Home - Software Defined Corporation

The auditor your IT department wishes it had.

Vendors make promises in spec sheets, SLAs, and proposals. We measure what was actually delivered and hand you one number.

Request an Investigation Read it in plain English →

Not software-defined networking — a different sense of the phrase.

One number: MFid

MFid (Measured Fidelity) is how much of what was promised you actually got, on a scale from 0 to 1. A 1.00 means every promise held. The weakest promise dominates the score, so one broken commitment cannot hide behind ten kept ones.

Collect the promises

Every measurable claim in the contract, the datasheet, or the deck. What is written down, not what was implied.

Measure what was delivered

Independent measurement over a real window of time. Not the vendor’s own dashboard.

Compare, promise by promise

Each claim gets a verdict. The verdicts roll into one score, and the report lists exactly where the gaps are.

The formal definition

Fidelity is the correspondence between a claim (what something says it is) and its referent (what it actually is). A buyer’s exposure to a wrong decision is the consequence of that decision multiplied by one minus the fidelity of the claim it rested on. The market spends heavily on reducing consequence and almost nothing on measuring fidelity, because there has been no instrument for it. MFid is the instrument.

The firm’s name asks the question the instrument answers: how software defined is your stack? — to what degree does its stated definition determine its behavior? That number is the Software Defined Index (SDI), and by definition SDI and MFid are the same number: one instrument, two names.

The theory → · The naming argument → · The ten intake questions →

Four questions, asked of every promise

A claim can miss reality in four distinct ways. Each dimension is scored from 0 to 1 against its own published rubric.

Dependability

Does it hold every time, not just on average?

“Responds in under 200 ms” — measured repeatedly, scored on spread and shortfall.

Efficiency

Does it deliver the number on the sheet, at the cost that was quoted?

“500K IOPS” observed at 340K scores 0.68 on that claim.

Observability

Can you see the gap yourself?

“Audit logs for every admin action” exposed only as a quarterly PDF scores low.

Integrity

Does the design actually serve the stated purpose?

“Zero Trust” running on a perimeter VPN scores low no matter how fast the VPN is.

MFid = (D × E × O × I)^1/4

Geometric mean of the four dimension scores. The weakest dimension caps the result.

Why a geometric mean?

An arithmetic mean lets strong dimensions pay for weak ones. A geometric mean does not: a single dimension at 0.5 caps the aggregate near 0.84 even if the other three are perfect, and any dimension at zero makes the whole score zero. That is deliberate — a broken dimension is a hole through which the buyer can be misled regardless of the others. Full derivation with worked numbers on the Theory page.

Can a score go above 1.0?

No. MFid scores match, not generosity: over-delivery is clipped at 1.0. Porsche is the canonical example — independent dyno runs aggregated at 992forum.com put the 992 Carrera S near 487 hp against a claimed 443 hp (source, retrieved 2026-05-28; a forum aggregate, not a single instrumented run). A vendor that under-reports as documented policy earns the Verified Underreporter annotation, not a number above the scale.

Full theory → Scoring rubrics → Worked example →

Who this is for

Any company that signs against a claim — cloud, SaaS, bandwidth, hardware, managed services. If it was promised in writing, the gap between promise and delivery is measurable, whatever the industry.

Running an MSP? That is the founder’s background and it shows. MFid tells you whether the trust your clients place in your stack, vendors, and team is earned — before a renewal, before an incident, before a client asks why something failed.

We score ourselves first

This website’s own uptime, latency, and reachability are scored with the same instrument and published live — alongside a list of the promises we make about ourselves that we have not instrumented yet. If our MFid slips, you see it before we explain it.

See our live MFid score →

Go deeper

Plain English — the whole idea, no formulas.
Theory — fidelity, the four dimensions, the math, the limits.
Methodology — evidence tiers, scoring rubrics, version history.
Walkthrough — a real public SLA scored end to end.
Compare — MFid next to SLOs, DORA, and ISO 25010.
Manifesto — the ethic: we don’t assert, we calculate.

Bring us a claim. We will tell you the number.

Vendor renewal on your desk, infrastructure refresh, platform decision? Score it before you sign.

Request an Investigation