"Johnny Yang, just getting one new product listed on all our channels takes two employees working full-time for six months. Can you believe that?"
Of course I can. I've seen it with my own eyes. No theory today. I'll walk you through one real case, and there's a good chance you'll see your own company in it.
First, who I am: I'm Johnny Yang. Seven years of back-to-back startups. My team has delivered 252+ implementations across 32+ industries. This case is one of the ones I remember best, because it took "multi-channel retail inventory hell" to the extreme.
A normal day at this company: 6 files open to list one product
This is a retail and e-commerce company with more than 10,000 SKUs. Every new product has to go up on their own website plus several marketplaces at once: Shopee, Ruten, MOMO (Taiwan's main e-commerce platforms) and others.
Sounds manageable? Look at the actual process. To list one product, an employee opens 4 Excel files + 2 documents and cross-checks back and forth between them. That's where the hell is hiding:
- Some files have the product name but no SKU;
- Some files have the SKU, but the brand name is typed any old way;
- The price might be in yet another file;
- And at the end, someone still has to manually check the part number in the ERP to confirm it's actually the same item being listed.
All of it runs on human eyes and human memory. You can guess the result. Often the customer has already ordered online before anyone notices the SKU is wrong. Then it gets worse: the system uses the wrong SKU to generate the sales order and shipping order, and both are wrong from top to bottom. Someone has to stop them by hand, recheck everything, and redo the whole thing.
And that's just listing and shipping. Their in-store POS, the ERP, and online inventory don't talk to each other at all. The store sells one, the website doesn't know. The website sells one, the store's stock count is off. The same batch of goods has three sets of books inside the company, and none of them match.
E-commerce listing alone took two full-time employees a full six months. And they ran like this for years and survived. Honestly, I respect that. Keeping the business alive under this much internal friction means their products and their market are genuinely strong. It also means a lot of profit and labor was being eaten by this process for nothing.
If this section felt uncomfortably familiar, get one thing straight first: this isn't a question of whether an inventory system or an ERP is better. It's a question of whether your systems are connected. I explain the line between the two in inventory system vs ERP: what's the difference. Start there for the basics.
Why "hire more people, try harder" never fixes it
Most owners facing this react the same way: "Hire one more person." "Tell everyone to be more careful." I get it. But that road goes nowhere.
Because the real problem isn't too few people or people not trying. The process itself is structurally built to produce errors. As long as the same product data is scattered across 6 files and bridged by someone's brain, errors aren't accidents. They're guaranteed. One person has a good day and makes fewer mistakes. The next day they're on leave, someone else takes over, and it's a different disaster. Adding headcount just makes the errors happen faster and in larger numbers.
This is the same root cause as companies that run on Excel until it breaks. Excel isn't the problem. When data has to be synced by hand across multiple files, the person becomes the bottleneck of the system. I listed the warning signs of that tipping point in how many people can Excel actually run a company for. Check yourself against it.
How we solved it: we didn't replace the ERP, we connected it with APIs
This is the key part, and the core idea I want you to remember.
One fact tells you how hard this problem was: this company had already tried to fix it. Before they found us, they had hired several different teams to overhaul things, and the money burned added up to more than NT$5,000,000. Every attempt took the same approach: "implement a brand-new system." The result? Every single one failed. Money gone, staff worn out, morale ground down, and the problem still exactly where it was.
Why did they keep failing? Because the direction was wrong from the start. Their ERP had been running for years, tied to a pile of processes and historical data. Tearing it down and starting over touched too much and carried too much risk. The migration alone could have shut the company down. So every "switch to a new system" attempt ran out of steam halfway and fell apart.
A lot of owners think a failed ERP implementation is their own fault, or that their staff wouldn't cooperate. It isn't. The vendor's direction was wrong. In situations like this, most system vendors reflexively tell you to "switch everything over to ours," and that is exactly the path most likely to fail and burn the most money. (I covered the other common reasons in why ERP implementations fail.)
So we didn't touch their existing ERP. Here's what we did:
- Connected their existing ERP through APIs. The ERP keeps running as before. We made data flow in and out automatically instead of telling them to throw it away and start over;
- Replaced the store POS with one that can sync, and added AI auto-entry and website traffic routing. Every in-store movement shows up in the same inventory in real time, nobody retypes data by hand, and in-store customers get pointed back to the website along the way;
- Added our automated workflows so that inventory and listings across the website and every e-commerce channel are fully automated. Product data is created once and synced to every channel automatically. SKUs, names and prices no longer depend on someone cross-checking files.
The result? All they do now is review and confirm. Everything else is automatic.
This is what I keep telling owners: you don't necessarily have to replace your system. Often it's enough for me to connect it with APIs and automate the repetitive work. This ties closely to the difference between a regular ERP and an AI ERP. The system should be doing the work for you, not waiting for you to wait on it. I wrote about that in 5 differences between a regular ERP and an AI ERP and what an AI ERP can do for you.
The numbers: what was actually saved
I know owners want numbers. After the rebuild, here is my rough estimate of what changed:
| Item | Before | After |
|---|---|---|
| Listing one product | About 30 minutes | About 1 minute |
| Checking SKUs on one order | About 10 minutes | About 1 minute |
| E-commerce listing staff | 2 employees full-time for six months | Review and confirm only |
| Store / website / e-commerce inventory | Three separate records that don't match | One inventory, synced automatically |
| Staff hours saved per year | — | Roughly NT$2,000,000+ |
Listing went from 30 minutes per product to 1 minute. Order checking went from 10 minutes per order to 1 minute. By a quick estimate, staff hours alone save more than NT$2,000,000 a year.
