From more than 1,200 Dreher Consulting engagements we know: the model is methodologically uncontested, yet in selection practice it is often reduced to spreadsheet folklore. This entry frames the Nutzwertanalyse from the perspective of managing directors facing an ERP selection, a make-or-buy decision, or a variant assessment.
What Is a Weighted Scoring Model? — Methodological Definition and Distinction
The weighted scoring model is a multi-criteria evaluation method that weights quantitative and qualitative criteria and aggregates them into a comparable total utility score per alternative. It is a discipline of decision preparation, not an arithmetic automaton. The DACH reference is Christof Zangemeister; the engineering frame is VDI 2225 Part 3.
The weighted scoring model is the structured evaluation of alternatives along weighted criteria. It combines quantitative magnitudes (licence cost, maintenance, implementation effort) and qualitative magnitudes (usability, vendor stability, industry fit) onto a single point scale. Weighted single evaluations aggregate into a total utility score for pairwise comparison.
The methodological reference in the German-speaking world comes from Christof Zangemeister: Nutzwertanalyse in der Systemtechnik (1970, fourth edition 1976). Zangemeister defines the method as the “analysis of a set of complex action alternatives with the aim of ordering them according to the preferences of the decision-maker with respect to a multi-dimensional objective system”. This definition remains the methodological reference point today.
Three terms must be separated to handle the method precisely:
- Knock-out criteria (must-have requirements) — criteria a vendor must fulfil to enter the evaluation. GoBD compliance, industry-specific validations, CSRD data provision. Applied as a filter before scoring, not weighted.
- Evaluation criteria (wishes) — criteria differentiated through point allocation and weighting. Functional scope, usability, TCO, vendor service.
- Weighting — the percentage share of a criterion in the result. In the two-stage matrix you first weight the criterion groups, then within each group the single criteria — sum at each level equals 100 per cent.
Distinction from related methods matters in the Mittelstand. Cost-benefit analysis reduces evaluation to monetary magnitudes; SWOT yields no comparable numerical value. Saaty's Analytic Hierarchy Process (AHP) extends the model with pairwise comparisons and a consistency check (CR below 0.1). The technical-economic evaluation per VDI 2225 Part 3 provides the German engineering frame and explicitly requires separating minimum requirements from wishes before point allocation.
Why the Weighted Scoring Model Matters in the DACH Mittelstand in 2026
ERP selection, make-or-buy, and variant assessment demand the structured combination of qualitative and quantitative criteria. From our experience, the Nutzwertanalyse is the 2026 DACH standard for exactly this decision class — provided it is led with discipline, not reduced to a polished spreadsheet exercise.
First, rising selection complexity. Cloud, on-premise and hybrid, ERP standard software, industry-specific ERP and bespoke ERP compete with dozens of vendors each. The Trovarit study “ERP in Practice 2024/25” is the established vendor-neutral user-satisfaction survey in the DACH region, and it shows that vendor service quality is regularly criticised. Consequence: vendor service belongs visibly weighted in the model.
Second, rising regulatory density. CSRD reporting, GoBD requirements, industry-specific validations — non-negotiable and therefore knock-out. Treating them as evaluation criteria lets an inadmissible vendor rise to the top on soft points — a systematic methodological error.
Third, shifting stakeholder logic. The selection team today consists of management, IT leadership, business unit heads, process owners, and external consultants. Consensus emerges not from intuition but from defensible evaluation. From our experience, the model in 2026 is less an arithmetic engine than a discipline tool for a heterogeneous team.
The methodological frame has been available for decades — Zangemeister 1970/76, VDI 2225 Part 3 (1998-11), Saaty's AHP. What is contested is not the method but its disciplined application.
Practice Example: ERP Selection at a Mittelstand Machine Builder (Anonymised)
In an ERP selection project with a Mittelstand machine builder we showed what a cleanly run model delivers — and what a pre-filled spreadsheet does not. Three vendors, up-front knock-out filtering, two-stage matrix, five anchor examples per qualitative criterion, closing sensitivity analysis. Result: a defensible decision in ten person-days instead of four months of debate.
From over 1,200 Dreher Consulting projects the pattern is robust: without a clean separation of knock-out criteria the weighted scoring model fails regularly — regardless of how well the rest of the method is set up. In our experience the knock-out discipline is the single highest-leverage intervention.
A Mittelstand machine builder from southern Germany with around 260 employees called us after an inconclusive ERP pre-selection. Three vendors, three spreadsheets each weighted by a different team member, three winners, no consensus. Management demanded a defensible basis.
