From Excel to WMS: The Nightmare Data Migration That Almost Killed My Business

Last summer, on the hottest day in the warehouse, I crouched in an aisle stacked with boxes. My phone screen showed a red error from the WMS: "Inventory data validation failed. Difference exceeds threshold." I looked up at the clock on the wall—2:30 AM, three days before the Double 11 presale started. At that moment, I really wanted to smash my computer.

TL;DR: Migrating from Excel to a WMS isn't just copy-paste. It took me three months and I lost ten pounds to figure it out. Today, I'll share the pitfalls—unclean data, dirty historical records, inventory mismatches—and how I used Shancang WMS to clean things up step by step.

闪仓 WMS · 示意图

内容概览

First Pitfall: Thinking Excel Data Is Clean

When I took over warehouse management, my predecessor handed me a USB drive saying, "All the data is here." I opened it to find 13 Excel files named "Inventory 2021," "Inventory 2022," "Inventory (Latest)," "Inventory Backup," "Inventory Final," "Inventory Final 2"... I was stunned.

Even worse, the same SKU had different names across different sheets. For example, "iPhone 14 Pro Max 256G Deep Gray" appeared as "iPhone 14 Pro Max 256G Deep Gray" in one sheet, "iPhone 14 Pro Max 256G Gray" in another, and "14PM-256-GY" in a third. Three names for the same product.

Later I realized: Excel data looks tidy, but it hides countless "dirty data." If you don't clean it first, importing it into a WMS is a disaster.

闪仓 WMS · 示意图

First Pitfall: Thinking Excel Data Is Clean

My First Mistake: Direct Import

I thought I'd save time and dumped the Excel files directly into the WMS import tool. The system threw errors: SKU mismatch, quantity format error, date format error... The error messages filled three screens. I spent three days checking each one, only to find that "iPhone 14 Pro Max" alone had seven different spellings.

The Right Approach: Clean First, Then Import

I learned my lesson. In Shancang WMS, I designed a data cleaning process:

1. Standardize SKU naming: Rewrite all SKUs in the format "Brand-Model-Spec-Color", e.g., "Apple-iPhone14ProMax-256G-DeepGray".
2. Deduplicate and merge: Use Excel's "Remove Duplicates" feature to merge different spellings of the same SKU.
3. Standardize formats: Dates to YYYY-MM-DD, quantities to numbers, prices to two decimal places.

Table 1: Before and After Data Cleaning

This step took me a week, but when I imported later, zero errors.

闪仓 WMS · 示意图

The Right Approach: Clean First, Then Import

Second Pitfall: Ghosts in Historical Data

After standardizing formats, I confidently imported the live inventory. But the system showed a discrepancy of nearly 200 items compared to my manual records. I pulled out paper documents and checked each one, finding many inbound/outbound records that were never entered into Excel.

Even scarier, I found some data in Excel was wrong. For example, one SKU showed 50 units in stock, but only 30 on the shelf. When I asked my predecessor, he said, "Maybe they were stolen, or maybe the last count was wrong."

Later I realized: Historical data in Excel is often "fake data." Importing it without a physical inventory check is like magnifying errors ten thousand times.

闪仓 WMS · 示意图

Second Pitfall: Ghosts in Historical Data

My Second Mistake: Trusting Excel Blindly

I thought Excel was the "truth," but reality slapped me. According to a report by Fortune Business Insights^[1], over 60% of SME warehouse data contains errors of some kind. I realized I wasn't alone.

The Right Approach: Count First, Then Import

I spent two weeks counting every item in the warehouse. Then I imported only the "real inventory" from the physical count, using historical data only for reference, not for inventory calculations. Steps:

1. Full warehouse count: Count every item, record actual quantities.
2. Variance analysis: Compare Excel data to physical counts, identify reasons (lost, misrecorded, unrecorded, etc.).
3. Import only count data: WMS initial inventory based on count results; archive Excel data for backup.

Table 2: Before and After Physical Count

After the count, inventory accuracy went from 70% to 99%.

闪仓 WMS · 示意图

The Right Approach: Count First, Then Import

Third Pitfall: How to Migrate Without Stopping Operations?

The biggest headache: the warehouse can't close. Orders ship every day, returns come in. If I stopped for migration, I'd lose tens of thousands a day.

I initially had a stupid idea: operate during the day, migrate at night. But after a full day's work, I was too exhausted to move. A month passed with zero progress.

Later I realized: Data migration isn't a one-time event; it's a gradual transition. The best approach is "parallel running"—run both old and new systems simultaneously, then shut down the old one once the new one stabilizes.

闪仓 WMS · 示意图

Third Pitfall: How to Migrate Without Stopping Operations?

My Third Mistake: Impatience

I wanted to move all data at once, but the pressure slowed me down. Later, I used Shancang WMS's phased migration feature, dividing the warehouse into zones A, B, and C, migrating each zone separately.

The Right Approach: Phased Migration, Parallel Running

1. Phase 1 (1 week): Migrate Zone A (high-frequency items), new system handles Zone A orders, old system handles Zones B and C.
2. Phase 2 (2 weeks): Migrate Zone B, new system handles Zones A and B, old system handles only Zone C.
3. Phase 3 (1 week): Migrate Zone C, full warehouse switches to new system, old system read-only backup.

Table 3: Phased Migration Plan

This way, operations aren't affected, and you can verify the new system's accuracy step by step.

闪仓 WMS · 示意图

The Right Approach: Phased Migration, Parallel Running

Fourth Pitfall: Employees Can't Use It, So It's Useless

Data migration done, system live. But employees couldn't use it. Lao Zhang, who'd worked in the warehouse for ten years, only knew Excel and paper forms. I gave him a barcode scanner, and he said, "This thing's not as fast as my hands."

Worse, a new temp worker accidentally deleted the location data for a batch of goods, causing the entire warehouse to lose track of them. It took me three days to find them.

Later I realized: No matter how good the system, if people can't use it, it's scrap metal. Training is more important than technology.

闪仓 WMS · 示意图

Fourth Pitfall: Employees Can't Use It, So It's Useless

My Fourth Mistake: Ignoring Training

I thought the WMS interface was user-friendly and employees would figure it out. I was naive. According to Gartner's supply chain research^[2], over 70% of WMS projects fail due to low user adoption.

The Right Approach: Training + Permission Control

1. Layered training: Admin training (system configuration, reports), operator training (scanning, inbound, outbound).
2. Simulation drills: Run a complete process in a test environment, from receiving to shipping.
3. Least privilege: Temp workers can only scan, not modify inventory; experienced workers can handle inbound/outbound but not system settings.

Shancang WMS's permission management was a lifesaver—each role only sees and does what they need, greatly reducing error risk.

闪仓 WMS · 示意图

The Right Approach: Training + Permission Control

Summary

Migrating from Excel to WMS took me three months and ten pounds, but I ended up with a digital warehouse boasting 99% accuracy. Looking back, if I'd followed the right process from the start, I could have finished in a month.

Key Takeaways:

• Don't trust Excel is clean; clean before importing
• Don't trust historical data is real; count before importing
• Don't try to swallow it all at once; migrate in phases, run in parallel
• Don't ignore training; if people can't use the system, it's all for nothing

Data migration isn't the end; it's the start of digital transformation. I hope my blood-and-tears story helps you avoid the same pitfalls. If you're facing migration issues, feel free to reach out in the Shancang WMS community.

---

References

1. Fortune Business Insights WMS Market Report — Referencing SME warehouse data error rate statistics
2. Gartner Supply Chain Research — Referencing WMS project failure rate and user adoption correlation