DEV Community

klauski24
klauski24

Posted on

I Built Sổ Lãi, a Practical Profit Tracker for Vietnamese Online Shops

#devchallenge #githubchallenge #webdev #productivity

GitHub “Finish-Up-A-Thon” Challenge Submission

This is a submission for the GitHub Finish-Up-A-Thon Challenge

What I Built

I built So Lai, a local-first profit tracker for small online shops in Vietnam.

The goal is simple: help a seller answer the question they often cannot answer from platform revenue alone:

Is my shop actually profitable after product cost, marketplace fees, shipping subsidies, discounts, ad spend, returns, and pending COD?

So Lai is not trying to be a full POS, CRM, or inventory suite. It focuses on the painful middle layer that many small sellers still manage through scattered spreadsheets:

  • Product cost by SKU
  • Orders from Shopee, TikTok Shop, Facebook, Zalo, or livestream sales
  • Platform fees, shipping cost, vouchers, and discounts
  • Ad spend by channel and SKU
  • Return/cancellation status
  • COD received vs. pending
  • Net profit by order, product, and sales channel

The app runs locally with Node.js and JSON storage, so it does not require paid APIs or cloud setup.

GitHub repo: https://github.com/klauski24/so-lai

Demo

Run locally:

git clone https://github.com/klauski24/so-lai.git
cd so-lai
npm start
Enter fullscreen mode Exit fullscreen mode

Open:

http://127.0.0.1:4182
Enter fullscreen mode Exit fullscreen mode

What the demo shows:

  • A Vietnamese-language dashboard called Sổ Lãi
  • Shop profile setup
  • Clear explanation of where the numbers come from
  • CSV import for products, orders, and ad costs
  • Manual order entry
  • Profit analysis by channel and SKU
  • Alerts for loss-making products, high COD pending, and high return rate
  • CSV and Markdown report export

Screenshots are included in the repository:

  • so-lai-desktop.png
  • so-lai-mobile.png

The Comeback Story

The first version was too vague.

It started as an English-named dashboard called ProfitLens. It had some useful calculations, but it did not feel practical yet. The biggest problems were:

  • The app used Vietnamese currency but had an English product name.
  • There was no place to define the shop.
  • It was not clear where the data should come from.
  • The dashboard looked like a demo, not something a seller could actually use.

I reworked the project into So Lai, a more grounded tool for Vietnamese sellers.

The before/after arc became:

  • Before: a generic dashboard with sample data
  • After: a local-first shop profit workbook with shop setup, CSV import, real data source guidance, and exportable reports

The most important improvement was adding a realistic input workflow. A shop can now start with CSV files exported or copied from tools they already use:

  • Marketplace order reports
  • Product cost spreadsheets
  • Ads manager exports
  • COD reconciliation records

That made the project feel much closer to a real small-business workflow.

What Changed

I finished the project by adding:

  • Vietnamese branding and UI copy
  • Shop profile fields
  • CSV import endpoints for products, orders, and ad costs
  • CSV paste and file upload UI
  • Template CSV examples inside the app
  • Order-level profit calculations
  • Product-level and channel-level summaries
  • COD pending tracking
  • Return/cancellation handling
  • Markdown report export
  • README with a practical workflow for real shop data
  • Desktop and mobile screenshots

The current data model is intentionally small:

shop
products
orders
adCosts
Enter fullscreen mode Exit fullscreen mode

That kept the app focused and usable instead of turning it into another oversized management system.

My Experience with GitHub Copilot

GitHub Copilot helped me move from idea to finished MVP quickly.

It was especially useful for:

  • Designing the CSV import workflow
  • Building the product/order/ad cost data model
  • Writing practical profit calculations
  • Refactoring the UI copy from a generic English dashboard into a Vietnamese seller-focused tool
  • Creating sample CSV formats
  • Improving the README so another person can actually run and test the project

The best part was using Copilot as an iteration partner. The project improved when I kept asking whether the tool was realistic enough for an actual seller, not just whether the code worked.

What I Learned

The biggest lesson: a business tool is not useful just because it has charts.

It becomes useful when the user can answer:

  • Where do I put my real data?
  • What format does the data need?
  • What decision does this dashboard help me make?

For So Lai, the answer is:

  • Put in products, orders, COD status, and ads through CSV.
  • Use the dashboard to find products and channels that are secretly losing money.
  • Export the result as a report for monthly review.

What's Next

The next practical step is preset imports for real platform exports:

  • Shopee order CSV mapping
  • TikTok Shop order CSV mapping
  • Facebook/Zalo manual order templates
  • COD reconciliation import

After that, I would add a monthly comparison view so sellers can see whether changes in pricing, voucher strategy, or ad spend actually improved profit.

Thanks for reading!

Top comments (0)