We set up an ERP-specific weighted scoring model in two weeks. First the separation: GoBD compliance, industry-specific validation for machine building, full CSRD data provision — three knock-outs as filter before scoring. One vendor dropped immediately because CSRD provision was not committed. Two remained.
In phase two we defined criterion groups with the team. Four groups: core functional coverage (35 %), TCO over five years (25 %), vendor and service (20 %), implementation and migration risk (20 %). Within each, six to eight single criteria with weights.
In phase three we drafted five anchor examples per qualitative criterion (1 = not fulfilled, 3 = fulfilled with adaptation, 5 = fulfilled out of the standard). Point allocation ran in a moderated workshop; results were documented per criterion audit-ready.
In phase four we ran three scenarios — function-dominant, TCO-dominant, vendor-risk-dominant. In two of three the same vendor stayed on top. In one the ranking flipped marginally and triggered a sharpened PoC. Management decided on the basis of the documented model — ten days of disciplined method instead of four months of internal debate.
What Weighted-Scoring Guides Don't Say
Three points rarely stated clearly in weighted-scoring guides or business textbooks but decisive in ERP selection practice. Contrary to most ERP textbook accounts, the bottleneck is rarely the math — it is the organisational discipline. From over 1,200 engagements we know the pattern.
1. The Two-Stage Weighting Matrix Is Not Optional, but Mandatory
Most guides recommend single-stage weighting (all criteria weighted directly, sum 100 per cent). Fine for three to five criteria — for an ERP selection with 30 to 80 criteria it yields two systematic distortions: functional detail dominates by sheer count, and criterion groups blur. The two-stage matrix solves both: upper stage between groups (function, TCO, vendor, risk), lower stage within each group. From Dreher practice: four to six groups above, six to ten criteria per group below.
2. Knock-Out Criteria Belong Before Scoring, Not Inside It
The most important separation in weighted scoring is routinely ignored: must-have requirements and evaluation criteria are weighted in the same table. The result: a vendor failing GoBD can still reach first place on excellent usability — methodologically a scandal, in Mittelstand practice everyday. The VDI 2225 Part 3 demands this separation of minimum requirements and wishes explicitly as standard procedure. We recommend three to seven knock-outs per ERP selection — statutory obligations, industry-specific validation, technical minimums such as multi-client capability. They act as a filter, not as a point supplier.
3. Sensitivity Analysis Is a Methodological Duty, Not a Decorative Add-On
The central weakness of every filled-in model: result robustness against small weight shifts is not tested. Consulting decks treat sensitivity analysis as “optional” — in every serious methodology it is mandatory. From Dreher practice: three scenarios at the end of every model — function-dominant, TCO-dominant, vendor-risk-dominant (each +10 percentage points on the group). If the leader stays stable in two of three, the decision is defensible. If it flips, the team runs a second round with a sharpened PoC.
Our reading
A weighted scoring model without up-front knock-out separation, without two-stage weighting, and without sensitivity analysis is a filled-in spreadsheet, not the method. The three disciplines are not negotiable if the model is to carry an ERP decision.
How We Apply the Weighted Scoring Model to ERP Selection — Four-Phase Approach
We set up the weighted scoring model for ERP selection in four phases — objectives and knock-outs, weighting matrix, evaluation with anchor examples, sensitivity analysis. Typical effort: eight to twelve person-days for a light model with three vendors and 30 to 50 criteria; more for corporate-style variants.
In independent ERP selection consulting we view the model not as a spreadsheet exercise but as part of the selection procedure. The following four-step approach has proven itself:
- Objectives and knock-out criteria. Clarify the strategic objective (growth, internationalisation, compliance, consolidation) and derive three to seven knock-outs, each documented audit-ready. We often pair this with the ERP requirements specification (Lastenheft).
- Two-stage weighting matrix. Four to six criterion groups at the top (function, TCO, vendor, risk, industry fit, architecture) and six to ten single criteria per group below. Weights sum to 100 per cent at each level. Workshop with the selection team, pairwise comparisons on the Saaty 1–9 scale, consistency check.
- Evaluation with anchor examples and calculation run. Five anchor examples per qualitative criterion — 1 = not fulfilled, 3 = fulfilled with adaptation, 5 = fully fulfilled out of the standard. Point allocation per vendor in a moderated workshop, supported by PoC results and the documented requirements base. Total utility score = points × weight, summed across all criteria. We have operationalised this disciplined reading of the Nutzwertanalyse as SCOReX® — a scoring framework for ERP selection in the DACH Mittelstand.
