"Johnny, we want to do digital transformation too. Should we buy a system first?"
I hear this thirty or forty times a year. I give the same answer every time: Not yet.
I'm not telling you not to transform. The order, "buy a system first," is itself the most expensive mistake. You haven't figured out where the company is leaking or where it's stuck, and you're already paying for a system that costs hundreds of thousands of NT dollars. The usual result: the system gets bought, people keep using Excel, veteran staff keep working from memory. Money spent, morale down, and in the end everyone agrees that "digital transformation doesn't work."
The problem was never that the system doesn't work. You just haven't reached the point where buying one makes sense.
Who I am, so you know why I get to say this
I'm Johnny Yang, founder of Brightstream Technology. I've been building companies for seven straight years. I haven't been doing anything else in that time. I've led my team through process and system implementations, one company at a time. The count is now 252+ implementation projects across 32+ industries: manufacturing, trading, food, hardware, logistics, services.
I'm not listing this to show off. I want you to know that what follows isn't jargon copied from a book. It's what I learned walking small and mid-sized businesses through the pitfalls one by one, losing time with them, and climbing out. I don't sell courses. I don't sell dreams. I just want you to skip the detour I've watched too many people take.
The short answer: step one isn't buying a system, it's finding where your company still runs on people holding it together
If you only remember one sentence from this post, remember this one:
The first step in digital transformation is finding the places in your company held up by one person's memory, one private Excel file, or one veteran employee's unspoken know-how.
Those places are your real risk. They're also where your transformation should make its first cut.
Think about it. Quotations have to go through Old Wang, because only he knows which customers get a discount. Whether inventory is accurate depends on whether the warehouse auntie felt like updating it today. Whether the books are right means waiting for the accountant to stay up three nights reconciling at month-end. These steps, the ones only one person knows, the ones that half-paralyze the company the day that person quits, are what digital transformation has to fix. A system is just a tool. The tool goes in last, not first.
How to actually take the first step: 5 steps
Step 1: map your processes and mark the duplication and the breaks
Don't hold meetings talking about "we need to transform." Take a sheet of paper. Follow one order from the moment it comes in, through shipping, to payment. Draw every stop: whose hands it passes through, what tool records it, where the data gets handed off.
You'll see two things right away: duplication (the same record typed into three places by three people) and breaks (half a conversation on LINE, verbal instructions, sticky notes, nothing a system can catch).
Mapping processes isn't paperwork. It's the first time you actually see how your company runs. Most owners are shocked when they finish: so this is how we really do things.
Step 2: separate what should be automated from what you should leave alone for now
Once you've mapped everything, you'll want to fix it all at once. Don't.
Some steps are repetitive, rule-based and high-volume. Those are the best candidates to digitize first, with the highest return: quotations, inventory, purchasing and sales. Other steps look messy but are actually held together by human judgment that keeps you flexible, like special terms for long-time customers or juggling rush orders. Forcing those into a system makes things worse.
Automate what is repetitive and rule-based, not everything. If you can't tell the two apart, you'll buy a system that locks you in. If Excel is still the main way your whole company is managed, read where Excel management hits its limit first. You'll see more clearly what Excel can handle and where it will eventually break.
Step 3: start with the one point that hurts most, not everything at once
The worst cases I've seen all start with the owner announcing "we're going fully digital." Then new systems, new processes, new tools, and the whole company relearning everything, all at the same time. Three months later the company falls apart and goes back to the old way.
The right way: start with the step that hurts most, goes wrong most often and causes the most arguments. Fix only that one. If inventory never matches, fix inventory first. If quotations keep going out wrong, fix quotations first. Get one point working, let employees see the benefit, and only then can you push the second.
Digital transformation is eaten one bite at a time, not swallowed whole. People who rush end up unable to eat a single bite.
Step 4: clean your data before you talk about which system
This step is the most boring. It's also the one most people skip, and it's where they fail.
If your customer records, part numbers and inventory counts are already wrong, duplicated, or one item is written three different ways, importing them into any system gets you an expensive set of wrong data. Garbage in, garbage out.
