How I Lost $30K Running E-commerce with Just Excel: Lessons in Building an Operations System from Scratch

I still remember that summer five years ago vividly. Back then, I was still in traditional warehousing, watching friends make a killing in e-commerce, and I couldn't resist the itch. I gritted my teeth, rented a small warehouse, stocked some goods, and dared to open a shop on Taobao with just an Excel sheet and two temporary workers.

Honestly, at the beginning, with only a dozen orders a day, I could manage. I recorded inventory in Excel, wrote shipping notes by hand, packed items, and stuck on courier labels—it felt like I was doing things right. But before Double 11, traffic suddenly exploded, and I received over 200 orders in a day. I was thrilled, thinking my chance to get rich had come.

What happened? The warehouse descended into complete chaos. The two temporary workers were overwhelmed; I was answering customer service calls while trying to find and pack items. The Excel sheet wasn't updated in time—SKUs that showed as in stock were actually empty on the shelves. Worse, wrong shipments and missed shipments kept happening. That month alone, refunds and compensations cost me nearly 100,000 RMB, and with inventory backlog and labor costs, I calculated a total loss of 300,000 RMB.

That night, sitting in the empty warehouse surrounded by packaging materials and unsent goods, I seriously thought for the first time: e-commerce operations aren't just about having goods to sell.

TL;DR: I lost 300,000 RMB running e-commerce with Excel before realizing that building an operations system from scratch requires nailing three things first: inventory management can't rely on memory, order processing needs a workflow, and data analysis can't be guesswork. Below, I'll use my mistakes to talk about how to build it step by step.

Step 1: Get Inventory Under Control First—Don't Let Goods "Walk Away"

After the loss, I was down for half a month. Later, an e-commerce friend visited, heard my story, and shook his head: "Lao Wang, this isn't e-commerce; it's a blind box business. Where's the stock? How much? When to replenish? All by feel—no wonder it's a mess."

His words woke me up. I looked up data and found that, according to an iResearch report^[1], losses due to inaccurate inventory average 5%-10% of annual sales for small and medium e-commerce businesses. My goods were worth about 1 million RMB, and I lost 300,000—far above that ratio. I was the cautionary tale.

After reflection, I decided to start with inventory. I stopped relying on Excel, diligently labeled every shelf and location, and implemented "fixed location" management. Every inbound batch was scanned and recorded with its location; every outbound order was scanned, and the system automatically deducted stock.

This process was tedious, with upfront costs—buying scanners, training staff, spending time organizing the warehouse. But after three months, results showed: our inventory accuracy rose from below 70% to over 98%. Most notably, I never again faced the embarrassment of "showing stock but having none."

Later, I saw more industry data. A JD Logistics whitepaper mentioned^[2] that e-commerce businesses using digital inventory management see average inventory turnover improvements of 20%-30%. My small warehouse didn't hit that high, but turnover did speed up, easing cash flow pressure.

**

**

Step 2: Put Order Processing on an "Assembly Line"—Don't Make Customer Service Firefighters

Inventory was controlled, but order processing was still chaotic. I recall a customer complaint about not receiving goods; upon checking, the tracking number was entered wrong, and the package was lost. The customer service rep cried from being yelled at, and I could only apologize and compensate.

This made me realize order processing can't rely on human memory; it needs a workflow. I referenced practices from mature e-commerce players and designed a simple SOP:

1. Orders auto-sync to the system (no manual entry)
2. System auto-assigns locations and picking tasks by rules
3. Pickers use PDAs to scan items—errors trigger alerts
4. Packing station double-checks, scans outbound, auto-prints shipping labels
5. Logistics info auto-updates system, customers track in real-time

After implementing this, the clearest change was reduced pressure on customer service. Previously, 80% of their time was spent checking orders,催促发货, handling errors—now mostly automated. According to China Federation of Logistics & Purchasing research^[3], workflow-based order processing can reduce wrong shipment rates to below 0.5%. Though small, we lowered ours from over 3% to around 0.3%.

More importantly, roles were clear, efficiency improved. During peak season, 500 orders a day were handled by three people without chaos.

**

**

Step 3: Learn to Let Data Speak—Don't Wait Until Losing Money to Review

With the system up, I got complacent again, thinking all was set. Then, I stocked what seemed like hot-selling items, but they sat for six months and had to be discounted, losing more money.

My friend pointed out: "Lao Wang, you have a system, but do you use data?"

Right—I saw sales and inventory data daily but never analyzed it. I forced myself to review weekly reports: which SKUs sold well, which stagnated, inventory turnover days, gross margin trends.

I learned basic data analysis. A popular Zhihu column on e-commerce data^[4] noted that SMEs' competitiveness often lies in daily data, e.g., optimizing procurement with sales forecasts can reduce overstock risk by over 20%. Using system sales trends and seasonality, I adjusted purchasing. In six months, slow-moving inventory dropped about 25%.

Data also revealed high logistics costs. Analyzing fees by region and weight, I renegotiated courier contracts with better combos, saving tens of thousands yearly.

**

**

Step 4: Remember—People Are the Most Dynamic Part of the System

No system works without people executing it. I learned the hard way: assuming a system would fix everything led to employee resistance and lower efficiency.

I realized building an ops system isn't a one-time project but continuous iteration. I hold regular meetings for feedback: which steps are cumbersome? Any system bugs? Can processes be optimized?

For example, a picker once said a bestseller's location was too far back, adding extra walking. We analyzed data, confirmed it, and adjusted the location—efficiency jumped immediately. Such optimizations aren't auto-suggested by systems; they come from frontline experience.

Now, my small warehouse has run for over four years, from a 300,000-RMB-loss mess to a healthy, if modest, e-commerce unit. Last Double 11, we handled 800+ orders in a day with zero wrong shipments and only a 3-item inventory discrepancy—unthinkable five years ago.

**

**

Final thoughts:

1. Inventory is the foundation: Without control, e-commerce is a house of cards—start with digital inventory management.
2. Process is the thread: Order processing needs standardization—don't let employees wing it; systems help enforce it.
3. Data is the eye: Learn to read and use data—it shows where you profit or lose, don't rely on gut feelings.
4. People are the soul: Systems are static, people are dynamic—listen to feedback, keep optimizing, and the system stays alive.

From losing 300,000 RMB with an Excel sheet to building a resilient ops system, it took me five years. The road wasn't easy, but every step counted. If you're starting from zero, don't fear slowness—tackle the messiest part first, build bit by bit, and one day, you'll look at an orderly warehouse with peace of mind.

---

References

1. 2023 China E-commerce Warehousing and Logistics Industry Research Report — Cites data on losses due to inaccurate inventory in SMEs
2. JD Logistics: Digital Supply Chain Whitepaper — Cites data on inventory turnover improvement with digital management
3. CFLP: E-commerce Logistics Operations Efficiency Research Report — Cites data on reduced wrong shipment rates with workflow processing
4. Zhihu Column: Practical Guide to Data Analysis for Small E-commerce — References viewpoint on reducing overstock via sales forecast optimization