VROF

Work

Collecting the sellers of products sold under a brand name without permission

To support a public cultural foundation's copyright and trademark protection work, we automatically collected the sellers of products sold in online shopping malls under the foundation's brand name. It counts by seller, not by product.

Sellers identified

Before

Checking search results one by one

After

Automatic collection across all pages

Structure diagram of changing the counting unit from product to seller. On the left are rectangles representing ten products, and the branches coming out of them converge on a single seller on the right.
Even if one seller uploads ten products, counting by seller makes it one place. The number of targets for action does not get inflated.
Client
Public cultural foundation
Industry
Public · Culture
Completed
2025
Duration
2 weeks
Service
Workflow automation
  • Collect all result pages
  • Product-name matching
  • Deduplication by seller
  • Excel export

Background

The cultural products the foundation makes carry a brand. And products sold under that brand name without official permission get listed in online shopping malls. Protecting copyright and trademark starts with knowing where they are being sold.

What we saw on the ground

Mall search results are split across pages, and the same seller keeps appearing repeatedly under different products. When a person scans them by eye, the later pages get missed, and sellers already checked earlier get checked again. The quality of the check depends on how many pages the search results span.

The problem we defined

What needed to be found was the seller, not the product. Counting by product catches the same place multiple times, so how many places actually require action cannot be known. Changing the counting unit was the core of this work.

What we built

  • Walk all result pages — search by brand name and go page by page, 40 items per page, all the way to the end
  • Product-name matching — keep only products whose name actually contains the brand name
  • Deduplication by seller — the same seller is recorded only once. The number of targets for action is exactly the number of sellers
  • Excel export — the product name, seller address, and position within the search results are kept together so they can be revisited

What we guarded against

  • Pace yourself between pages — each page turn waits a random amount of time. Not hammering the site is the way to keep reaching the later pages without getting cut off
  • Go all the way to the end — in the search results, spots that are not official sellers are packed toward the back, not the front. Seeing only the first few pages means missing what you were supposed to find
  • Duplicates are filtered by seller — even if one place posts ten products, only one record remains. The number of targets for action does not get inflated

Result

The commissioned scope was completed and delivered.

Other work made the same way is gathered at web data collection.

Sources

The collection script

We offer consultations to help you save time Free consultation · Replies within 1 business day

Contact