A system won't clean bad data for you. It just makes bad data move faster and go wrong more neatly. Before any system goes in, spend the time: standardize part numbers, deduplicate the customer file, do one full physical inventory count. It isn't sexy, but it decides whether everything you invest afterward is wasted.
Step 5: only now do you discuss whether to implement ERP
After the first four steps, you're finally standing at the door of whether to buy a system. By now you know where it hurts, whether your data is clean, and which processes belong in a system. Now when you look at ERP, you'll understand it, ask the right questions, and be able to compare options.
If you're at this step and want to understand what ERP actually is and how an SME should choose one, I wrote out the full logic in the complete guide to ERP implementation for SMEs. Many people think inventory software and ERP are the same thing. They're very different. The difference between inventory systems and ERP will help you tell them apart, so you don't buy a small tool thinking it's a big system, or spend big-system money on a small need.
ERP is the result of finishing the first four steps, not a shortcut for skipping them.
Which stage are you at? A one-minute self-check
| Your current situation | Your stage | What to do next |
|---|---|---|
| Everything runs on veteran staff's memory, LINE and sticky notes | Not started | Do step 1: map your processes |
| Lots of Excel files, each kept separately, often don't match | Half-digital, real mess | Clean the data + find the point that hurts most |
| Small software on one step, everything else by hand | Partially digitized | Connect the data so systems aren't isolated islands |
| Want an integrated system but don't know how to choose | Ready for ERP | First understand costs and how to select |
Find your row and you'll know which part of this post to spend your effort on, instead of trying to solve everything at once.
The two most common traps
Trap one: transforming for the sake of transforming. A competitor installed a system, "digital transformation" sounds trendy, so you think you need one too. The motive is "others have it," not "this is where it hurts." Nine out of ten of these fail. Transformation should fix a specific pain, not the owner's anxiety. CRM is ground zero for this. So many companies bought one and nobody uses it. I put it bluntly in why CRM implementations fail. Read it. A lot of those mistakes you can still avoid.
Trap two: trying to change everything at once. I said this above. I'm saying it again because it has sunk too many companies. People's attention is limited. Employees' patience is limited. Your cash flow is limited. Doing it all at once means betting that three things succeed at the same time. The odds are too low to be worth it. Fix one point at a time and build from there. That's actually the fastest way.
If you run a service business, a restaurant or a retail store, where people are the capacity, step one usually isn't inventory. It's staff scheduling. That's what burns your managers' time every week and directly drives labor costs. I wrote about it separately in scheduling driving your service business crazy? Three principles of staff scheduling.
FAQ
Q: Does digital transformation require spending big on a system? Not necessarily, and I'd advise you not to rush into spending big. The early steps, mapping processes and cleaning data, cost almost nothing, yet they decide success or failure. When it's really time to spend, know exactly where the money goes. I laid out ERP cost structure in how much ERP implementation actually costs. Understand first, then spend.
Q: Our company is small. Do we need digital transformation? Yes. But for you, "digital transformation" might just mean "stop letting the company run on one person's head." Small is actually easier: less baggage, and a small boat turns quickly. Sort out your processes early, and your own mess won't hold you back when you grow.
Q: Cloud or our own servers? Should we think about that now? That's a question for later, during selection. Don't get stuck on it at step one. When you really need to compare, I wrote up the differences and who each one suits in choosing between cloud ERP and on-premise ERP. Read it when you get there.
Q: Everyone's talking about AI. Should I go straight to an AI ERP? Stand steady before you run. AI is an accelerator. It speeds up processes that are already in order. If the process itself is a mess, AI just makes you messy faster. To see how traditional ERP and AI ERP differ and whether you're ready, read the difference between ERP and AI ERP.
One last honest word
Digital transformation isn't something you buy. It's something you have to see clearly. Once you see where your company is being held together by people, you're already ahead of half the owners rushing to pay.
If you get stuck halfway through mapping, or want someone to walk through your processes with you, I offer a free 30-minute diagnostic. No sales pitch, no system to sell. Just helping you see where to take the first step. If you want to talk, pick a time slot with Johnny Yang. Choose online and it's booked.




