If it has a spec, we measure whether it tells the truth.

The Software Defined Corporation

Every vendor makes a claim. Every product ships with a spec sheet. Every SLA is a promise. We treat all of them like a witness statement — and check the receipts.

Our Mechanical Firmware Index (MFid) is a single number, between 0 and 1, that tells you how much of a claim is real. An MFid of 1.0 means the thing does exactly what was promised. Anything less is the gap between the pitch and the truth. We track that gap across your vendors, hardware, software, and processes — and we use it to decide what stays, what goes, and what gets renegotiated.

What is MFid? How We Use It

A child can understand this.

If someone tells you a cookie has 10 chocolate chips, and you count 7 — that's an MFid of 0.7. The cookie is 70% honest. Do that enough times with the same brand, and you stop trusting the box.

Now scale that to your business. Every vendor, every product, every contract is making claims. Most are not lies — they are slow erosions. A little less bandwidth than promised. A little more latency than the datasheet. A response time that quietly drifts. Add them up across a year and you are running on a fiction.

MFid is the math that catches it. Three tiers of evidence:

Scientific Calculation ✓ What physics and architecture say is possible. The ceiling no claim can exceed.
Published Specification ✓ What the vendor put in writing. The contract, the datasheet, the SLA.
Measured Reality ✓ What we observe in your environment, under your load, on your worst day.

The score writes itself.

What we actually do

We are not your IT department. We are the auditor your IT department wishes it had. Our work falls into four motions, and MFid is the spine of all of them:

🔎 Investigate

We score every vendor, contract, and component already in your stack against what it claimed to be.

⚖️ Police

When the gap is wide enough to matter, we put it in writing. Vendors get the data. So do their account managers.

🛠️ Improve

We tune, replace, or renegotiate until the system gets as close to its own spec as physics allows.

📈 Track

MFid is a heartbeat, not a snapshot. We watch the score over time so drift never becomes disaster.

The KPI Your Board Should Be Asking About

MFid = (D × E × O × I)^1/4

Determinism. Efficiency. Observability. Intentionality. One number. No fluff.

What the math usually finds

< 1.0

Almost Everything

In a market built on marketing, perfect fidelity is the exception, not the rule.

10–40%

The Honesty Gap

The typical distance between what a vendor sold you and what they actually delivered.

Dimensions Scored

Behavior, resource use, visibility, and purpose — measured separately, then combined.

Number That Matters

A board can argue about anything except a falling MFid.

If it ships with a spec, it can be scored.

Marketing teams want you to believe their category is special. It is not. Anything with a published number can be tested against a measured one:

Hardware

Servers, Storage, Network Gear

A drive rated for 500K IOPS that delivers 340K is not "close enough." It is a 32% lie at scale.

Vendors

ISPs, Cloud Providers, SaaS

A 99.99% SLA that lands at 99.7% is ten times more downtime than you bought. The vendor knows. Now you do too.

Processes

Workflows, Support, Onboarding

A 4-hour response target with an 11-hour average is not a target. It is a wish written on a slide.

Software

Applications, APIs, Databases

Software, like any machine, must be reliable, repeatable, and truthful. Anything else is a defect with a roadmap.

Compare what they wrote down to what we observe. Score it from 0 to 1. Repeat next quarter. That is the entire trick.

Why this firm exists.

The Porsche Standard

Porsche routinely under-reports horsepower. The 911 ships with more power than the brochure admits. That is the rarest kind of honesty in engineering: a number that errs against the seller. Almost no one else does this. We score everyone else against the standard Porsche set.

MFid is a KPI, not a feature

MFid belongs on the same dashboard as revenue and churn. It tells you whether the systems and vendors your business runs on are honest. A falling MFid is a leading indicator. By the time it shows up in revenue, you are already late.

Vendor selection, by math

We recommend vendors with the highest measured MFid — not the loudest deck, not the best dinner. As your environment grows, we re-score them. When a vendor's MFid drops, our recommendation drops with it. No loyalty to logos.

We score ourselves first

Our own response times, our own resolution rates, our own promises — measured by the same framework, published to the same client. If our MFid slips, you see it before we explain it. That is the only honest way to sell honesty.

Intentionality is the line

The deepest fidelity gap is not slow software. It is software doing what it was not asked to do. In an era of autonomous systems, an unscored Intentionality dimension is how careers and companies end. We measure it. Most don't.

Stop buying brochures. Start buying behavior.

If your next vendor decision, infrastructure refresh, or platform bet is going to ride on a spec sheet, get a second opinion from someone whose only job is to score the spec sheet against reality. Existing clients include Montecito Bank & Trust; additional references on request.

How We Work Read Investigations Request a Score

What an MFid investigation looks like

mfid-analyzer

$ mfid investigate --client=acme --scope=stack

🔍 Pulling vendor contracts, datasheets, and SLAs...
📊 Comparing claims to measured behavior...
⚖️  Scoring each component on the [0,1] scale...

MFID INVESTIGATION REPORT — Q1
=========================================================

Stack-wide MFid: 0.81  (Honest, with leaks)

Hardware:
├─ Server CPUs:        0.91  ✓ within 9% of rated throughput
├─ NVMe IOPS:          0.72  ⚠ 360K of 500K rated — 28% gap
├─ Switch backplane:   0.88  ✓ 88% sustained
└─ Memory bandwidth:   0.94  ✓ in spec

Vendors:
├─ Cloud provider:     0.96  ✓ 99.96% vs 99.95% SLA
├─ ISP bandwidth:      0.71  ⚠ 710 Mbps of contracted 1 Gbps
└─ SaaS uptime:        0.89  ✓ 4 of 5 SLAs met

Processes:
├─ Helpdesk response:  0.68  ⚠ 6.2h avg vs 4h target
└─ Patch compliance:   0.85  ✓ within window

🎯 Findings filed against:
• ISP — 29% bandwidth shortfall, contract clause 4.2 triggered
• Storage vendor — firmware regression suspected
• Internal process — helpdesk staffing gap, peak hours

No upsell. No "courtesy" findings. Just the score.