Furniture Product Data Quality Scorecard

“The data needs work” is hard to act on. This scorecard shows where the work is.

Use it on a sample from one supplier or range. Include straightforward products, highly configurable products and recently added SKUs. The purpose is not to manufacture a flattering score. It is to find the gaps that block publication, confuse customers or create rework across channels.

How to score a range

Give each check one of three scores:

• 0 — missing or wrong
• 1 — present, but incomplete, inconsistent or unverified
• 2 — complete, structured and checked against an approved source

Mark a check N/A only when it genuinely does not apply to the product or market. Do not average the result until you have looked at the zeros. One missing safety document or a broken variant structure can matter more than several polished descriptions.

1. Identity and variant structure

Stop and fix it if: the variant structure is wrong. Copy and imagery attached to the wrong hierarchy will have to be rebuilt later.

GS1 says each product variation needs its own GTIN where GTINs are used, while Google expects variants to be submitted separately and grouped with a shared item_group_id.[1][2]

2. Customer-facing product facts

AI-drafted copy can score 2 after a person has checked and approved it. Unreviewed copy scores 1, however fluent it sounds.

3. Dimensions, materials and logistics

Stop and fix it if: a customer cannot judge fit, access or the material being purchased.

Google separates assembled product measurements from shipping measurements in its product data specification.[3] Your source record should do the same.

4. Image and asset coverage

Furniture Connect can run colorway, packshot and lifestyle production across hundreds of SKUs.[4] It does not remove the approval decision. Generated images should be accepted, amended or rejected by a reviewer, then linked to the exact variant with confirmation.

Channel format is another explicit check. A square marketplace image or web banner ratio must be requested and reviewed; it is not silently inferred.

5. Compliance and evidence

Stop and fix it if: required evidence is missing or cannot be matched to the item being sold.

UK upholstered furniture and US upholstered furniture, clothing storage units and composite wood products are covered by different rules.[5][6][7][8] Use the furniture product data model to decide what the schema must hold, then confirm current obligations with your compliance lead.

6. Channel readiness and ownership

A product can be complete for the website and not ready for a marketplace or print catalog. Score the intended destination, not an abstract idea of completeness.

Read the result

Avoid pretending the number is more scientific than it is. Use the score to rank the work.

• Mostly 2s: the range is well structured. Investigate the remaining zeros and put a date on future rechecks.
• A mix of 1s and 2s: the data exists, but the team is carrying verification and consistency risk. Assign owners and sources.
• Several 0s in one section: that section is the real project. Fix it before expanding the range to more channels.
• Any stop condition: hold the affected products until the structure, critical facts or required evidence are resolved.

Compare results by supplier and range. The pattern is more useful than a single overall percentage. One supplier may consistently omit logistics data; another may provide good specifications but no variant-level imagery. Those require different responses.

Turn the score into an action list

Work in this order:

1. Correct product and variant identity.
2. Resolve facts that affect fit, use and delivery.
3. Link required compliance evidence.
4. Close image gaps and complete manual accuracy review.
5. Prepare channel outputs from the approved record.
6. Improve copy and search content from verified inputs.

Then add the missing requirements to the next supplier intake pack. That is how the scorecard improves catalog onboarding instead of becoming another report nobody opens.

Frequently asked questions

Sources

1. GS1: How many GTINs are needed for products with multiple sizes and colors?
2. Google Merchant Center: Item group ID for product variants
3. Google Merchant Center: Product data specification
4. Furniture Connect: AI for furniture at catalog scale
5. UK legislation: Furniture and Furnishings (Fire) (Safety) Regulations 1988
6. US CPSC: Flammable Fabrics Act guidance, including upholstered furniture
7. US CPSC: Clothing storage unit requirements
8. US EPA: Formaldehyde standards for composite wood products

---

If you want to see what this looks like on a real range, bring a supplier file to a Furniture Connect demo. We will show you how the catalog surfaces gaps and keeps approved product data and imagery connected.