Choosing an ERP system and implementation partner
ERP selection is the process of establishing which ERP system and which implementation partner fit your processes, your risks and your ambitions. It ends in a decision you can account for, not a preference.
What to test a system on
What to test a partner on
When you do not need a process
Two decisions, not one
An ERP selection holds two decisions that are usually treated as one: which software you will run, and who will help you put it in. They fail in different ways, which is exactly why they deserve separate judgement.
Why those two decisions carry so much weight: Gartner predicts that more than 70 percent of recently deployed ERP initiatives will fail to fully meet their original business goals. That is rarely technology breaking, and more often goals that were never made specific enough.
The system
Sets what is possible. When it does not fit, you notice it in the number of exceptions you have to have built.
Read about choosing a systemThe partner
Sets what you actually get. When they do not fit, you notice it in how long problems stay open.
Read about choosing a partnerJudge both as one, and you find out which of the two was the problem only after go-live.
An ERP selection is therefore not a comparison of feature lists. Nearly every mainstream system ticks nearly every box. The difference lies in what works as standard within your way of working, and what only works once it has been adapted.
What are the steps in an ERP selection?
The order runs from business question to decision. The weight sits deliberately before the demonstrations: skip steps one and two, and you spend the rest of the process justifying a choice that was already made.
Establish what the business actually needs
Start with the processes that earn the money, not with a feature list. Record which steps pinch today, which exceptions genuinely occur, and which of those disappear once the process changes.
Without that distinction you are buying your current mess a second time.
Separate requirements from wishes
A requirement is something the business cannot run without. Everything else is a wish. Skip that distinction and you end up with a list on which every system scores about the same.
The choice then falls back on instinct, or on whoever demonstrated best.
Build the long list on evidence
Select candidates on demonstrable fit for your sector and your size. Include systems you have never heard of.
A list of familiar names only confirms what people already believed.
Run demonstrations on your own scenario
Do not let vendors give their standard demonstration. Hand them your own orders, your own exceptions and your own month-end close.
Only then do you see where the standard stops and configuration begins.
Judge the partner separately
The system sets what is possible, the partner sets what you actually get. Test consultant availability, sector experience, and how they handle things going wrong.
Ask about a project that ran into trouble, not only about the successes.
Record the decision with its reasoning
Write down why the choice fell as it did, which risks were known, and which assumptions still need testing.
That document is what explains, later, why certain concessions were made.
What do you test a system and a partner on?
Each topic carries two different questions. If you can answer only one of them, you are not yet in a position to decide on that topic.
Two questions per topic
| Topic | About the system | About the partner |
|---|---|---|
| Process fit | Does the system cover your core process as standard, without custom work? | Does the partner know this process from their own projects in your sector? |
| Exceptions | How does the system handle your unusual orders and customers? | Does the partner advise dropping exceptions, or building them? |
| Data | Which data are you bringing across, and what structure does the system demand? | Who owns data quality, you or the partner? |
| Integrations | Are the connections to your existing systems available as standard? | Who maintains those connections after go-live, and under what agreement? |
| Life after go-live | What can you configure yourself, and where do you stay dependent? | Do the same people stay involved, or does the team change after handover? |
| Cost balance | Which items sit inside the licence, and which arrive on top of it? | How does the hourly rate relate to the guidance you will actually need? |
When do you need a selection process?
Not every ERP question calls for a selection. Establishing that up front saves a great deal.
A selection process fits
- Your processes have changed and the system no longer follows them
- You can no longer see what custom work costs you each year
- The vendor is ending support, or forcing a migration
- You are growing into a scale or a country the system was never built for
Something else fits better
The system fits, the configuration does not
Complaints are about screens, permissions and reports, not about what the system can do. A configuration project solves that for a fraction of the effort.
The bottleneck sits in the process
If the same exception returns every week because nobody is allowed to refuse it, new software moves that problem to a new screen.
A choice has been made that nobody has tested
A second opinion on that choice is worth more than a full process. Test the assumptions, not the candidates.
Going deeper per topic
Both decisions have a page of their own, because there is more to say about each than fits here.
Frequently asked questions
- What does ERP selection involve?
ERP selection is the process of establishing which ERP system and which implementation partner fit your processes, your risks and your ambitions.
It ends in a reasoned decision rather than a preference. The reasoning is the real deliverable: it is what explains, later, why certain concessions were made.
- What is an ERP system?
An ERP system is standard software that supports an organisation’s core processes as one connected whole: purchasing, stock, production, sales and finance.
Its strength is that connection: an order changes stock, planning and the figures at once. That same connection is why a wrong choice keeps costing you for years.
- How do you choose the right ERP implementation partner?
Judge the partner separately from the system, on sector experience, consultant availability, and how they handled setbacks on earlier projects.
Ask about a project that ran into trouble and what they changed as a result. A partner who can only name successes has either done little or learned little.
- How long does an ERP selection take?
The timeline is set by the availability of your own people, not by the number of candidates on the list.
The weight sits before the demonstrations. Get your requirements sharp and the rest moves quickly. Skip that and you stall in comparisons that decide nothing.
- What does an ERP project cost?
That depends on your size, your number of users and how much falls outside the standard. The licence is rarely the largest item; implementation and the custom work around it weigh more together.
So do not ask for a figure but for the breakdown: what the price includes, what is charged per user, and what counts as extra work. A quote without that breakdown cannot be compared with another one.
- Can you run an ERP selection yourself?
Yes, provided you free up someone who knows the processes, has the time, and has no stake in the outcome. That last condition is the hard one.
The question is not whether you can, but who organises the challenge. External guidance earns its place the moment those three conditions stop coinciding.
Further reading on ERP selection
This page describes the process. The articles below each take one part of it and go deeper.
Have your selection tested
ERP Company guides selection processes from the client side and reviews choices that have already been made. We are platform-independent and hold no partnership with software vendors, so we have no stake in the outcome.
Where we help
- Testing your requirements before you go to market
- Guidance on choosing a system and a partner
- A second opinion on a choice already made