Picture this. You're sitting in a meeting room. Three people across the table, sharp suits, beautiful slides. Halfway through, one of them says: “Johnny Yang, don't worry about any of this. Just leave it all to us. Your team doesn't need to do any testing, feedback, or optimization. We'll handle it. You just wait for go-live.”
If your first reaction is “Finally, someone who can carry this” — please, calm down for three seconds first.
Any vendor who promises “your team won't need to be involved at all” is almost certainly lying.
I'm Johnny Yang, head of Brightstream Technology. Over the past seven years I've led my team through 252+ system implementations across 32+ industries — manufacturing, trading, retail, food service, services. This post isn't selling a course or pushing a package. It's every pit I've fallen into, watched others fall into, or pulled clients out of, laid out in one place. Because picking the wrong vendor is more fatal than buying the wrong system: a great system is useless with the wrong vendor. Being good at selling is not the same as being able to deliver.
1. The conclusion first: implementations fail at vendor selection, not system selection
Plenty of owners spend months comparing systems, quotations, and feature lists. They end up picking a decent system — and the implementation still fails. It never goes live, and staff keep sneaking back to Excel.
Where's the problem? You treated it as buying a product.
An ERP implementation has never been a purchase. It's a co-creation project between your staff across departments and the vendor: interview → test → feedback → optimize, together. You're not buying software. You're buying whether someone is willing to move your workflow into the system with you.
And the market is full of vendors who are great at selling and have zero plan to actually do the work. Their decks beat everyone's, they close contracts faster than anyone, but when it's time to get hands dirty, roll up their sleeves, and untangle your processes one by one with you, they're gone.
So the point of vendor selection isn't who sounds best. It's who will actually get in and do the work with you. This post starts with how to spot the ones who only sell. If you want to understand why implementations fail in the first place, I wrote a fuller piece: The real reasons ERP implementations fail. Read them together.
2. Four red flags: screen out the dangerous vendors in the first meeting
2-1. Red flag one: promising “zero involvement from you, leave it all to us”
Saying it again, because it matters that much.
Think about what an ERP touches: payroll, bonuses, accounting, inventory — all core financial processes. How you calculate pay, where your bonus tiers sit, how costs get allocated, how stock in and out gets recognized. That logic lives only in your team's heads and a few old Excel files. No vendor can know it on day one.
Someone who hasn't even asked about your bonus formula — what gives them the right to promise it will work without you being involved?
That promise means one of two things: they don't understand, or they're lying. You want neither. People who have actually done this will tell you straight: “Sales, accounting, warehouse — every stage needs someone from your side working through it with us. It will cost you time, but it can't be skipped.” Sounds exhausting. That's what a project that actually succeeds looks like.
2-2. Red flag two: only salespeople ever show up
First meeting: a salesperson. Second meeting: still a salesperson. Ask a detailed question, and the answer is “Let me check with the engineer/consultant.”
Danger. It means the people who will actually do the work won't appear in front of you until the deal is signed. Every promise you heard came from someone who won't be working on your project.
The healthy version: partway through the talks, the consultant or PM who will lead your implementation shows up, and asks “dirty” questions — how you do warehouse stock counts, where month-end closing gets stuck, which department drags its feet the most. People who ask dirty questions are the ones getting ready to do the work.
2-3. Red flag three: the whole pitch is a feature list
“Our system has inventory and order management, production management, BI reports, mobile approvals, 380 feature modules…”
Sounds rich, right? But notice: the whole time, they're talking about what the system has. They never ask what your work looks like.
This is the dividing line I most want to call out, and the next section is all about it. For now, remember: a vendor who shows features and never asks about process is selling a catalog, not implementing anything for you.
2-4. Red flag four: the post-signing blame triple
You can't see this before signing, but you can ask about it up front and get it in writing. The blame triple is the vendor's three standard responses once the system turns out to be hard to use:
- Asking you to pay more — “That's a separate customization fee.”
- Saying it's not in the contract — “You didn't buy that module.”
- Saying the system doesn't do that — “That module never had that.”
Three moves, and every “hard to use” lands back on you. By the time you notice, the money is paid, the system has you locked in, and there's no way back.
