Tools and providers
Part of Before you sign a shopping feed contract, check these tools and providers
Shopping feeds tool implementation explained for multichannel shops
Practical steps for implementing a shopping feeds tool in England: scope, data mapping, feed rules and quality checks, with a decision table and common questions.
What to take away
- Implementation runs in seven stepsscope the channels, map attributes, build rules, schedule exports, pilot, hand over and review.
- Hosted feed platforms such as DataFeedWatch, Channable and Productsup make sense once you serve two or more channels.
- UK consumer feeds must quote the price a shopper actually pays, so VAT-inclusive figures and delivery charges belong in the feed.
- Accept the build only when fewer than 2 per cent of items are disapproved.
- Name one owner for product data and one approver for rule changes.
The seven implementation steps in order
The seven implementation steps
- Scope the channels and the product set each one needs.
- Map source fields to destination attributes on one sheet.
- Build and name rules that clean descriptions, brands and variants.
- Schedule exports to run after stock and price jobs finish.
- Pilot 20 to 50 products and compare what each channel receives.
- Hand over with acceptance criteria, a runbook and two training sessions.
- Review feed health after thirty days, then monthly.
Scope the channels and map attributes
List every destination: Google Merchant Centre, Microsoft Shopping, eBay, Amazon, Meta catalogues and any affiliate network. Record the product set each one needs, whether that is the full catalogue or a seasonal range.
Feed failures trace back to missing attributes more often than to the tool itself. One named owner for titles, images and prices prevents most of them.
Build a mapping sheet with your source field on the left and the required feed attribute on the right.
Source field to feed attribute
Source field
- Product title
- title
- Image URL
- image_link
- Price
- price
- Stock level
- availability
- Brand
- brand
Feed attribute
- Product title
- Image URL
- Price
- Stock level
- Brand
Google's image link requirements set the format, size and availability rules that decide whether an item can appear at all.
If you are still weighing suppliers, the criteria worth settling before mapping begins are set out in shopping feeds software selection in England.
Example: mapping one variant-heavy product
A T-shirt held as one parent SKU and twenty child rows still needs twenty feed rows. Each row carries its own item_group_id, colour, size, GTIN and image link.
Write one rule per attribute and reuse it everywhere. Two rules describing the same thing will disagree eventually.
Choose the tool type and schedule exports
A spreadsheet handles one channel with infrequent updates. Once you serve several channels or change prices daily, a hosted platform earns its cost by automating rules and scheduling.
Schedule exports after the nightly stock and price jobs finish. Overlapping schedules produce duplicate rows and rejected items.
Choose the tool type
| Step | Decision | Sensible default |
|---|---|---|
| 2 | Attribute source | Source field, with a platform rule for gaps |
| 3 | Where rules live | In the platform, never in hand-edited rows |
| 4 | Export timing | After the nightly stock and price job |
| 5 | Pilot size | 20 to 50 products covering every attribute type |
| 6 | Acceptance | Under 2 per cent disapproved, all channels live |
Apply UK price and listing rules
Google Merchant Centre expects the price attribute to match the price on your landing page. For most UK retailers that means a VAT-inclusive figure, with delivery charges declared separately.
A feed price that differs from the page can count as a misleading omission under the Consumer Protection from Unfair Trading Regulations 2008 (legislation.gov.uk).
Listing text and promotional claims also fall within the ASA advertising codes, so leave out superlatives your product data cannot support.
When disapprovals appear, how UK merchants fix Google Merchant Centre feed disapprovals works through the usual causes.
Hand over with acceptance criteria and a runbook
Define finished before you sign off: every channel live, fewer than 2 per cent of items disapproved, and a date when your team owns the schedule. Ask the provider to document the mapping logic they built, not only the result.
Book two training sessions. One covers price and stock uploads, the other covers disapproval handling. Staff turnover is the main reason feeds break quietly months after launch.
The runbook fits on one page: who approves rule changes, how to pause a feed, and where the mapping sheet lives.
Review feed health after go-live
Set a thirty day review with the provider covering disapproved items, delivery errors and any channel warning still open. Agree which side fixes what.
Then move to a monthly check of active products, approved items and missing attributes. Climbing disapproval counts usually point to a change in your source system rather than a tool fault.
Common questions
How long does implementation take?
For one channel and a catalogue under a few thousand SKUs, allow two to four weeks from kick-off to a stable feed. Variant-heavy, multi-channel builds take longer because mapping and testing need more rounds.
Do we need a feed platform, or will a spreadsheet do?
A spreadsheet suits one channel with infrequent updates. Two or more channels, daily price changes or large variant counts justify a platform, and contract terms are compared in shopping feeds tools and supplier guide for 2027.
What causes most feed rejections?
Missing or malformed attributes, especially images, GTINs and prices. Validate inside the export so errors surface before the channel sees them.
Who owns the feed after launch?
Day-to-day ownership sits with whoever manages product data, plus a named approver for rule changes. Split those roles and fixes stall.
