Service · Google Shopping Feed
Your product feed isn't admin. It's part of your media strategy.
As Google Ads automates, advertisers hold fewer direct controls. Google increasingly decides who sees the ad, when, what you bid, where it appears and which product gets the impression. That makes the information you give the system more valuable, and for ecommerce most of it starts in the feed.
Your website is built for customers. Your product feed is built for machines. Google doesn't walk your navigation, sit with your buying team or attend the Monday trading meeting. It receives structured product data, and from that it tries to understand your catalogue.
01 / Two layers
Most feed optimisation stops at layer one. We want both.
Layer 1 · Product information
What is this product?
Titles that reflect how customers search rather than how merchandising named the product. GTINs, MPNs and identifier consistency. Product types that preserve your own hierarchy. Accurate Google product categories. Category-relevant attributes. Images that actually explain the product before the click. Descriptions with information density, not keyword density. Price and availability accuracy. And Merchant Center health, because a beautiful feed Google refuses to serve is of limited commercial use.
Layer 2 · Commercial information
What is this product worth to the business?
Margin, contribution, stock depth, stock age, sales velocity, customer value, lifecycle, seasonality, commercial priority and the product's BOI® job. Carried into Google Ads through custom labels, supplemental feeds and rules, so classification changes campaign structure, targets, budgets and reporting.
Availability tells Google whether it can sell the product. Inventory tells us whether we should push it.
Longer isn't better
Feed optimisation occasionally becomes a competitive sport in making titles unreadably enormous. The most valuable information should appear early. More words isn't optimisation. Better information is.
Stock without velocity misleads
500 units selling 100 a day is very healthy. 500 units selling 3 a week is a potential problem. Context changes what the number means.
Labels aren't the strategy
Adding HIGH_MARGIN to a product does not magically improve Google Ads. Someone still has to decide what counts as high margin, and whether margin alone justifies more investment. Data classification supports judgement, it doesn't replace it.
Rules need governance
A badly designed automated feed rule can efficiently damage thousands of products at once. Automation remains impressively impartial about whether the instruction was sensible.
Category matters
Fashion needs size, colour and variant-level stock context. Automotive needs fitment and compatibility. Beauty needs shade, format and concern. Furniture needs material, dimensions and room. There isn't one universal title formula.
The feed doesn't fix everything
A better title cannot rescue an uncompetitive price, poor reviews, slow delivery or a weak offer. Sometimes feed analysis shows the advertising is working fine and the product simply isn't competitive. That's useful information too.
02 / Where feeds fail
Six quiet failures sitting in most ecommerce Shopping feeds
01 · Bland titles
"Blue T-Shirt" loses to "Nike Dri-FIT Running T-Shirt - Men's Blue". Query match collapses, CTR halves, CPC rises.
02 · Missing GTIN / MPN
The single biggest cause of suppressed impressions. Google quietly demotes the SKU; you never see the auctions you lost.
03 · Inconsistent product_type
Hierarchy mismatch breaks Performance Max asset group targeting and listing group splits in Shopping.
04 · One blanket image
Merchant Center Next rewards multiple angles. Single hero shots leak eligibility on lifestyle and discovery surfaces.
05 · No custom labels
No labels means no BOI®. The bidder cannot tell a hero from a clearance line and treats them identically.
06 · Native feed only
Shopify and Magento default exports miss critical fields. Without supplemental feeds, GTIN-brand-MPN triplets stay broken.
How we fix it
Each failure has a documented repair, written into the feed
Every line below is something a client can read in Merchant Center after the engagement. No theory, no slides - the feed itself is the deliverable.
01 · Problem
Bland titles
Before
Blue T-Shirt
After
Nike Dri-FIT Running T-Shirt - Men's Blue - UK Medium
The fix
We rewrite titles using a brand-type-attribute-variant formula tuned to the queries your category actually receives, then A/B the new structure in a supplemental feed so the impact is measured, not assumed.
02 · Problem
Missing GTIN / MPN
Before
GTIN blank, identifier_exists set to false
After
GTIN repaired, MPN backfilled, brand triplet validated
The fix
We pull a coverage report, source missing identifiers from manufacturer data or barcode lookups, and write the repaired values via a supplemental feed so the source system stays untouched.
03 · Problem
Inconsistent product_type
Before
Mixed depth, free-text categories, no hierarchy
After
Five-level taxonomy aligned to Google product categories and PMax asset groups
The fix
We rebuild product_type as a controlled hierarchy so listing groups, asset groups, and value rules can be split cleanly by category, margin, and BOI® job.
