1) Get the foundation right: Merchant Center health + product data quality
Merchant Center Next changes how you navigate—but not what wins auctions
If you’ve been running Shopping for years, the biggest operational change recently is the platform experience: Merchant Center Next is rolling out broadly, and some familiar areas have new names. For example, “Feeds” are now “Data sources,” and diagnostics are surfaced more directly in a “Needs attention” workflow. The practical takeaway is simple: spend less time hunting menus and more time acting on the highest-impact issues Merchant Center is prioritizing for you.
Fix “Needs attention” issues first (they silently cap visibility and ROAS)
Before you touch bids, budgets, or creative, make sure products are eligible to serve. Product-level warnings can limit performance; disapprovals stop products from showing entirely. Also watch for pre-emptive disapprovals when price/availability don’t match between your site and your product data—these are especially common during promotions, rapid inventory changes, or when structured data on the site is inconsistent.
A very pragmatic rule: if you’re troubleshooting a sudden drop in impressions or a spike in CPC, always check product approval status and policy/quality issues the same day. Shopping performance can look like a bidding problem when it’s actually a feed eligibility problem.
Optimize the attributes that most directly influence matching, relevance, and segmentation
For Shopping, your “keywords” are your product attributes. Titles are the single biggest lever for matching and click-through rate, so treat them like ad copy: keep them accurate, aligned to the landing page, and focused on core descriptors people actually use (brand, product type, key variant like size/color, and model where relevant). Avoid promotional fluff in titles. If you’re using AI-generated titles, use the structured title attribute designed for that purpose rather than forcing AI text into your standard title field.
Next, make your identifiers and brand data clean enough that the system can confidently understand what you’re selling. Brand should reflect the true manufacturer (and be consistent in one language/alphabet), and products that have a manufacturer-assigned GTIN should include it—this isn’t optional for “some brands” anymore; it’s a general requirement when a GTIN exists.
Finally, don’t overlook shipping. In retail accounts, inaccurate shipping settings create downstream conversion-rate problems (surprise costs at checkout) and can also trigger data-quality friction. Keep shipping configuration as close as possible to what your website charges, and understand that item-level shipping attributes can override account-level settings—this is powerful, but it can also create inconsistent behavior if your rules aren’t intentional.
2) Choose the right campaign type and structure it for control (without fighting automation)
Performance Max vs Standard Shopping: pick based on control needs, not habit
Most retailers should assume Performance Max is the primary Shopping format because it can access more inventory and formats across channels, while Standard Shopping remains more narrowly focused on Shopping placements. The “right” choice depends on whether you need the additional reach and cross-channel formats of Performance Max, or whether you specifically want the more constrained behavior and levers typical of Standard Shopping.
In practice, I still see Standard Shopping used successfully for specific situations—like tightly controlled testing, highly segmented bidding approaches, or when you want a simpler, shopping-only footprint. But for growth-oriented ecommerce with solid conversion tracking and enough volume, Performance Max is usually where the platform’s latest capabilities concentrate.
Use listing groups to control product coverage (and custom labels to make segmentation scalable)
Inside a Performance Max asset group, listing groups determine which products are eligible to serve. Think of them as your product coverage map: you can include/exclude products, and you can subdivide by attributes like category, brand, item ID, condition, product type, channel, and custom labels.
Here’s the mistake I see most often: advertisers create an overly complex listing group tree with hundreds or thousands of subdivisions, then wonder why management becomes impossible. There is a hard limit per asset group, and very large listing-group counts are explicitly not a best practice—performance can suffer. If you want durable control, build your segmentation strategy around a small set of meaningful custom labels (for example margin tiers, seasonality, price bands, or inventory status), then subdivide listing groups using those labels rather than exploding the tree by item ID.
Be intentional with Final URL expansion and asset group themes
With a Merchant Center data source attached, Performance Max can send traffic to product URLs included in listing groups, and it can also expand beyond your provided final URL domain depending on whether Final URL expansion is enabled. This matters because it changes where the system is allowed to land shoppers—and landing page quality is a major driver of conversion rate, which directly impacts Smart Bidding’s ability to hit your goals.
