What an ERP implementation really means
Why this is a business redesign wearing a software project's clothes.
An ERP implementation is not a software installation. The software is the easy part — you can have ERPNext running in an afternoon. What takes ten weeks is deciding how your business actually works, writing that down, and agreeing on it with the people who do the work.
That reframing matters because it predicts where projects fail. They rarely fail on technology. They fail because nobody could say, definitively, who is allowed to approve a purchase over ₹2 lakh; or because two departments each believed they owned the customer master; or because "we'll figure out the returns process later" turned into a month of reconciliation after go-live.
The three things you are actually buying
A single source of truth. Today the answer to "how much stock do we have?" depends on who you ask. After implementation there is one answer, and it is either right or visibly wrong — which is an improvement, because a visibly wrong number gets fixed.
Enforced process. An ERP makes some things impossible. You cannot invoice against a cancelled order; you cannot receive 200 units against an order for 100 without someone approving the variance. Every one of those constraints is a decision you are making about how the business runs.
An audit trail. Who did what, when, and what it changed. This is worth little until the first time you need it, and then it is worth the entire project.
What you are not buying
You are not buying a faster version of your current process. If the current process is broken, automating it produces broken outcomes at higher speed — this is the single most common way a technically successful implementation still fails.
You are also not buying the end of spreadsheets. Some spreadsheets are a sign of a gap in the system; others are genuinely the right tool. Deciding which is which is part of discovery, not something to be declared at kickoff.
How to read this guide
Sequentially the first time. The chapters follow the order the work actually happens in, and later decisions depend on earlier ones — master data design depends on the entity structure, which depends on what discovery found. Skipping ahead to configuration is the most expensive shortcut available.