When to replace Excel with an ERP or system: 5 signals you can't put off

Excel works fine, so should you switch? The answer isn't about headcount. It's whether you've hit these 5 signals. Includes a self-diagnostic table you can fill in yourself, and one reminder many people miss: don't rush into a big system.

Johnny Yang18 min read
When to replace Excel with an ERP or system: 5 signals you can't put off

“Johnny, my Excel works fine. Why would I switch?”

I hear this about thirty times a year. Every time, I don't rush to argue. I ask one question first: “Last week, did you run into a case where the same record didn't match between two files?”

Nine out of ten pause, then smile wryly: “Yeah... the item number got mismatched again.”

Good. Now we have something to talk about.

This article isn't here to scare you into switching systems, and it's definitely not here to sell you a big platform costing hundreds of thousands of NT dollars. I'm Johnny Yang of Brightstream Technology. Over the past few years my team has done 252+ implementation and optimization projects across 32+ industries — multi-channel e-commerce, traditional manufacturing, food trading, and old-line factories that are hidden champions in their niche. About seven years in total. After seeing that many, here's the honest thing I most want to tell you:

Whether to replace Excel with an ERP or a system has never been about “how many people your company has.” It's about whether your operations have already hit the tipping point of “relying on someone's memory” and “relying on someone manually syncing across multiple files.”

Headcount is surface. I've seen three-person teams that should switch, and I've seen companies with revenue in the hundreds of millions of NT dollars and fifty-plus people where Excel can still be forced to hold — but holding doesn't mean no risk. More on that later. Today I'll break this down: five signals you can't put off any longer, a diagnostic table you can fill in yourself, and one reminder a lot of people miss — don't rush into a big system.

The conclusion first, in case you don't read to the end

If you only take away one sentence, take this one:

When the correct answer to something depends on what's in one person's head, or on someone manually cross-checking several Excel files to piece it together, Excel has hit its limit.

Everything else expands on that sentence. Excel itself is a good tool. I still use it every day. Its problem isn't whether it can calculate. It's that it can't guarantee “one record, one version,” and it can't sync automatically across multiple people, files and processes. Once those two things start breaking, even your most diligent employee can't patch it, because it's a problem with the tool's structure, not with people not trying hard enough.

OK, let's go through the signals one by one. Check them against your own company as you read, and count how many you hit.

Signal 1: The same data starts having several “correct versions”

This is the most classic first crack, and the easiest one to miss.

Have you had this happen: a quotation needs the price list, and accounting has one copy, sales has another, and the boss has a third on his desk. The prices on all three are different, and everyone insists theirs is the latest. Or product data: the warehouse Excel lists one spec, the website back end lists another, and customer service hand-copies yet another for the customer.

I worked with a multi-channel client with over 10,000 SKUs. Just listing one new product meant opening four Excel files and two documents to cross-check. Why so many? Because the data was scattered: some files had the product name but no item number, some had the item number but a messy brand column, and the price sat in yet another sheet. To assemble one complete, correct record for a product, staff had to flip through six files like a jigsaw puzzle and piece it together by hand.

Guess what happened. Item numbers were often mismatched. And nobody noticed at the time. It usually blew up only after the customer placed an order, and then the whole document was voided and redone.

What does this mean? It means your data no longer has a single source of truth. When one record exists in multiple versions, “which one is right” becomes something people have to judge, remember, and guess. And as long as people are guessing, one day they'll guess wrong — the only question is which order goes wrong today. I go into this pain in more detail in Where Excel management hits its limit. Read it alongside this one.

Signal 2: You have to cross-check several files to get one answer (human VLOOKUP)

The second signal is the first one's sibling, but it goes a step further.

If signal 1 is “the data has several versions,” signal 2 is “even getting one answer means manually searching across files.” The typical scene: you want to know how much customer A actually shipped this month and how much they still owe you. Your employee opens the shipping Excel, the reconciliation Excel and the inventory Excel, then does VLOOKUPs by hand, or just compares them line by line by eye.

I call this “human VLOOKUP.” What makes it dangerous is that it usually looks like it works — the employee is fast and finds the answer quickly. But have you counted how many times a day this happens? Five minutes each, twenty times a day, is a hundred minutes. Over a month, that's thirty-plus hours. Not to mention the error rate, which climbs exponentially as the files and data grow.

The 10,000-SKU client above is a living example. It's not that they couldn't use Excel. The opposite — they were very good at it. Good enough to force a listing process out of four files and two documents. But that's exactly the problem: when daily work requires “a skilled person” plus “a cross-checking method only they understand,” what you depend on isn't a system. It's that person.