How do you protect yourself? Before signing, nail down: if it doesn't work as expected after go-live, how it gets adjusted, who's responsible, and whether it costs extra. How to negotiate fees and where the hidden costs are, I've laid out here: ERP implementation cost: where the money actually goes.
3. Feature training vs. workflow integration: the line between used and unused
I'm spending the most space on this section, because it's the real source of 90% of implementation failures.
Many vendors only teach you what features the system has. They don't help you build your workflow into the system. Those two things are worlds apart.
Feature training looks like this: “Click New here, enter the part number here, press Save here, this is the report export.” Done. Here's an operations manual. “Enjoy go-live.”
Workflow integration looks like this: “Right now, for a purchase order, who opens it first? Who signs after that? After signing, why does it get printed and typed in again? Can that step go? I'll connect it so it goes out for approval in one click, and your warehouse staff scan the barcode on receipt to post it back directly.”
See the difference? The first throws the system at you and leaves you to figure out how to cram your work into it. The second takes your work apart and shapes the system to fit it.
Vendors who only teach features run, at bottom, a “dump and run” business: hand you the system plus a manual, then stop answering. It's the classic packaged-software profit model: they sell licenses, not results. They don't need your implementation to succeed. They just need you to pay.
And what you want was never “a system that runs.” You want staff who actually use it, and are faster with it than before. The gap between those two gets filled by whether someone is willing to get in and build your workflow into it.
The test is simple. In the meeting, hand them your company's most painful process. See whether they rush to say “our system has feature XX for that,” or first ask back, “Why does your process run this way?” The second one is someone who can deliver.
4. Why implementation must be cross-department co-creation, no exceptions
Some owners think: can't I just make the call myself and let the vendor handle it? I'm paying, after all.
No. And the reasons are hard ones.
ERP touches payroll, bonuses, accounting, inventory — core processes that cut across multiple departments. Payroll and bonuses involve HR and finance. Inventory involves the warehouse and purchasing. Cost involves production and accounting. How each department actually works, and every small “why we do it this way,” only the frontline people know.
If you don't pull these people in to interview, test, and give feedback together, the vendor is guessing your processes out of thin air. Guess wrong, and on day one of go-live payroll is miscalculated, inventory doesn't reconcile, and the whole company is yelling at you.
So the right implementation rhythm looks like this:
- Interview: the vendor sits down and goes department by department to understand the current state and the pain points.
- Test: the configured processes go to the staff who will actually use them, run on real data.
- Feedback: staff say out loud “this part isn't smooth, that part is missing a step.”
- Optimize: the vendor goes back, fixes it, tests again, until it runs smoothly.
It's a loop, run for several rounds. It's a co-creation project, not a product delivery. A vendor with no intention of running this loop with you — what do you need them for?
One more thing: if your company is still small and you're unsure whether to implement, or how far to go, start with this: ERP implementation guide for SMEs. Get clear on your needs before you look for a vendor, and you won't get led around.
5. The biggest trap in custom-built systems: moving a bad process over as is
This section is for owners who say “our processes are special, we have to go custom-built.”
Custom-built work isn't the problem. The problem is how many vendors do it: they take your existing overcomplicated, step-heavy, bad process and move it into the new system without changing a single step.
The result? Before, filling in Excel meant jumping through seven fields, three stamps, and entering data twice. Now it's seven fields, three approval clicks, and entering data twice in the new system. Not one minute of work saved, and you're now maintaining one more system and paying one more license fee.
That's digitizing a bad process. It's the most expensive waste in implementation.
Real custom-built work uses the rare opportunity of an implementation to fix the process, not just move it. Why does that step need data entered twice? Can it auto-fill? Why three stamps? Are two of them pure ritual? A good vendor helps you cut the process before customizing, down to what's truly necessary, then digitizes it.
The difference between a vendor who asks “can this step go?” and one who only asks “how many fields do you want?” is your team's working hours for the next three years.
