What goes wrong
Why ERP projects fail
I sell implementations and this page still exists, because a project that goes wrong costs you money and costs me my reputation. These are the ways they fall over, in the order I have run into them, and what can be done before signing so they do not happen.
Six ways to fall over
Nobody decided how the work would run
The system gets configured while the process is still being argued, and every argument redoes the configuration. Avoided by writing and signing the processes before touching a screen.
The scope was never closed
Every meeting adds something "small". Added up, those small things are a second project. Avoided with a written scope and one rule: what is not in it enters by a decision, not by a conversation.
The people who know have no hours
The project ends up defined by those who do not do the work, and the system ends up configured for work nobody does. Avoided by reserving those hours, in writing, before starting.
The data was dirty
Duplicate customers, products with three names, balances that do not add up. Migrating that multiplies the cost and contaminates the new system. Cleaning is the company's job and it starts before, not during.
Custom code was written where none was needed
Every custom development has to be written, tested and maintained every year a new version ships. Adopting the standard is cheaper today and much cheaper in three years.
Training happened once, on go-live day
People go back to their spreadsheet in the second week and the system becomes an expensive file. Training is support over weeks, not one session.
Signs you are heading there
If you recognise two or more of these, it is worth stopping and sorting things out before spending more.
The vendor never asked to see how you work
Anyone quoting an implementation without having seen your processes is quoting something else: their own.
Nobody told you no to anything
A vendor who accepts everything is not listening, they are selling. Good projects have several things ruled out in writing.
The go-live date was fixed before the scope
When the date rules, what gets cut is testing and training, which is exactly what holds the go-live up.
There is nobody on your side who can decide
Without someone in the company with the authority to settle arguments, every question turns into a week.
What to do before signing
Three things, and none of them costs what getting it wrong costs.
Know which stage you are at
Fifteen minutes of questions tell you whether the company can absorb the project today. It is free.
Write and sign the processes
The scope stops being a conversation and becomes a document both sides sign.
Ask the uncomfortable questions in writing
What is out of scope, who cleans the data, and how many training hours are included. With those three answers the proposals become comparable.