Asset groups should be themed. If you sell multiple categories (for example, running shoes and winter coats), separate asset groups usually outperform a “one giant asset group” approach because your creative, audience signals, and merchandising messages can stay coherent. If you don’t provide video, the system may auto-generate one; if that’s not acceptable for your brand, upload your own videos to maintain creative control.
3) Make Smart Bidding work: measurement, bidding strategy, and guardrails
Pick a bidding goal that matches how your business makes money
For ecommerce, value-based bidding is typically the most profitable path once conversion value tracking is reliable. “Maximize conversion value” aims to drive as much revenue/value as possible within your budget, and you can add a ROAS target when you need ROI discipline. If you don’t have trustworthy values (or every conversion is roughly equal), “Maximize conversions” is the better fit, optionally with a CPA target.
Be careful with targets early on. Aggressive ROAS targets are one of the fastest ways to “throttle” a Shopping campaign—especially if your feed is new, your conversion volume is low, or your price competitiveness is mixed. A common expert approach is to start with a more flexible target (or no target), stabilize volume, then tighten the target once performance is consistent.
Use the right diagnostics for budget-limited questions
When you’re running automated bidding, some legacy columns and habits can mislead you. For example, for Maximize conversions, certain impression share budget metrics aren’t compatible in the way many advertisers assume. If you’re trying to understand whether budget is constraining growth, use the tools designed to simulate and forecast budget opportunity rather than relying on incompatible columns.
Add guardrails the modern way: brand exclusions first, negative keywords sparingly
Performance Max can now be controlled with negative keywords for Search and Shopping inventory, including at campaign or account levels, but this is a restrictive lever that can harm performance if you use it like a traditional Search campaign negative list. For most retailers, brand exclusions are the cleaner approach for preventing spend on your own brand terms (including variants and misspellings), while negative keywords should be reserved for clear brand-safety needs or truly irrelevant queries you never want.
4) An optimization loop that actually moves the needle: search terms, merchandising, and ongoing cleanup
Turn search term visibility into feed and creative improvements
Performance Max reporting has evolved to provide stronger visibility into what people searched before seeing your ads, bringing capabilities that historically felt “Search-only” into the Shopping workflow. Use this data as a merchandising feedback loop: identify the terms that convert at strong rates and make sure your titles, product types, imagery, and promotional messaging align with those intents. When you find queries that clearly don’t fit your business, exclude them carefully via negatives or brand controls—and monitor performance afterward to ensure you didn’t block valuable discovery.
Use feed rules to scale improvements without rebuilding your ecommerce platform
If your catalog is large, manual edits don’t scale. Attribute rules (and related feed tooling) let you improve titles, fill gaps, and standardize formatting at scale by prepending/appending structured details or transforming values conditionally. This is one of the fastest ways to raise relevance across thousands of SKUs without asking engineering to rebuild your data layer.
Quick diagnostic checklist (use this before changing bids)
- Check product eligibility and prioritize fixing “Needs attention” items that are disapproved or at risk due to price/availability mismatches.
- Confirm shipping configuration matches checkout behavior and that item-level overrides aren’t unintentionally conflicting with account-level rates.
- Verify listing groups include the right products, and avoid bloated subdivision trees; lean on custom labels for scalable segmentation.
- Review search terms and apply exclusions only where you have high confidence (brand exclusions first, negatives second).
- Evaluate whether your bidding goal (conversion value vs conversions) matches your tracking maturity and business model before tightening ROAS/CPA targets.
If you follow this sequence—eligibility and data quality first, then structure and coverage, then bidding and guardrails, then reporting-driven iteration—you’ll avoid the most common “busy work optimizations” and focus on the levers that consistently improve Shopping visibility and conversion efficiency over time.
