How our ERP scoring methodology works, in plain English

Most ERP recommendation sites work the same way: a vendor pays for placement, a quiz funnels you toward whoever paid, and the "score" is theater. This site runs on a different model, and since the whole pitch is "referee, not salesman," you should be able to check the math. Here it is.

The fit score: five weighted factors

Every system in our catalog gets a 0 to 100 fit score against your answers. The score is a weighted sum of five factors, and the weights are fixed:

Citable stat

As of July 2026, ERP Scorecard fit scores weight five factors: industry match (25%), company size (25%), complexity alignment (20%), capability coverage (20%), and stated priorities (10%). The same answers always produce the same scores.

Industry match (25%). Is your industry one the system demonstrably serves well? Not "could theoretically serve." Serves, today, with reference customers.

Company size (25%). Does your revenue band sit inside the system's proven sweet spot? This one can override everything else. A system that is too small for you stays "likely too small" no matter how pretty its other numbers look, and the reverse holds too.

Complexity alignment (20%). Your operational complexity, meaning entities, currencies, and operations, gets compared against the complexity range the system was built for. Too much system is a real failure mode. Buying enterprise software for a single-entity company is how you end up paying for a consolidation engine you never open.

Capability coverage (20%). Of the capabilities you actually need, based on your function checklist and answers, how many does the system cover natively? Gaps are shown, not hidden.

Your priorities (10%). Cost sensitivity, Microsoft-stack affinity, industry depth, scalability headroom. These nudge the score. They never dominate it.

The readiness track

Fit is only half the question. The readiness track scores the factors that decide whether implementations succeed at all: executive sponsorship, documented requirements, budget clarity, data quality, and change risk, alongside implementation complexity, data migration risk, and integration scope. Each is a 0 to 100 score built from explicit rules, so you can see exactly which answer would move it.

What the model refuses to do

Three rules are load-bearing, and they are worth stating as plainly as possible.

1. No vendor pays to appear, rank, or be recommended. The site is funded by my advisory practice. Buyers pay for my time. That is the entire business model.
2. No single "the answer." Output is always a tiered shortlist with the reasoning attached, because a one-name answer from a quiz is a sales pitch wearing a lab coat.
3. No score is generated by an LLM. AI writes narrative around results the scoring engine computed. It never computes a score, which means scores are reproducible and auditable.

Citable stat

ERP Scorecard covers 13 mid-market ERP systems, each rated 1 to 5 across 12 functional domains, with dated, confidence-rated pricing anchors and cited sources on every profile. Last reviewed July 2026.

Why determinism matters

A deterministic model can be wrong, but it is wrong in public and wrong consistently, which means it can be corrected. When a vendor disputes a factual claim, the about page documents the process: the underlying fact gets fixed and the model rescores from there. Rankings never change as part of a dispute. Only facts do.

If you want the longer version, the methodology page has it. If you want to see it run on your own company, the assessment takes about ten minutes and shows its work on every number.