What does this mean? It means your operational efficiency is locked to people manually moving data around. The whole point of a system is to turn questions like “how much did customer A ship this month and how much do they owe” into two clicks and a guaranteed-correct answer. When basic questions like that still need someone searching across files, and still come out wrong, that's a clear signal the tool needs upgrading. Inventory is where this shows most. I break it down for you in What's actually the difference between inventory management and ERP.

Signal 3: One record is entered two or more times, and errors are common

The third signal is one many owners look at every day without realizing it's the disease.

Watch your staff. A customer places an order. They type it into the order Excel. Then, to notify the warehouse, they type it into the shipping Excel. Then, for reconciliation, they type it into the finance Excel. The same order, the same product name, quantity and amount, entered by hand three times in one day.

Duplicate entry is wasteful enough on its own, but the real damage is this: every extra entry is one more chance to mistype. And the three files aren't linked. Change file A, and file B doesn't follow. A customer changes the quantity at the last minute, the employee updates the order file but forgets the shipping file, and the warehouse ships the old number. Wrong goods go out, a complaint comes in, and you get returns, reshipments, reissued orders — the whole chain of costs.

The nastiest thing about these errors: they aren't big errors. They're lots of small ones. Each looks like a minor slip — “oops, an extra zero,” “oops, this item number is missing a digit.” But add them up and you're redoing dozens of orders a month, your staff's morale is worn down, and you can't even find a culprit to blame. Because there isn't one. The process itself is producing the errors.

What does this mean? It means your process is full of repetitive work that a computer should be doing but people are doing instead. The right way to handle a record is “enter it once, and it flows through the whole process automatically”: the order is confirmed, inventory is deducted automatically, the shipping slip is generated automatically, the receivable is booked automatically. If you're still having people manually move the same record to three places, you're not just wasting hours. You're systematically producing errors.

Signal 4: Only one person and one private Excel file can make sense of it, and when they take leave everything stops

The first three signals are about efficiency and errors. The fourth is about risk — the kind that's most often underestimated.

Almost every company has this person. Maybe an accountant with ten-plus years in, maybe the veteran running the warehouse, maybe the owner's wife herself. They have one, or several, Excel files that make anyone else dizzy, but they can run them with their eyes closed. The sheet hides a pile of rules nobody else knows: why this column is filled in this way, what this color means, why this formula takes this detour. The whole logic exists only in their head.

Normally nothing goes wrong. But when they take leave, that whole step stalls. If they quit or retire, it doesn't stall. It just goes dark.

Let me tell you about a case that stuck with me. A hidden-champion company, doing very well, revenue in the hundreds of millions of NT dollars, one of the top players in its niche. When I went in, their core operations still ran on paper plus Excel, and how the whole thing ran depended entirely on the memory of a few senior veterans. Which customer has what special rules, which part can't be mixed with which, how to set a quotation, how to work out delivery dates — all in the veterans' heads. Not one word written down as process.

Can it hold up? Sure. It's held for decades and makes serious money. But what I saw wasn't stability. It was an unexploded bomb: if the succession of any one of those veterans goes badly, or one of them retires and leaves, a large chunk of the company immediately can't be handed over. That's not fearmongering. It's a matter of time.

What does this mean? It means your company's most critical knowledge assets aren't in the company's hands. They're in a few people's heads and private files. One of the real values of systematizing is turning tacit knowledge that only one person understands into a written process that anyone can take over and run. This is tied directly to succession and handover. I wrote a separate piece, How to digitize your operational knowledge before succession and retirement. If your company has a “key person” like this, I strongly recommend reading it.

Signal 5: Problems are always found after the mistake, never before

The last signal is the one I think best separates “Excel is enough” from “time to switch.”

Let me ask you: of the problems your company runs into, are more of them stopped beforehand, or found only after something has gone wrong?

For companies on Excel, the answer is almost always the latter. Because Excel is passive. It won't warn you that “this stock is about to run out,” “this item number doesn't match the database,” “the amount on this order looks off.” It just sits there. It takes whatever you enter, wrong or not. So problems always blow up at the very end of the line: the customer orders, then you find the item number was wrong. You ship, then you find inventory was actually negative. You reconcile at month-end, then you find you missed issuing several invoices.

Back to the 10,000-SKU example — why were mismatched item numbers always found only when the customer ordered? Because before that, there was no checkpoint to tell you “this one is wrong.” A manual cross-checking process has no such thing as real-time validation. The error flows freely all the way to the customer, and the customer catches it for you. Think about it: making your customer your quality control is an expensive arrangement.