Let AI handle
the Google Ads grunt work
Let AI handle
the Google Ads grunt work
Optimizing Google Shopping ads is often less about constantly tweaking bids and more about getting the fundamentals right: clean Merchant Center health (fixing “Needs attention” issues, price/availability mismatches, and shipping settings), strong feed attributes (especially titles, brand, and GTINs), and a campaign structure that fits your goals (Performance Max for broader reach when tracking is solid, or Standard Shopping for more controlled tests). From there, dialing in listing groups and a small set of durable custom labels helps you steer performance without creating unmanageable segmentation, while search term insights and careful brand exclusions/negative keywords keep spend focused. If you want a lightweight way to turn this kind of checklist into ongoing, concrete actions, Blobr connects to Google Ads and runs specialized AI agents that continuously spot opportunities (like landing page alignment or ad asset improvements) and surface prioritized recommendations—so you can stay in control while reducing the day-to-day manual work.
1) Get the foundation right: Merchant Center health + product data quality
Merchant Center Next changes how you navigate—but not what wins auctions
If you’ve been running Shopping for years, the biggest operational change recently is the platform experience: Merchant Center Next is rolling out broadly, and some familiar areas have new names. For example, “Feeds” are now “Data sources,” and diagnostics are surfaced more directly in a “Needs attention” workflow. The practical takeaway is simple: spend less time hunting menus and more time acting on the highest-impact issues Merchant Center is prioritizing for you.
Fix “Needs attention” issues first (they silently cap visibility and ROAS)
Before you touch bids, budgets, or creative, make sure products are eligible to serve. Product-level warnings can limit performance; disapprovals stop products from showing entirely. Also watch for pre-emptive disapprovals when price/availability don’t match between your site and your product data—these are especially common during promotions, rapid inventory changes, or when structured data on the site is inconsistent.
A very pragmatic rule: if you’re troubleshooting a sudden drop in impressions or a spike in CPC, always check product approval status and policy/quality issues the same day. Shopping performance can look like a bidding problem when it’s actually a feed eligibility problem.
Optimize the attributes that most directly influence matching, relevance, and segmentation
For Shopping, your “keywords” are your product attributes. Titles are the single biggest lever for matching and click-through rate, so treat them like ad copy: keep them accurate, aligned to the landing page, and focused on core descriptors people actually use (brand, product type, key variant like size/color, and model where relevant). Avoid promotional fluff in titles. If you’re using AI-generated titles, use the structured title attribute designed for that purpose rather than forcing AI text into your standard title field.
Next, make your identifiers and brand data clean enough that the system can confidently understand what you’re selling. Brand should reflect the true manufacturer (and be consistent in one language/alphabet), and products that have a manufacturer-assigned GTIN should include it—this isn’t optional for “some brands” anymore; it’s a general requirement when a GTIN exists.
Finally, don’t overlook shipping. In retail accounts, inaccurate shipping settings create downstream conversion-rate problems (surprise costs at checkout) and can also trigger data-quality friction. Keep shipping configuration as close as possible to what your website charges, and understand that item-level shipping attributes can override account-level settings—this is powerful, but it can also create inconsistent behavior if your rules aren’t intentional.
2) Choose the right campaign type and structure it for control (without fighting automation)
Performance Max vs Standard Shopping: pick based on control needs, not habit
Most retailers should assume Performance Max is the primary Shopping format because it can access more inventory and formats across channels, while Standard Shopping remains more narrowly focused on Shopping placements. The “right” choice depends on whether you need the additional reach and cross-channel formats of Performance Max, or whether you specifically want the more constrained behavior and levers typical of Standard Shopping.
In practice, I still see Standard Shopping used successfully for specific situations—like tightly controlled testing, highly segmented bidding approaches, or when you want a simpler, shopping-only footprint. But for growth-oriented ecommerce with solid conversion tracking and enough volume, Performance Max is usually where the platform’s latest capabilities concentrate.
Use listing groups to control product coverage (and custom labels to make segmentation scalable)
Inside a Performance Max asset group, listing groups determine which products are eligible to serve. Think of them as your product coverage map: you can include/exclude products, and you can subdivide by attributes like category, brand, item ID, condition, product type, channel, and custom labels.
Here’s the mistake I see most often: advertisers create an overly complex listing group tree with hundreds or thousands of subdivisions, then wonder why management becomes impossible. There is a hard limit per asset group, and very large listing-group counts are explicitly not a best practice—performance can suffer. If you want durable control, build your segmentation strategy around a small set of meaningful custom labels (for example margin tiers, seasonality, price bands, or inventory status), then subdivide listing groups using those labels rather than exploding the tree by item ID.
