medium-sized shops. It assumes a few hours a month, more in busy seasons.

How often	What to check	Why
Daily (a few minutes)	Spend pacing, Merchant Centre alerts, any sudden drop in impressions or sales	Catch broken feeds, tracking failures and runaway spend early
Weekly	Search terms and negatives, product-level spend, stock levels on top products	Stop waste and make sure budget goes to products you can actually sell
Monthly	ROAS targets against margins, product segment refresh, title improvements, price insights	Keep targets realistic and products in the right groups
Quarterly	Conversion tracking audit, campaign structure review, seasonal planning, feed attribute review	Prevent slow drift and prepare for busy periods

If you only have time for one thing each week, make it the search terms review and a quick look at Merchant Centre diagnostics. Those two catch most problems before they become expensive.

Edge cases worth planning for

Small catalogues

If you sell a few dozen products, heavy segmentation does more harm than good. Keep one or two campaigns and put the effort into titles, images, pricing and reviews. Every product matters, so feed quality has a direct effect on results.

Very large catalogues

With thousands of products, you cannot hand-check every title. Work on rules instead: title formulas by category, supplemental feeds for overrides, and custom labels applied in bulk. Focus manual effort on the products that bring in most of your revenue.

Seasonal products

Smart Bidding learns from recent history, so it can be slow to react when demand changes sharply. For short, predictable events such as a planned sale, Google Ads offers seasonality adjustments that tell Smart Bidding to expect a change in conversion rate. Use them sparingly and only for short periods. For longer seasons, start building budget a few weeks before demand peaks.

New accounts or new shops

With no conversion history, Smart Bidding has nothing to learn from. Expect the first weeks to be a learning period. Keep targets loose, structure simple and budgets realistic, and judge results over a sensible timeframe rather than day by day.

Products with very different margins

If some products make 15% margin and others 60%, one target will either overspend on the low-margin items or underspend on the high-margin ones. This is the clearest case for splitting campaigns by margin band using custom labels.

Mistakes that undo good work

• Changing too much at once. If you change targets, structure and titles in the same week, you will not know what worked.
• Judging on too little data. A bad weekend is not a trend. Look at weeks and months.
• Trusting reported ROAS blindly. If tracking includes VAT, shipping or orders later refunded, the figure is flattering you.
• Restructuring constantly. Each rebuild resets some of the learning. Make structural changes deliberately and then leave them alone.
• Ignoring the website. Slow pages, unclear delivery costs or a fiddly checkout will lose sales that the ads paid for.
• Letting the feed go stale. Prices, stock and product ranges change, and the feed has to keep up.

If you are not sure which of these apply to your account, a free Google Ads audit is a sensible starting point. It will show where the biggest gaps are, so you can tackle them in order.

Common questions

How often should I change my Shopping bids or targets?

For Smart Bidding campaigns, monthly is usually enough, with modest changes each time. Frequent changes stop the algorithm from settling. The exception is when something has clearly broken, such as a tracking failure or a sudden collapse in spend.

How long does it take to see results from optimisation?

Feed fixes such as getting products approved can have an effect quickly, once Google reprocesses the feed. Changes to targets and structure take longer because Smart Bidding needs time to relearn. Give a significant change at least a couple of weeks before judging it, longer on lower-traffic accounts.

Should I run Shopping alongside Search campaigns?

Often, yes. Shopping is strong for product searches, while Search campaigns can capture brand searches, service-style queries and terms where a text ad works better. They need to be managed together so they are not competing for the same traffic. This is part of good Google Ads management rather than a separate job.

Is my feed or my campaign the problem?

If products have few impressions, start with the feed: titles, categories, approvals and identifiers. If products get plenty of clicks but few sales, look at price, landing pages, delivery costs and whether the searches are relevant. If spend is high and the return is low across the board, check tracking and targets first.

Can I do this myself or do I need help?

Many owners can run the weekly routine themselves once it is set up properly. Where help usually pays off is the initial tracking audit, feed restructuring and margin-based campaign setup, which are technical and easy to get wrong. If you are comparing options, our Google Ads pricing page sets out what is included at each level.