Your product feed is the foundation of every Shopping channel. A poorly optimised feed costs you impressions, clicks, and conversions. This guide covers everything UK merchants need to know — from title structure to AI enrichment to custom labels.
Product titles are the single most impactful feed element. Include brand, material, colour, size, and product type in the right order for your category. Trajekt's rules let you rewrite titles at scale.
🖼️
Image quality
Feeds with high-quality images on clean backgrounds get significantly more clicks. Minimum requirements: 500x500px for Meta, 250x250px for Google. Larger is always better.
🔢
GTIN coverage
GTINs (EAN barcodes) dramatically improve product match quality in Google Shopping. Trajekt maps barcodes from your source automatically and flags missing GTINs.
🏷️
Custom labels
Segment your catalogue into bidding tiers with custom_label_0-4. High margin, clearance, seasonal — set different bids per segment in your Shopping campaigns.
🤖
AI enrichment
Trajekt's AI Rules can rewrite product descriptions, enrich titles with missing attributes, and generate channel-specific copy at scale — for thousands of products overnight.
📂
Category accuracy
Correct google_product_category assignment improves visibility in category searches. Trajekt includes rules to map your product types to the right Google taxonomy IDs.
FAQ
Frequently asked questions
Product feed optimisation is the process of improving the quality, completeness, and accuracy of your product data to increase visibility and performance on Shopping channels. It includes enriching titles and descriptions, mapping to channel taxonomies, adding missing attributes like GTINs, filtering underperforming products, and setting custom labels for campaign segmentation.
Product title has the biggest single impact. Google matches your products to queries based on title more than any other field. Including brand, key attributes (colour, material, size), and product type in the right order for your category can increase impressions by 30%+ and significantly improve CTR. Test title variations systematically.
While not technically mandatory for all products, GTINs (Global Trade Item Numbers — EAN/barcodes) dramatically improve Google Shopping match quality. Products with correct GTINs get better placement and quality scores. Products without GTINs must declare identifier_exists: no and may have reduced visibility. If your products have barcodes, always include them.
Custom labels (custom_label_0 through custom_label_4) let you add your own segmentation data to products. You use these labels to create product groups in your Shopping campaigns and set different bids per segment. For example: custom_label_0 for margin tier (high/medium/low), custom_label_1 for seasonality (spring/autumn/evergreen).
At minimum, daily updates for pricing and availability. Twice daily is better for fast-moving stock. Google requires updates at least every 30 days to prevent products expiring. ChatGPT accepts updates every 15 minutes. Meta and TikTok support hourly updates. Price mismatches between feed and website are the most common cause of disapprovals.
Feed management covers the technical side: creating, formatting, and delivering product data to channels. Feed optimisation goes further — improving that data to perform better. Enriching titles, mapping categories, adding missing attributes, filtering underperformers, and A/B testing variations. Trajekt handles both: the technical plumbing and the optimisation tools.
Good Shopping descriptions are factual, specific, and keyword-rich without being spammy. Include key product specs (dimensions, materials, weight), use cases, compatibility, and care information. Avoid marketing language like "amazing" or "best quality". For AI search channels like ChatGPT, conversational factual descriptions perform best. Trajekt's AI Rules can rewrite descriptions at scale.
google_product_category is a numeric ID from Google's product taxonomy. It helps Google categorise your products for better query matching. For example, 5672 for shirts, 1604 for outerwear. Providing specific, accurate categories (not just top-level) improves visibility in category-based Shopping searches. Trajekt's rules engine can map your product types to taxonomy IDs automatically.
Common causes: price mismatch (feed price differs from website), missing required fields (especially GTIN for products that should have one), low-quality images, landing page mismatch (title/description differ from feed), availability mismatch, Google policy violations (prohibited products, deceptive content). Trajekt validates against common disapproval causes before submission.
Primary product image should show the product clearly on a white or neutral background, with the product taking up at least 75% of the image. No promotional overlays, watermarks, or borders. Additional images can show different angles, lifestyle shots, or product in use. Google supports up to 10 additional images; Meta supports multiple; ChatGPT supports up to 4. Higher quality = better CTR.
Trajekt's AI Rules let you apply Claude AI to specific fields across your entire catalogue. For example: rewrite all product descriptions in a clean, factual format suitable for ChatGPT; extract colour and material from title and populate separate colour and material fields; or generate better Google Shopping titles based on product type rules. AI credits are used per product processed.