Shopify themed background image

Case Study

Catalog management shouldn't be a full-time job!

2PRICETAGS × AUTOBIZZ

2pricetags was importing raw marketplace scrapes straight into Shopify Plus: options in the wrong fields, prices missing, customs codes wrong, in-stock products showing Sold Out. AutoBizz replaced the manual work with one connected pipeline that turns supplier data into import-ready listings.

“The Team can now trust the data to be well organized, complete and easy to make changes to”

Client

2Pricetags

Sector

E-commerce

Solution

Product Catalog Creation

Timeline

2025-2026

AutoBizz built 2PriceTags a catalog that maintains itself!

AutoBizz built an automated catalog operations system on Shopify Plus. A BigQuery process turns raw supplier data into complete listings, a landed-cost model reprices every product weekly, and a Python and Gemini LLM based process assigns customs codes against the official Bahamas tariff schedule. The result is 20,163 live SKUs with correct options, tags, prices and HS codes, and a documented procedure to repeat the steps for a new set of products.

  • Shopify
  • BigQuery
  • SQL
  • Python
  • Gemini API

The Situation

Twenty thousand products, one manual process.

The catalog grew faster than the process behind it. Every new batch of supplier data meant fixing options, rebuilding tags, recalculating prices and hunting for customs codes by hand, and the knowledge of how to do it sat in one person's head.

Services We Provided

  • Data Automation Engineering
  • Pricing Automation
  • AI Classification
  • Hs Code Assignment

Old Routine

  1. 01Receive raw supplier scrapes with mixed-up options and no clean structure
  2. 02Re-map options, tags and categories by hand
  3. 03Recalculate landed cost for every product, every week (8 to 10 hours)
  4. 04Look up an HS code per product, with roughly half of the existing codes wrong
  5. 05Build import files and fix whatever Shopify rejected

Individually, the steps were manageable. Together they meant days of data entry for every launch, and nobody on the team had time left to grow the catalog.

None of these problems threw an error. They just leaked revenue quietly.’

Time taking manual process of raw data to processed information

The Challenge

The errors weren't loud. They were expensive.

Each defect looked small on its own row. Across 20,000 products, they added up to lost sales, blocked imports and customs risk

How the catalog works by giving the finalized listings
  1. 1. Day zeroRaw supplier data lands with options in the wrong fields and prices missing.
  2. 2. Days laterIn-stock products show Sold Out and images fail to import. Nobody can say why.
  3. 3. Later still820 products sit at $0.00 in the data, one launch away from selling for free.
  4. 4. Too LateA wrong HS code changes the duty, the landed cost and the price, and creates customs risk.
  5. 5. Trust spentThe customer meets the error first: the wrong price, the wrong availability, a delayed clearance.

System Architecture

The Listing Process Board

The flow chart of ingestion to processing of data making it shopify ready

SEE THE DELAY BEFORE IT BECOMES A CUSTOMER PROBLEM

The Turning Point

Automated catalog operations that work for people

2pricetags didn't need a cleaner spreadsheet. They needed a system that fixes defects at the data level, prices every product from the same model, assigns customs codes against the official source, and catches problems before they go live. We stopped the revenue leaks first, then built the machine, then handed it over.

Issues with the Raw supplier data

The Solution

One process for the whole catalog

AutoBizz rebuilt catalog operations around one idea: remove manual work wherever a system can make the decision faster and more accurately.

01

The Repair Crew

Fixed false “Sold Out” flags, blocked image imports, misfiled options and junk tags.

02

The Listing Factory

SQL turns raw supplier data into complete Shopify listings. Nothing is typed by hand.

03

The Price Engine

Cost, freight, duty and fees feed one model. Vendor costs sync weekly and every price updates.

04

The Safety Net

Caught 820 products at $0.00 before launch. The check now runs on every pass.

05

The Customs Desk

Gemini matches each product to the Bahamas tariff and returns a code, a reason and a confidence score. Low scores go to manual review.

06

The Navigator

24 marketplace categories became 27 subcategories in 4 departments, tagged to bypass Shopify’s 500-product collection cap.

07

The Handover

Metafields, navigation, search and written SOPs, so the client’s team runs it.

What Changed

From manual upkeep to a self-running catalog

Before

In-stock products showed Sold Out

Now

Every live product is buyable

Before

Prices maintained by hand across 20,163 SKUs

Now

Weekly cost sync and automatic reprice

Before

HS codes roughly 50% accurate

Now

AI-assigned codes checked against the official tariff schedule

Before

Raw marketplace categories

Now

27 subcategories across 4 departments

Before

820 zero-price products hidden in the data

Now

Zero-price detection on every run

Before

The process lived in one person's head

Now

Written SOPs the client's team executes

Measureable results and their metrics

The Results

Measurable improvements across the catalog

20,163

Catalog rebuilt

Every live SKU re-optioned, re-tagged, re-priced and given a valid HS code.

90%

HS code accuracy

Up from roughly 50%, measured against the official Bahamas tariff schedule.

820

Caught pre-launch

Zero-price products flagged before going live. Source data gaps, not free products.

300+

Hours saved per annum

Landed cost now recalculates automatically every week.

References

Documentation this build relies on

  1. 01

    BigQuery Documentation

    The SQL layer behind the listing process.

    View source
  2. 02

    Shopify Product CSV Import

    The import format the process generates.

    View source
  3. 03

    Bahamas HS Tariff Schedule

    The official classification reference.

    View source

Get Started

Ready to automate your catalog?

If your team is still fixing options, repricing by hand or guessing at customs codes, AutoBizz can turn that process into a connected, self-running process.

FAQ

Frequently asked questions

Still have questions? Book a call and we'll answer anything.

Can AI assign HS codes accurately?

Yes, when it works from the official tariff schedule. Our process gives the model the schedule as its reference, returns an 8-digit code with a reason and a confidence score, and sends low-confidence results to manual review.