But I'll be honest about one thing: the money saved isn't really the point.
The real point is this. When your employees no longer spend every day on repetitive, mechanical work where one mistake means redoing the whole batch, that exhaustion, that mental drain, is gone. Repetitive, meaningless work grinds the soul out of people. Hand it to the system, and people have energy left for work that actually needs judgment and a human touch. That value doesn't fit in the table above, but it's the most expensive thing on the list.
The real payoff sits behind the cost savings: members and repeat purchases
If integration only matched up inventory and automated listings, it would be a strong cost-cutting tool and nothing more. But once the store POS, website and e-commerce channels are truly joined into one system, something more valuable happens: online and in-store members become the same person.
Before, everything was split. A customer paying by card in the store was one identity. Ordering on the website was another. Buying on Shopee was a stranger. You had no idea that the regular who kept coming into the store was the same person as some account on the website. So every sale was a one-off. The customer bought and disappeared into the crowd.
After integration, this company can do this: membership cards, points redemption and coupons all work and can be redeemed on every channel.
It sounds ordinary. The effect is big. It makes one thing possible: a customer can try the product in the store first and touch it before buying. Back home, their membership is still there and their online coupon still works, so they naturally come back to the website to buy again. Offline handles experience and trust. Online handles convenience and repeat purchases. The two stop competing and start handing off to each other.
That produces two concrete business results:
- LTV (customer lifetime value) gets longer. A customer isn't done after one purchase. Points, coupons and membership keep pulling them back, and one purchase becomes a long-term relationship.
- Dormant old customers get woken up. Leads who bought once years ago and were never heard from again can now be woken up again and again through cross-channel membership and promotions, turning from dead data into loyal fans who keep buying.
That's why I say inventory integration is only the entry ticket. Members and repeat purchases are the real payoff. You think you're fixing an operations problem called "listing is exhausting." What you end up with is a growth engine that keeps turning old customers into revenue. Plenty of vendors who only want to sell you a system will never even bring this up.
Pitfalls you might hit, and the one question to ask before signing
If you want to integrate multi-channel inventory too, here are two pitfalls to flag up front:
Pitfall 1: thinking "buying a new system" will fix it automatically. It won't. If your underlying product master data (SKU, name, price) is already a mess, the most expensive system in the world just moves the mess to a new home. Clean up the master data before implementation. Cut corners here and everything after it fails. I wrote up the most common ways this goes wrong in why ERP implementations fail.
Pitfall 2: getting talked into "tear it all down and rebuild." Unless your current system is truly beyond saving, connecting your existing system through APIs is usually lower risk, faster to launch, and cheaper.
So before you sign, ask the vendor one question. That single question filters out half the vendors who aren't a fit:
"I don't want to replace my current ERP. Can you connect to it through APIs and only automate the part that hurts most right now: multi-channel listing and inventory sync?"
Vendors who can do it will sit down and seriously discuss how to connect. If they only want to sell you a whole new system and lose interest the moment you say "I'm not replacing my ERP," you have your answer.
For a fuller way to decide where to start, read the complete ERP guide for SMEs first. It lays out the whole path.
One more thing to be clear about: sometimes multi-channel numbers don't match and it has nothing to do with the channels. Your books simply don't match what's on the shelves. However well the systems are connected, if the opening inventory you feed in is wrong, what comes out is still wrong. I listed six real sources of count discrepancies in stock counts never match? Six sources of book-vs-physical gaps. Do one round of reconciliation before integrating. Get the order backwards and you'll think the integration was done badly.
FAQ
Q: I don't have 10,000 products. Do I still need inventory integration? It's not about how many products you have. It's about whether you have multiple channels each keeping their own books. Even with a few hundred products, if your store, website and e-commerce stock don't match and you keep overselling or missing orders, you're already paying the cost of that friction. You just haven't been pushed to the breaking point yet.
Q: Do I have to replace my current ERP or POS? Not necessarily. An ERP can usually be connected through APIs and left alone. The POS depends on whether yours can connect to anything outside it. If it's an old, completely closed POS that won't let data out, you may need to swap it for one that syncs. That's so inventory can update in real time, and it's worth it.
Q: Is this kind of integration risky for my data? Will it be locked up with you? With good integration, the data should still belong to you. Before connecting anything, ask clearly: can the data be fully exported, in what format, and can you take it with you if you stop working together? I also stressed how important this is in cloud ERP vs on-premise ERP: how to choose.
Q: How long does implementation take? Will the company grind to a halt halfway through? Because we connect through APIs and leave the existing ERP alone, the risk is usually much lower than replacing everything, and the company keeps running while we connect. What really takes time is cleaning up the master data at the start. The more solid that work is, the smoother everything after it goes.
If you read this and realized it's describing your company — store, website and e-commerce each keeping their own books, listing and order checking grinding your team down every day — we should talk. I offer a free 30-minute process diagnostic. Pick a time online and book directly. Tell me about the part that's most stuck right now. My team has done 252+ projects across 32+ industries, and I'll tell you straight: what can be connected with APIs, what should be automated, and roughly how much labor it could save you. No sales pitch, no pushing you to switch systems. Just helping you see clearly where the friction is.




