Cantürk Çelikok
All projectsE-commerce App

Misrepresentation Audit: Merchant Center Audit App

A dashboard app that audits T-Soft stores against Google Merchant Center's misrepresentation rules, lists issues by severity and generates a cleaned-up product feed.

Misrepresentation Audit summary screen: risk score, scores for 7 areas and the issues to fix first
01

Problem

For stores listing products on Google Shopping, one of the biggest risks is having the Merchant Center account suspended for "misrepresentation". The causes are usually small but scattered: no phone number on the contact page, no stated return period, a product page price that doesn't match the feed, claims like "miracle", discounts written into titles, products in the wrong category. Finding them one by one is almost impossible for a store owner.

02

Solution

I built an app that installs into the T-Soft dashboard. It scans the store from two sides: through the Admin API for products, variants, categories, brands and contact details, and through the storefront, the way Google sees it, for the home page, policy pages, product pages and structured data. The result appears inside the dashboard as a risk score, separate scores for 7 areas and a list of what to fix first, with why each issue matters and how to fix it.

03

Outcome

Store owners see the issues up front and on one screen, instead of after their account has been suspended. New products are checked automatically, and the product data sent to Google is cleaned up. The app runs live on a test store and is ready for T-Soft App Market review.

Misrepresentation Audit is a dashboard app I built for e-commerce stores running on the T-Soft platform. It audits the store against Google Merchant Center's misrepresentation rules with 86 checks across 7 areas: identity and contact, policies, price and payment transparency, product data and stock, claims and counterfeits, brand and visual identity, and technical and security. The text in the logo is read with OCR and compared with the store name; logo contrast, colour legibility and promotional text on product images are detected too. Store categories are mapped automatically to Google's 5,595-category Turkish product taxonomy, helped by a dictionary of Turkish e-commerce terms. The app generates a Google product feed that strips promotional titles, fixes stock and barcodes, writes variants as separate rows with their colours and sizes, and leaves risky products out; pasting the address into Merchant Center is all it takes. A full scan runs every day, new products are checked within minutes, and the product edit page shows a card with that product's risk. It only asks for read access and never changes anything in the store. This app is not an official Google product and is not affiliated with Google.

Client

Own product

Technologies

Node.jsTypeScriptExpressT-Soft Admin APICheerioSharpTesseract.jsGoogle Merchant Center
Issue list: severity filters and an explanation of each issue
Screen mapping store categories to Google product categories
Google Merchant Center feed screen: feed address, statistics and automatic fix settings
Brand identity screen: logo contrast, logo text read with OCR and the store's colour palette
Automatic check results for newly added products

Other projects