An owner replacing a core system treats it as an IT project. A software budget, a vendor to manage, a timeline on a whiteboard. It is not an IT project. It is a knowledge project with an IT deadline attached, and the software is the easy half.
Here is what actually has to move. Pricing logic. Discount authority. Credit limits. Customer terms. What is actually on the shelf. All of it has to become configuration before go-live, and configuration can only be built from what is written down. Whatever nobody writes down either disappears at cutover or gets rebuilt as somebody's best guess, and the guess persists for years because eventually nobody remembers it was a guess.
The industry's own numbers point the same direction. More than a quarter of these projects run over budget, almost a quarter run over schedule, and the most common cause of schedule overrun is organizational issues, not technical ones. Six questions before the build starts.
- What does your new system need to know that only a person currently knows?
- Which of your prices could a new employee find without asking anybody?
- What are you about to migrate that nobody has ever verified?
- If your best operator resigned this week, what would the new system never learn?
- When the new system gives someone an answer that looks wrong, who do they ask?
- Who is teaching the system all of this, and what else is on that person's plate?
If more than two of those answers are a name instead of a document, the rest of this piece is what that costs at cutover.
The software is the easy half
Most modern systems can meet a mid sized company's baseline requirements. The industry data backs that up, since project trouble traces most often to unclear process ownership, weak adoption, and organizational issues rather than to the product. Fewer than a quarter of organizations put serious focus on managing the change at all. The part of the project that fails is the part that was never about software. A system can only be configured to run rules that exist somewhere outside a person, and in most businesses the rules that matter do not.
Your implementer cannot save you
The integrator you hire is good at the visible work. They will map the old customer table to the new one, field by field, competently. What they cannot do is tell you what your own people mean. Somewhere in your data is a code that has been carrying three different meanings for a decade depending on who typed it, and the person who knows which meaning applies has never had a reason to say it out loud. Migration specialists are blunt about this. Mapping is not the hard part. The hard part is the small share of records carrying undocumented business rules, and they only surface under real volume, after go-live. A new system does not fix an undocumented process. It exposes it, on a deadline.
After go-live, the guesses go quiet
During the build, a guess is at least visible. It is a question somebody asked and an answer somebody shrugged. After cutover it hardens into configuration, and configuration does not look like a guess. It looks like the system. Picture a company carrying its inventory records into the new system as the opening balance when those records have never once been verified against a physical count. From day one the system is confidently wrong, and the people who knew which numbers not to trust are now working around software that outranks them. Two years later nobody can explain why the system does what it does, because the reasons were never reasons. The window to fix this closes on the go-live date and does not reopen.
Migrate the knowledge first
The sequence is the whole game. The cheapest weeks of the entire project are the ones spent writing down what only people know, before configuration starts. Which prices, and why. Who can discount, and how far. What a customer's terms actually are, not what the invoices imply. A count somebody has physically verified. This is the argument of The Premature Hire wearing an implementation budget. Documented judgement has to exist before authority gets handed over, and a system holds more authority than any hire ever will. The new system is worth buying. What it cannot do is know your business by itself. Write down what only people know first, then configure. That is the difference between a system that runs your business and one that runs on a guess.