32 Client Shops. Roughly 27,000 Listings. Zero Manual Entry.
Addrox was running all of it out of Google Sheets. The errors were not an inconvenience, they were reaching client account health.
- Client
- Addrox, Lahore
- Tech Stack
- Custom headless scraper, TikTok Shop API, custom ERP
- Services Delivered
- Development, AI Automation
Results at a glance
The numbers first. The build underneath.
The challenge
Addrox manages TikTok Shop and Amazon accounts for clients. By the time they came to us, that meant 32 separate TikTok Shops, each carrying around 842 listings.
- Every one of those listings was being copied from eBay by hand and rekeyed into TikTok Shop.
Read the detail
Every one of those listings was being copied from eBay by hand and rekeyed into TikTok Shop. Inventory was then synced manually on top of that. Holding it together was a set of Google Sheets that had grown into the system of record for the entire operation.
Two problems came out of it. The first was time, and there was a lot of it. The second was worse: manual entry at that volume produces errors, and those errors were landing on client accounts. When a shop shows stock that eBay has already sold, orders get cancelled and account health takes the damage. Addrox was absorbing complaints for mistakes that were structural rather than careless.
The outcome
32 client shops synced from eBay to TikTok Shop with no manual entry
- Roughly 27,000 listings maintained automatically
- Google Sheets removed from the workflow entirely
Read the detail
Errors from manual rekeying eliminated, along with the account health damage they caused
Running in production for 13 months
32
client shops on one system
13
months live in production
Our solution
A scraper that reads a full eBay listing and posts it natively to TikTok Shop, and the ERP that runs all 32 client shops behind it.
- The scraper.
- The ERP.
- Eight modules, one login, 32 shops.
Read the detail
The scraper. A headless system that reads the full eBay listing, pricing, variants, stock levels and description text, then posts it natively to TikTok Shop in the structure TikTok expects. No CSV exports, no middleware subscription, no spreadsheet in between.
The ERP. Addrox does not just need listings synced, they need to run 32 client businesses. So we built the system they were trying to approximate in Google Sheets: order management, payout management, return reconciliation, supplier management, client payment tracker, accounting covering invoicing, expenses and profit and loss reporting per shop, task management, and Data Box, a private file storage and sharing module built into the system.
Eight modules, one login, 32 shops. The scraper keeps the listings honest. The ERP is where the business actually runs.
Before
- eBay listings copied into TikTok Shop by hand, shop by shop
- Inventory synced manually across 32 accounts
- Google Sheets serving as the system of record
- Manual errors reaching client account health
After
- eBay to TikTok Shop sync running continuously across all 32 shops
- Orders, payouts, returns, suppliers and accounting in one system
- Client reporting and payment tracking built in
- Thirteen months in production
Scale
32 client shops synced with no manual entry.
Accuracy
Manual rekeying errors eliminated from the workflow.
Operations
An ERP replaced the spreadsheet stack entirely.
“32 TikTok shops and we were running them from Google Sheets. Now listings, orders, everything sits in one place. Nobody is copying stuff by hand anymore.”
How we delivered it
The journey, architecture, and product surfaces behind the results.
Discovery
Mapped every field in an eBay listing against its TikTok Shop equivalent, and mapped the Google Sheets workflow to find what the spreadsheets were actually doing. The second exercise is what turned this from a sync tool into an ERP build.
Built for one shop first
The scraper and sync engine were built and proven against a single live shop before anything was rolled out. Building against live listings rather than test data meant the edge cases surfaced during development instead of after launch.
Rolled out to all 32
Once it held on one shop, it went out across the rest of the portfolio. It runs on every one of the 32 today.
The ERP on top
Order management, payouts, returns, suppliers, accounting, client payments, tasks and file storage. Six months of development in total.
Thirteen months live
Addrox has run 32 client shops on this system for over a year.
Before you book the call
Questions about this one
We don't manage 32 shops - does this only work at that scale?
No. The scraper and sync engine were built and proven on a single live shop first, then rolled out across the rest of the portfolio. The same approach applies whether you're running one shop or thirty.
How long has this actually been running, not just built?
13 months in production as of this case study, across all 32 shops.
Does this replace tools we already use, or require starting over?
For this client it replaced Google Sheets entirely, but the build connects to what you already run - eBay and TikTok Shop APIs here, your own stack elsewhere. Nothing needs to be ripped out before we know what's actually there.
Next Step
Same problem, different company?
Bring the bottleneck. In 30 minutes you leave with the order of operations - whether or not you hire us.