From idea to MVP in 4 weeks — Lessons from DankShop

Where do you start?

When the client first came to us, the initial request was "an e-commerce site like Shopee, but smaller." That sentence almost sank the project on day one.

First lesson: never start by comparing yourself to a giant platform.

Week 1 — Discovery & scope

Instead of jumping into code, we spent the whole first week on discovery:

• ✓What do users really need? Not features, but the problem they want solved.
• ✓What is the real MVP? We listed 40+ features, then crossed them out until only the 12 most essential remained.
• ✓Who are the first users? The client already had 3 shops selling manually through Zalo.

Result of week 1: a 15-screen Figma prototype and a clear MVP scope.

Week 2 — Foundation

The tech stack decision

We chose:

• ✓Next.js for the frontend — SSR gives good SEO from the start
• ✓NestJS for the backend — clear structure, easy to maintain
• ✓PostgreSQL — e-commerce data relationships need an RDBMS
• ✓Cloudinary for product images — no point building our own image CDN before we need it

The most important decision: payments

We did NOT build our own payment gateway. Integrating VNPay and MoMo through their existing SDKs took 1 day instead of 2 weeks. It was the best trade-off of the whole project.

Week 3 — Core features

We focused on the 3 most important flows:

• ✓Customers place orders — browse → add to cart → checkout → pay
• ✓Shops process orders — receive → confirm → ship → complete
• ✓Admin management — products, inventory, revenue reports

What we CUT that week:

• ✓Reviews & ratings (added later)
• ✓Vouchers & discounts (added later)
• ✓Chat with the shop (added later)

Week 4 — Polish & launch

The last week wasn't about adding features — it was about:

• ✓Fixing bugs from user testing with 5 real people
• ✓Optimizing performance (Lighthouse score from 62 to 89)
• ✓Deploying to AWS with basic auto-scaling
• ✓Writing a user guide for shop owners

Results 3 months after launch

• ✓3 shops using it from day one
• ✓+300% online revenue compared with selling manually through Zalo
• ✓Zero downtime in the first month
• ✓The client continued working with us on phase 2 features

The biggest lesson

A small scope doesn't mean low quality. The DankShop MVP had only 12 features, but those 12 features worked perfectly. Users don't need lots of features — they need the most important ones to work smoothly.

If you have a product idea and don't know where to start, talk to us about custom software development.