That's the most fundamental difference between a system and Excel: a good system catches errors at the moment you enter them — a nonexistent item number won't save, insufficient stock triggers a warning, an abnormal amount gets flagged red. It moves quality control from “fixing after the error” to “preventing before the error.”

What does this mean? It means every error cost you're carrying now is the most expensive kind — because errors that surface at the end of the line, in front of the customer, cost the most to fix and do the most damage to your reputation. If you find most of your company's problems are discovered after the fact, Excel's passive nature is already dragging you down.

Quick self-diagnostic: how many do you hit?

That's five signals, and it may still feel a bit fuzzy. So here's a table. Take three minutes, check it against your company, and tick the ones that apply.

No.SignalDoes this apply to you?Risk
1The same data (price list, product data, etc.) has two or more versions that “everyone thinks is the latest”☐Wrong version used, wrong price quoted
2Getting one answer (how much a customer owes, how many units of an item are left) means opening three or more files and cross-checking by hand☐Low efficiency, cross-checking errors
3The same order or record is entered by hand two or more times a day☐Rework, data out of sync
4There's a sheet or process only one or two specific people understand, and it stops when they take leave☐Handover gaps, succession time bomb
5Most of the company's problems are found after the error, not stopped beforehand☐Errors reach the customer, highest cost

How to read it:

  • 0–1: Your Excel is still in the healthy range. Keep using it, but save this table and check yourself once a quarter.
  • 2–3: Yellow light. It's not an emergency yet, but cracks have appeared. Time to start seriously mapping your processes and planning. Move now and it's your choice. Keep putting it off and you'll be forced to move.
  • 4–5: Red light. Your question isn't whether to switch. It's how long you can hold out before something big goes wrong. Honestly, if you've made it this far, it's mostly because your staff are carrying it for you, and that kind of carrying gives out sooner or later.

Busting a myth: “holding up” doesn't mean “no risk”

At this point I have to puncture the most common myth, because it makes too many owners miss the best moment to act.

A lot of people tell me: “Johnny, I get everything you're saying, but I've done it this way for ten-plus years and the company hasn't gone under. It's fine, isn't it?”

I completely understand that mindset. People easily mistake “nothing has gone wrong yet” for “nothing will go wrong.” But those two things are worlds apart.

Look back at the hidden champion with revenue in the hundreds of millions of NT dollars, still running on veterans' memory. Is it holding up? Yes, and holding up beautifully, making a lot of money. But is it free of risk? Not at all. It has staked the company's lifeline on the assumption that a few people won't suddenly leave, won't get sick, won't get poached, and won't misremember. The longer that bet runs, the more it costs to lose — because the bigger the company and the better the business, the bigger the payload of that unexploded bomb.

“Holding up” describes the present. “Risk” describes the future. A company can be holding up now and carry frightening risk for the future at the same time. There's no contradiction. The companies that really get into trouble often aren't the ones that were shaky from the start. They're the ones that were “fine all along, until one day they suddenly weren't” — because they treated every warning signal as noise.

I'm not trying to make you panic. I want you to turn “still holding up” from a reason not to switch into a reason why now is the calmest time to switch. Upgrade while you're still holding up, and you have time and room to do it slowly. Wait until you can't hold up, and you're fighting the fire while it's burning, at several times the cost.

This article is about switching from Excel to a system. If you already have a system running and just feel it's getting worse, that's a different set of signals and a different decision — whether and when to replace an old system, which I list in When should you replace an old system? Five signals it's time. The criteria in the two articles are different. Don't apply the wrong one.

But — don't rush into a big system

OK, if you've read this far and you're thinking “uh-oh, I hit three or four,” I'm going to hold you back first and say something you might not expect:

Don't rush out and buy a big ERP costing hundreds of thousands, or over a million, NT dollars.

In my line of work, I'm supposed to encourage you to buy as big as possible. But I've seen too many companies that hit only two or three signals get talked by a salesperson into an oversized system. The result? Ninety percent of the features go unused, the implementation is chaos, staff resist it, the owner regrets it. The money's spent, the pain's endured, and the problem still isn't solved. I've gone in to put out fires on cases like this about as often as I've done implementations from scratch.

The right order is this:

Step one: map your processes first, not pick a system first. A system is just a tool, and a tool exists to carry a process. If you can't clearly say what your processes look like, where they get stuck, and where work is duplicated, then no matter how expensive the system, it won't fit — because you're just moving a messy process to a more expensive place to stay messy. Spend a week or two drawing out your order flow, inventory flow and reconciliation flow. Just doing that, many owners spot a pile of redundant steps they can cut on the spot. I'm very blunt about this in The first step of digital transformation for SMEs. Read that one before you decide.

