Plain English - Software Defined Corporation

MFid, Explained Without the Math

What we actually do, in plain English. No formulas on this page — not one.

The whole idea in one question

Did you get what you paid for? That is the entire idea. Everything else on this site is the careful way of answering it.

When your company buys something — internet service, software, a support contract, hardware — the seller made promises in writing. MFid is a score from 0 to 1 that says how well those written promises match what you actually got. 1.00 means every promise was kept; the further below 1.00, the bigger the gap.

Think of a health-inspection grade for the things your company buys. The restaurant does not grade its own kitchen. Here, the “rules” are the seller’s own promises.

How the score is built

Three steps, none of which need math:

1. Collect the promises. We go through the contract, the spec sheets, and the sales material and write down every measurable promise. Not what the salesperson implied — what is actually in writing.
2. Measure what really happened. Not what the vendor’s dashboard says and not what anyone remembers — what independent measurement shows, over a real period of time.
3. Compare, promise by promise. Each promise gets its own verdict. Then the verdicts are combined into one score.

One rule governs the combining: the weakest promise dominates. A vendor cannot hide one badly broken promise behind ten kept ones. Like a chain, the score is only as strong as the weakest link — because in real life, the promise that breaks is the one that hurts you.

For the car people: it is a dyno

Bringing your systems to us is like taking your car to the dyno. The window sticker is a claim; the rollers are the truth. We strap your stack down under your load, on your worst day, and publish the sheet — what it actually made against what it was rated for.

A dyno does not care what the marketing department wrote, how fast the car is for its class, or how much the owner likes it. It measures what reaches the wheels and prints the graph next to the number on the sticker. That is MFid, with one honest difference: a dyno pull is a few seconds at wide-open throttle. We do not redline your vendors. We measure them the way they are actually driven — over a real stretch of time that includes the day it matters most.

Sticker horsepower: The vendor’s published spec, SLA, or contract. The claim under test.
Wheel horsepower: What we measure in your environment, independently. The verdict.
The dash readout: The vendor’s own dashboard. We do not score from it, for the same reason you do not dyno a car by reading its speedometer.
The dyno sheet: The report: every promise plotted against what was delivered, not just the peak number.
The dyno queen: Tuned to post a number on the rollers, slow on the street. A vendor that meets spec on a calm Tuesday and falls over at close of quarter. This is why the measurement window must contain your worst day — a score from a quiet week is the vendor’s score, not yours.
The Porsche that over-delivers: Reads higher than the sticker. Scores 1.00 — not more. The sheet notes the generosity; the score measures the match.

Why this is more trustworthy than a dyno number

Anyone who has argued about dyno numbers knows the problem. A Dynojet reads higher than a Mustang; SAE and STD correction give different answers; a number with no correction factor on the sheet is a brag, not a measurement. The credible sheet prints the dyno type, the correction, and the conditions, so the number can be compared to another one.

We hold our numbers to the same rule. Every published MFid carries its evidence tier (measured, estimated, or published spec), the percentage of the system it actually covers, the measurement window, and the methodology version it was scored under. A score without those is not a score — it is a draft, and the Methodology page says so. The correction factor is printed on every sheet.

The four questions we ask of every promise

Dependability

Does it work every time, or just on average?

A courier that averages “on time” but is an hour late every Friday is not dependable, even though the average looks fine.

Efficiency

Does it deliver what the sheet says, at the price you were quoted?

Not whether it is cheap — whether the stated speed, capacity, and price are the ones you actually get once overages and add-ons are counted.

Observability

Can anyone actually check?

A promise nobody can measure is worth nothing, and we score it that way. Trust that cannot be checked is not trust.

Integrity

Does it do what it says on the label?

No undocumented surprises, no quiet changes, no features that work differently than described.

All four are scored from the same evidence, and the weakest-link rule applies across them too: excellent pricing does not buy forgiveness for undependable service.

What you get

One score, and behind it the list of every promise with its own pass-or-gap verdict. A 0.91 means the promises held about 91 percent of the way, and the report says exactly where they did not. You can take that list into a renewal negotiation, an incident review, or a purchasing decision.

What the number means when you see one

Close to 1.00 — the promises are being kept. What you were sold is what you are getting.
Noticeably below 1.00 — there is a measurable gap somewhere, and the report names which promise and how big. The score never arrives alone.
Never above 1.00 — a vendor that over-delivers (Porsche famously ships more horsepower than the brochure claims — see the dyno section above) still scores at most 1.00. The score measures whether promises match reality, not whether the vendor was generous. Generosity gets a note in the report, not extra credit.

One score, one meaning, no grade inflation: a 0.91 means the same thing for a phone system as for a cloud contract.

What an audit looks like from your side

You bring the paperwork. The contract, the spec sheet, the proposal — whatever the seller put in writing. That is the whole entry requirement.
We ask ten questions. Published openly on the intake page: which promises matter most to you, what a failure costs you, what you can already measure. You rank the promises; you know your business, we do not.
We measure. Over a real window of time, against the promises as written.
You get the score and the list.

We hold ourselves to the same standard: our own promises about this website are measured and published live, scored with the same instrument we point at everyone else.

If you want the math after all

Theory — the full framework and its limits (starts with a five-bullet summary).
Methodology — the exact scoring rules, formulas, and evidence standards.
Walkthrough — a complete worked example scoring a real public service.

You do not need any of it to commission an audit or to read the result. The math exists so the score cannot be argued with — not so it cannot be understood.

Bring us a promise. We will tell you if it was kept.

A contract renewal, a vendor you suspect, a purchase you are about to make — the paperwork is all we need to start.

Request an Investigation