- Sensitivity analysis and decision workshop. Three scenarios — function-dominant, TCO-dominant, vendor-risk-dominant. Test ranking stability. If the leader flips in more than one scenario, a second round with a sharpened proof of concept (PoC). Results are not delivered as a report but discussed as a decision basis.
How we use the method in practice
From over 1,200 engagements we bring ranges — number of criteria per selection class, typical weights per criterion group, person-days per phase. The difference from corporate-style offers: we do not build a pre-fab tool model that adopts the tool's own criteria. We deliver an ERP-specific model tailored to the company's strategy, processes and industry.
Common Mistakes in Weighted Scoring
Four recurring failure patterns in Mittelstand weighted-scoring projects. All avoidable — if addressed before point allocation. Three are methodological, one organisational. In our projects, the organisational mistake is the most persistent of the four.
Mistake 1 — pseudo-objectivity through spreadsheet polish. A table with two decimal places suggests precision the evaluation does not contain. The method lives on discipline, not decimal places. Remedy: a five-point scale with anchor examples, sensitivity analysis as honesty test.
Mistake 2 — knock-out criteria inside the scoring. Statutory obligations such as GoBD or CSRD are point-scored rather than applied as filter up front. Result: an inadmissible vendor rises on soft points. Remedy: three to seven knock-outs explicitly before the scoring.
Mistake 3 — no operationalisation of qualitative criteria. “Usability good” is not an evaluation but an opinion. Remedy: five anchor examples per qualitative criterion, documented audit-ready — VDI 2225 Part 3 demands this explicitly.
Mistake 4 — team politics beats method. Weights are not negotiated in consensus but set by the loudest voice. Remedy: pairwise comparisons on the Saaty scale, consistency check, documented justification. If in doubt: an external moderator.
Our reading
In our projects the mistakes are rarely methodological — they are organisational. From our experience that is where independent consulting has its leverage.
Frequently Asked Questions
Both belong to multi-criteria decision methods (MCDM). In the Zangemeister reading the model weights criteria directly (sum 100 per cent) and scores each vendor on a point scale (1–5 or 1–10). AHP after Saaty adds pairwise comparisons on the Saaty 1–9 scale to derive weights, plus a consistency check (CR below 0.1). In practice we combine the two: weighted-scoring structure with a two-stage matrix, AHP logic for weight derivation within each group.
From 1,200+ Dreher engagements we know: a light model with 30 to 50 criteria reliably carries an ERP selection with three vendors. More than 100 criteria is worth the effort only for complex industries (plant engineering, medical technology, food). Distribution matters: four to six groups above, six to ten criteria per group below — flat lists lose the discipline of two-stage weighting.
No. A light model in the DACH Mittelstand runs on a spreadsheet. From our experience with machine builders, food producers, and traders, AHP specialist software pays off only when pairwise comparisons across more than twenty criteria are run systematically — rarely the case here. Methodological discipline matters more than the tool: clean knock-out separation, two-stage weighting, anchor examples, sensitivity analysis.
The sensitivity analysis tests whether the result is stable against small weight shifts. In ERP selection that is decisive, because the weights themselves are team-negotiation outcomes — not absolute. From 1,200+ engagements we know the pattern: in roughly two of three runs the ranking is stable, in one of three it flips under plausible weight shifts — then the team needs a second round with a sharpened PoC. Without sensitivity analysis this honesty test is missing, and the decision is only apparently precise.
Next Steps
Before a weighted scoring model for ERP selection, a short up-front clarification usually pays off: which strategic question should the model support, which knock-out criteria are non-negotiable, who carries the moderator role? Skipping that produces a table nobody trusts. If you want to deepen these topics, look into our ERP consulting services and our ERP selection area. We work independently and assess requirements along strategy, process and system architecture — not along a vendor preference.
We also recommend our wiki entry on the functional specification (Pflichtenheft). Current articles live in our insights section; an initial conversation runs via the contact page. Methodological depth: Zangemeister (1976), VDI 2225 Part 3, AHP and the Trovarit study 2024/25.
30 minutes with Dreher Consulting
A structured up-front clarification of your ERP weighted scoring model — strategic question, knock-out criteria, and the right weighting logic, vendor-neutral, from over 1,200 projects in the DACH Mittelstand.