Be intentional with Final URL expansion and asset group themes
With a Merchant Center data source attached, Performance Max can send traffic to product URLs included in listing groups, and it can also expand beyond your provided final URL domain depending on whether Final URL expansion is enabled. This matters because it changes where the system is allowed to land shoppers—and landing page quality is a major driver of conversion rate, which directly impacts Smart Bidding’s ability to hit your goals.
Asset groups should be themed. If you sell multiple categories (for example, running shoes and winter coats), separate asset groups usually outperform a “one giant asset group” approach because your creative, audience signals, and merchandising messages can stay coherent. If you don’t provide video, the system may auto-generate one; if that’s not acceptable for your brand, upload your own videos to maintain creative control.
3) Make Smart Bidding work: measurement, bidding strategy, and guardrails
Pick a bidding goal that matches how your business makes money
For ecommerce, value-based bidding is typically the most profitable path once conversion value tracking is reliable. “Maximize conversion value” aims to drive as much revenue/value as possible within your budget, and you can add a ROAS target when you need ROI discipline. If you don’t have trustworthy values (or every conversion is roughly equal), “Maximize conversions” is the better fit, optionally with a CPA target.
Be careful with targets early on. Aggressive ROAS targets are one of the fastest ways to “throttle” a Shopping campaign—especially if your feed is new, your conversion volume is low, or your price competitiveness is mixed. A common expert approach is to start with a more flexible target (or no target), stabilize volume, then tighten the target once performance is consistent.
Use the right diagnostics for budget-limited questions
When you’re running automated bidding, some legacy columns and habits can mislead you. For example, for Maximize conversions, certain impression share budget metrics aren’t compatible in the way many advertisers assume. If you’re trying to understand whether budget is constraining growth, use the tools designed to simulate and forecast budget opportunity rather than relying on incompatible columns.
Add guardrails the modern way: brand exclusions first, negative keywords sparingly
Performance Max can now be controlled with negative keywords for Search and Shopping inventory, including at campaign or account levels, but this is a restrictive lever that can harm performance if you use it like a traditional Search campaign negative list. For most retailers, brand exclusions are the cleaner approach for preventing spend on your own brand terms (including variants and misspellings), while negative keywords should be reserved for clear brand-safety needs or truly irrelevant queries you never want.
4) An optimization loop that actually moves the needle: search terms, merchandising, and ongoing cleanup
Turn search term visibility into feed and creative improvements
Performance Max reporting has evolved to provide stronger visibility into what people searched before seeing your ads, bringing capabilities that historically felt “Search-only” into the Shopping workflow. Use this data as a merchandising feedback loop: identify the terms that convert at strong rates and make sure your titles, product types, imagery, and promotional messaging align with those intents. When you find queries that clearly don’t fit your business, exclude them carefully via negatives or brand controls—and monitor performance afterward to ensure you didn’t block valuable discovery.
Use feed rules to scale improvements without rebuilding your ecommerce platform
If your catalog is large, manual edits don’t scale. Attribute rules (and related feed tooling) let you improve titles, fill gaps, and standardize formatting at scale by prepending/appending structured details or transforming values conditionally. This is one of the fastest ways to raise relevance across thousands of SKUs without asking engineering to rebuild your data layer.
Quick diagnostic checklist (use this before changing bids)
- Check product eligibility and prioritize fixing “Needs attention” items that are disapproved or at risk due to price/availability mismatches.
- Confirm shipping configuration matches checkout behavior and that item-level overrides aren’t unintentionally conflicting with account-level rates.
- Verify listing groups include the right products, and avoid bloated subdivision trees; lean on custom labels for scalable segmentation.
- Review search terms and apply exclusions only where you have high confidence (brand exclusions first, negatives second).
- Evaluate whether your bidding goal (conversion value vs conversions) matches your tracking maturity and business model before tightening ROAS/CPA targets.
If you follow this sequence—eligibility and data quality first, then structure and coverage, then bidding and guardrails, then reporting-driven iteration—you’ll avoid the most common “busy work optimizations” and focus on the levers that consistently improve Shopping visibility and conversion efficiency over time.