Step two: treat the actual problem, and don't fire a cannon at a mosquito. Once you've mapped your processes, you'll find many companies' pain points are quite concentrated. If your core pain is “inventory and orders don't match,” a well-designed inventory management system may solve 80% of your problems, and you won't need a full ERP at all. ERP is heavy equipment that ties finance, inventory, production and HR together. It suits companies with complex processes and lots of cross-department work. If your needs are still simple, forcing in an ERP is firing a cannon at a mosquito. Whether inventory management is enough or you really need ERP, I sort that out with real scenarios in two articles: What's actually the difference between inventory management and ERP and The complete ERP selection guide for SMEs.

Step three: if you're multi-channel, pay special attention to integration. If you run multiple channels like the 10,000-SKU client above, the core problem you need to solve is really “how to integrate inventory and product data across channels into one version of the truth.” That's a different level of problem. Don't apply single-channel thinking to it. I wrote a separate piece on this, How to integrate multi-channel inventory. Multi-channel owners, read it.

One sentence to sum up this section: switching systems isn't “buy the most expensive one.” It's “understand your own disease first, then prescribe just the right medicine.” Mapping processes is always step one. Choosing tools is step two. Get the order backwards and the money is wasted.

FAQ: the questions owners ask me most

Q: My company only has five or six people. Is it still too small to need a system? Headcount really isn't the criterion. None of the five signals above are about headcount. Five or six people, but you're already doing human VLOOKUP, entering data twice, and a key process is understood by only one person — then you need to deal with it. On the other hand, even with twenty or thirty people, if your processes are simple and you haven't hit those signals, Excel can still hold. Judging by the diagnostic table is far more accurate than judging by headcount.

Q: Isn't switching systems expensive and a hassle? Can SMEs even afford it? This is one of the biggest misconceptions. “Switching systems” doesn't mean “a million-NT-dollar ERP.” There are plenty of monthly-subscription inventory or lightweight systems built for SMEs with just the features you need, and the barrier is much lower than you think. What's really expensive is the hidden cost of never switching — the hours spent on rework, compensation for errors, the reputation lost to complaints, and that handover time bomb that could go off any time. Add those up once and you'll find not switching is the most expensive option. The key is to map your processes first, treat the actual problem, and not buy something oversized.

Q: My staff are older. Will they resist, or be unable to learn? They'll resist. That's normal. But the reason usually isn't that they can't learn. It's that they weren't guided properly, and they're afraid of being replaced. Where system implementation most often fails isn't the technology. It's the people. My approach: involve staff in mapping the processes from the start, so they feel it's “making our work lighter together,” not “the boss wants to replace me.” Pick tools with simple interfaces that stay close to how they already work. Guide them right, and a sixty-year-old veteran will use it smoothly — I've seen it too many times to count.

Q: So where should I actually start? Three steps. First, go through the diagnostic table above honestly and see how many you hit, red light or green. Second, if you hit two or more, spend a week or two drawing out your core processes (orders, inventory, reconciliation) and find where the real pain is. Third, take that process map with you when you decide between inventory management and ERP, and which spec. Diagnose first, map processes next, choose tools last. Get that order right and you won't fall into the trap.


Excel isn't the enemy. It's a good partner. Every tool just has its limits. If you're reading this article, you probably already have a vague sense that something isn't quite right — and that instinct is usually correct. If you hit two or three signals and it's making you uneasy, you're welcome to book a free 30-minute diagnostic with Brightstream Technology. Just pick a time that works for you online. We don't do sales pitches and we don't sell courses. We bring seven years, 252+ projects and 32+ industries of experience, and help you get clear on whether you should switch and how. You bring the problem, I'll give you a direction, and this half hour won't cost you a cent.

Found this useful? Pass it on
Prices published · free range estimate

So what does a build to your process actually cost?

From NT$300,000, priced against your spec, with same-scope changes free during the maintenance period. The prices and the estimator are in the first screen — get a range first, you do not have to decide anything yet.

Written scope, then a fixed price · staged acceptance · the code and the data are yours · or just book a free 30-minute diagnostic

Johnny Yang (熊董)
Johnny Yang (熊董)
Founder, Brightstream Technology

Serial founder leading a small team that builds systems into an owner's everyday work.

As a founder
7 yrs
Projects delivered
252+
Industries served
32+
About Johnny