Why this matters. An unstructured selection typically extends the project by six to eighteen months and increases the risk of later contractual disputes — a pattern Dreher has documented across more than 1,200 selection projects. Deloitte's selection guidance emphasises functional fit as the primary lever of long-term total cost of ownership.
When this approach fits — and when it does not
This approach fits when:
-
Executive sponsorship treats the selection as a strategic decision with sufficient preparation time;
-
Core processes are documented or scheduled for documentation as part of the selection;
-
The selection team can allocate at least twelve weeks of focused preparation.
This approach does not fit when:
-
The project is a purely technical migration without process changes;
-
A vendor has already been politically committed before evaluation begins;
-
Accountability between IT and the business has not been clarified.
How to do it — the four steps
The structured ERP selection follows four sequential steps. Each step produces a concrete artefact that feeds the next.
-
Step 1 — Build the process landscape: identify core processes, structure MECE, name a process owner for each. Output: twelve to twenty documented core processes with named accountability.
-
Step 2 — Weight the requirements matrix: capture three critical requirements per process, weighted by business impact. Output: a weighted requirements matrix with documented rationale.
-
Step 3 — Define the scoring model: agree evaluation criteria before any vendor demo. Output: scoring model with twenty-seven to forty weighted criteria.
-
Step 4 — Evaluate vendors and negotiate exit clauses: score against the model, negotiate contractual exit terms, document the decision basis in writing.
A common mistake and how to avoid it
A mid-sized manufacturer in Baden-Württemberg started with a feature list from the internet, signed a contract after eight months of evaluation, and discovered during implementation that three central processes were not in the specification. Six months of additional work resulted.
In practice, a checklist works only when it puts organisation before IT: processes first, tools second. Across more than 1,200 selection projects, this sequence demonstrably reduces implementation cost and lowers the risk of mid-project scope correction.
What to do next
If you want to test how defensible your current selection basis is:
-
Name three critical business decisions (e.g. order release, purchase approval, month-end close).
-
For each, write two or three scenarios — at least one of them an exception case.
-
Test those scenarios against your current requirements list. Where it fails, you know what is missing.
We can run this assessment with you and translate the result into a usable scoring model — shortening the vendor evaluation by weeks.
FAQ
A defensible selection works with twenty-seven to forty weighted criteria. More dilutes the decision; fewer fails to capture critical requirements.
Before the vendor conversations. A list written after the demos unconsciously adopts the language of the products seen and biases the evaluation.
At least twelve weeks for mid-sized organisations. Shorter preparations typically result in additional work surfacing during the implementation phase.
Related insights
- ERP Selection consulting — how Dreher runs the four-step process
- ERP Consulting and Specifications — turning the checklist into a contract-ready brief
- Wiki: Process Owner — the role that anchors Step 1
- Insights — perspective articles on ERP selection and digital transformation
- Answer Library — browse all questions on ERP, processes, and selection