04 · Problem
One blanket image
Before
Single hero shot, white background, no lifestyle
After
Hero plus three to five additional_image_link entries covering angles, scale, and context
The fix
We brief the imagery, push additional_image_link via supplemental feed, and validate eligibility across Shopping, lifestyle, and discovery surfaces in Merchant Center Next.
05 · Problem
No custom labels
Before
custom_label_0 through 4 empty
After
BOI® job, margin band, stock cover, hero flag, new-customer flag
The fix
We assign every SKU a commercial job, write it to custom_label, and wire value rules so the bidder pays what each SKU is actually worth to the business.
06 · Problem
Native feed only
Before
Default Shopify or Magento export, no enrichment
After
Primary feed plus supplemental feed for identifiers, labels and rewrites
The fix
We layer a supplemental feed on top of the native export so commercial signals and repaired attributes flow through without touching your ecommerce platform.
How we engineer a feed
Five layers from raw catalogue to commercially led bid signal
No black box. Every layer is a documented change you can read in Merchant Center and in the campaign.
01
Diagnose
9-point audit covering approval rate, attribute completeness, title structure, image quality, GTIN coverage, taxonomy, supplemental coverage, label readiness, and feed-bid alignment.
02
Engineer
Rewrite titles for query intent, repair GTIN / MPN / brand triplets, normalise product_type, attach supplemental feeds for the fields your platform misses.
03
Label
Assign every SKU a BOI® job - Scale, Profit, Protect, Recovery, Gateway - then write that job and the margin band to custom_label fields the bidder can target.
04
Bid
Wire value rules and asset group structure to the labels. The bidder now sees commercial role, not just price and conversion value.
05
Monitor
Weekly review of stock cover, margin drift, returns, and approval health. Jobs and labels shift with the business; the feed is treated as a living asset.
How we decide
BOI®
BOI® (Bid On Intent) gives every SKU one commercial job at a time. Bidding follows that job rather than a blended target that quietly averages winners and leakers.
- Scale
- Profit
- Protect
- Recovery
- Gateway
Platforms
Works with the stack you already run
Platform
Shopify
Native integration via Google channel; supplemental feed for GTIN/MPN backfill and BOI® labels.
Platform
WooCommerce
Direct connector or middleware; rule-built titles for variant-heavy catalogues.
Platform
BigCommerce
Native channel app; supplemental feed for attribute gaps and custom_label injection.
Platform
Magento / Adobe Commerce
Custom export for complex catalogues; middleware for rule-heavy stores.
Platform
Headless / custom stack
Direct Content API for Shopping; supplemental layer for commercial signals.
What it means for clients
A repaired feed changes the P&L, not just the dashboard
Ranges are taken from JudeLuxe engagements over the last 24 months. Your starting point dictates where in the range you land - we model both before signing.
+18-42%
Impression share recovered
Once GTIN, MPN and product_type are repaired, suppressed SKUs re-enter the auction and previously invisible demand becomes addressable.
-15-30%
Average CPC
Stronger titles and tighter query match mean the bidder competes on intent-rich searches, not broad mismatches you were quietly funding.
+22%
Blended Shopping ROAS
Typical lift on a feed that moves from native-only to engineered, measured 60-90 days after BOI® labels are wired into value rules.
Cleaner P&L
Margin-weighted spend
Budget follows commercial role - heroes protected, clearance accelerated, recovery lines bid down - so the catalogue stops cross-subsidising loss-makers.
For the founder
Spend stops cross-subsidising loss-making SKUs. Cash conversion improves because clearance and ageing stock are bid for velocity, not vanity ROAS.
For the ecommerce manager
Hero SKUs are protected, new launches get learning room, and PMax stops spraying budget across categories you cannot defend on margin.
For the marketing director
Reporting moves from blended ROAS to margin-weighted contribution by SKU job. You can answer the board's question: "what did this spend buy us?"
Verified Case Study
Global D2C Homeware
-22% spend, +18% net profit
Feed-led Shopping rebuild. Zombie SKUs with high returns excluded from the feed, audience signals tightened to high-intent buyers, bidding aligned to contribution margin not platform ROAS.
-22%
Ad spend cut
+18%
Net profit uplift
Day 1
Profitable
Continue reading
Related services and reading
Frequently asked
Google Shopping feed questions, answered straight
Go Deeper
Next step
Get the 9-point Shopping feed audit. Free, no obligation.
We score titles, attributes, GTIN coverage, supplemental gaps, image quality, taxonomy, custom labels, approval health, and feed-bid alignment. You get the diagnostic whether we work together or not.