6. Quick comparison: vendors who sell vs. vendors who deliver
| Aspect | Vendors who only sell | Vendors who actually deliver |
|---|---|---|
| Opening promise | “Your team doesn't need to be involved at all” | “Every department needs to send someone to work through it with us” |
| Who shows up | Salespeople, start to finish | Consultant/PM appears early, asks dirty questions |
| Pitch focus | Feature lists, module counts | Takes your process apart, asks about your pain points |
| Custom-built approach | Moves the old process into the system as is | Cuts the process first, then digitizes |
| Delivery | System + manual, dump and run | Interview → test → feedback → optimize loop |
| When go-live isn't smooth | Pay more / not in the contract / no such feature | Adjust it together until it runs smoothly |
| At its core | Sells licenses, ignores results | Works with you until you get results |
Print this table, put it on the desk at your first meeting, and check them against it. You'll pretty much know which kind is sitting across from you.
7. Before you sign, ask these two questions
These two questions are more useful than a hundred-page contract. Ask them face to face and watch how they answer.
First: “Testing, feedback, and optimization during implementation — how exactly do you plan to do that with us?”
A vendor who delivers will immediately give you a concrete working rhythm: who does the interviews, how often you test, who they need from your side, roughly how many rounds. If they can't answer and just say “trust us, leave it to us,” that's the fake promise from earlier.
Second: “If one day I stop using you, can I take all my data with me?”
This one exposes everything. A vendor with something to hide, who wants you locked in, goes pale the moment they hear it. An honest vendor tells you straight: the data format, how to export it, whether it can connect to other systems. Your data is your asset, not the vendor's hostage.
8. How Brightstream does it: we don't force-replace your ERP, we add automation through API integration
After all these other people's pitfalls, a quick word on how we work. Treat it as a comparison group.
Many in this industry walk in and want you to “replace everything, re-implement from scratch,” because the bigger the replacement, the more they make. But forcing out an existing ERP touches too much — data has to migrate, staff have to relearn, risk shoots up. Plenty of companies die right at this step.
Our route is different: if your existing ERP still does the job, we don't force a replacement. We use APIs to connect your systems, then add the automated workflow you're missing, so the steps that used to mean manually moving data, reconciling numbers, and entering things twice across systems run on their own. What's left for your staff is just checking and confirming.
Lower risk, faster go-live, less staff resistance. If you want to know what AI and automation can actually do on top of an ERP, read this: What AI ERP can actually do.
Frequently asked questions
Q: The vendor says “we've done hundreds of implementations, don't worry.” Is that a guarantee? Hundreds of implementations is their experience, not a guarantee of your success. What matters isn't how many they've done, it's how they plan to do yours. No amount of experience saves you if what shows up is the dump-and-run routine. Asking about their working rhythm is more useful than asking for case counts.
Q: My company is small. Do we really need to be this careful, with cross-department co-creation and all? The smaller you are, the less you can afford to lose. One failed implementation can set you back half a year. The upside of a small company is fewer departments and faster communication, so co-creation is actually easier. Instead of chasing “fast, cheap, no hassle for you,” aim for “done right the first time.” Start with the ERP implementation guide for SMEs to get your bearings.
Q: Packaged software is much cheaper. Should I stay away from it and go custom-built? No. Packaged software isn't inherently bad, and expensive custom-built work guarantees nothing. The key is always whether the vendor is willing to get in and build your workflow into it. Some packaged systems, paired with a team that does the work, get implemented beautifully. Some pricey custom-built projects still move a bad process over as is. For custom-built or not, cloud or on-premise, see Cloud vs. on-premise ERP: how to choose.
Q: My only problem is messy inventory. Do I need a full ERP? Not necessarily. Many owners lump inventory management and a full ERP together and end up using a sledgehammer to crack a nut. Whether you need an inventory system or a full ERP, I wrote a whole piece on it: Inventory system vs. ERP: what's the difference. Sort that out before you decide. It can save you a lot of money.
Choosing a vendor comes down to choosing a partner willing to get things right with you, not a supplier who drops a system on you and disappears. If you're evaluating an implementation, or some vendor has already left you furious, we offer a free 30-minute process diagnostic — no system pitch, no contract talk. We sit down, listen to where you're stuck, and see whether there's an easier path. Whether you hire us afterward doesn't matter. At least you'll know which questions to ask. If that helps, pick a time for a 30-minute talk with Johnny Yang and book online.




