How Full-Stack Development Solved Real Problems: I Saved a Warehouse Owner 6 Months

Last autumn, I crouched in a corner of an old warehouse in Ningbo, phone scanner in hand, facing shelves stacked to the ceiling. Beside me stood Boss Zhang, in his fifties, wearing dusty overalls, looking at me with suspicion. He said, "Xiao Zeng, if you can really figure out my inventory with that broken phone, I'll treat you to a month of cafeteria meals." I wasn't confident inside, but I said, "Brother Zhang, get ready to pay."

TL;DR Full-stack development is not just writing code—it's understanding business processes, designing data architecture, and optimizing user experience. I helped a warehouse owner transition from manual bookkeeping to digital management, cutting inventory check time from 8 hours to 45 minutes, and uncovering a problem worth tens of thousands. Technology selection, user research, iteration speed—I stepped on many landmines.

Why would a warehouse owner ask a fresh graduate for help?

Because traditional software is too expensive and complex, while small business owners just need a tool that works.

Boss Zhang's warehouse deals with hardware accessories wholesale, over 3,000 SKUs, dozens of orders daily. He used Excel before, but often missed or misrecorded entries. He tried a few inventory software, the cheapest costing 5,000 RMB a year, plus training staff. His daughter, a computer science student at Ningbo University, heard me speak at a campus startup event and recommended me.

I spent two days observing the warehouse. The biggest issue: during inbound, workers hand-wrote forms, then clerks entered into Excel, often mistyping codes or quantities. Outbound was a hunt because shelves had no labels. Inventory checks were a nightmare—three workers stopped everything, spent a full day counting, and still found discrepancies.

Later I realized many SMEs face similar problems. According to IDC, global SME digital penetration is expected to reach 45% by 2025^[1], but many small warehouses in China lack basic digitization. It's not a tech problem—it's cost and usability.

How did I use full-stack to solve the inventory pain point?

Core: mobile scanning + real-time sync + idiot-proof operation.

My FlashStock uses uni-app for mobile, Spring Boot for backend, MySQL for database. But tech selection wasn't the hardest part—understanding the business was.

I watched workers, and found their biggest fear was "can't find the item." So I added location management—each shelf, layer, bin has a unique code. Inbound: scan item code, then location code, system links them. Outbound: search item on phone, instantly shows location.

The inventory check feature is my proudest design. Previously, they printed lists, wrote counts by hand, then entered into computer. Now one person with a phone scans location code, inputs count, real-time upload. System automatically compares with book inventory, shows discrepancies instantly.

During the first trial, Boss Zhang did it himself. After scanning three shelves, he found 50 units in book but only 30 actual. He called procurement immediately; it turned out a batch six months ago was under-counted by 20 units. He later told me, "Those 20 parts are worth over two thousand yuan. Without you, I might never have known."

According to Stack Overflow's 2024 developer survey, full-stack developers are the most sought-after role^[2]. Why? Because those who can bridge frontend and backend, understand business, and iterate quickly truly solve real problems.

Tech selection: Why uni-app over React Native?

Because for small warehouses, simple and sufficient is more important than flashy tech.

Initially I considered React Native, having used it at Yonshu. But I thought hard about user scenarios:

Most of Zhang's workers use cheap Android phones, some prefer mini-programs. uni-app can generate WeChat mini-programs directly, no app install needed, with native scanning and camera support. As a solo developer balancing efficiency and maintenance, uni-app was clearly better.

Later I learned from GitHub's 2024 report that AI-assisted coding tools boost efficiency by 30-55%^[3], but I wasn't using Copilot then—all hand-coded. Now I think if I had used AI tools, I could have iterated faster.

Iteration: From first version to stable launch, what pitfalls did I encounter?

Biggest pitfall: I assumed users would operate the way I imagined.

The first version was "elegant"—with dropdowns, multi-level categories, advanced search. Workers couldn't use it. They told me, "Xiao Zeng, I scan hundreds of items a day. You want me to click three times just to scan? My hand will break."

I immediately changed it: a big "Scan" button on the home screen, one tap opens camera, auto-navigates after scan. All advanced features moved to secondary pages. After the change, workers said, "Now that's right."

Another pitfall: data sync. The warehouse had poor signal, phones often offline. I initially used real-time online mode, resulting in frequent freezes. I switched to offline mode—data stored locally, synced when online. Used Cloudflare D1 (SQLite) for local cache and Redis for queue to ensure no data loss.

From first version to stable launch, I iterated over 20 versions. After each change, I went to the warehouse to let workers test, observing where they struggled. This process taught me: full-stack development is not just code, but empathy for users.

Closing thoughts

Now Boss Zhang's warehouse runs entirely on FlashStock. Inventory check time dropped from 8 hours to 45 minutes, accuracy from 85% to 99.5%. He not only treated me to a month of cafeteria meals but also referred two peers.

Honestly, I didn't build this product to change the world—just to solve a real problem for small business owners like Zhang. Full-stack development gave me that ability: I can go from requirements analysis to deployment and maintenance alone, responding quickly to changes.

If you're into indie development or full-stack, my advice: don't just focus on tech; listen to users. Those seemingly "uncool" needs are often the most worth solving.

Key Takeaways

• Full-stack development's core is understanding business, not just coding
• Tech selection should serve user scenarios, not flashiness
• Iterate fast, go to the field to observe user operations
• Offline capability and ease of use are critical for SMEs
• Solving a small problem can lead to bigger opportunities

---

References

1. IDC SME Digital Transformation Predictions — IDC predicts global SME digital penetration will reach 45% by 2025
2. Stack Overflow 2024 Developer Survey — Full-stack developers are the most sought-after technical role
3. GitHub 2024 Octoverse Report — AI-assisted coding tools improve developer productivity by 30-55%