SandurTech
Framework · Decision Intelligence Layer

One clear number for how the network is really doing

Ten explainable indices that turn scattered plant data into a shared language every plant — and every level, from the shift floor to the leadership table — can read the same way.

10 indicesOne consistent way to read delivery, quality, cost, and capacity
Every plant, one languageThe same number means the same thing everywhere
Traceable, not opaqueEvery score can be followed back to the record behind it
ONRINRICAIMIIQRITEEIYMEISCEIWCVIDFAIDIL
Why plants need this

Everyone already has a number. Rarely the same one.

Ask three people at a multi-plant steel operation how the network is doing on delivery, or quality, or cost, and you'll often get three different answers — pulled from three different spreadsheets, calculated three different ways. None of them is wrong. That's the problem.

1

Signals are scattered

Delivery data sits in one system, quality in another, energy and yield somewhere else again. Nobody sees the whole picture at once.

2

The same word means different things

"Reliable" at one plant isn't calculated the same way as "reliable" at another — so comparing plants becomes a debate about methodology, not performance.

3

By the time it's visible, it's already cost you

Without a shared, current view, a slipping number often surfaces in a monthly review — long after the decision that could have changed it.

The idea

One shared language, built from what you already measure

The Decision Intelligence Layer doesn't ask a plant to collect anything new. It takes the records already being kept — orders, heats, downtime, quality checks, vendor deliveries, inventory, forecasts — and turns each into one clear, comparable number, calculated the same way, everywhere, every time.

Each of the ten indices answers one plain question about the network — is it delivering in full and on time, is it running at the capacity it should, is it making what quality demands, is it using energy and material well, are vendors and inventory behaving, is the forecast holding up. Together they cover the questions a plant head, a planner, and a shift in-charge each ask in their own words.

No index is a black box. Every score is built from ordinary arithmetic on real records, so it can always be explained — never just asserted.

Ten lenses, one picture

What each index is actually asking

Grouped the way a plant would naturally think about them — not by how they're calculated, but by what they're for.

Delivery & reliability
ONRI

On-Time-In-Full Reliability

How reliably the network delivers complete orders, on time — weighted so it reflects the customers who matter most.

NRI

Network Reliability

How much of the network's capacity stays available and predictable, once stoppages and disruptions are accounted for.

CAI

Capacity Availability

How much usable production capacity the network can actually call on, after subtracting time genuinely lost.

Manufacturing & quality
MII

Manufacturing Integrity

How much of what was planned to be made is actually made, at usable quality.

QRI

Quality Reliability

How much output passes quality the first time, without a complaint or a claim following it out the gate.

TEEI

Thermal & Energy Efficiency

How efficiently heat and energy are used, measured against what each process should reasonably need.

YMEI

Yield & Material Efficiency

How much usable product comes out of the material that went in.

Supply chain & capital
SCEI

Supply Chain Efficiency

How reliably vendors and materials arrive on time and at the right cost, with sound alternatives when they don't.

WCVI

Working Capital Visibility

Whether inventory across the network is sitting where — and in the quantity — it should be.

Planning
DFAI

Demand Forecast Accuracy

How closely what was planned to sell matches what actually sold — plant by plant, month by month.

How it's used

See it, don't just report it

The point isn't a monthly slide with ten scores on it. It's a way of asking a question and getting an answer you can act on the same day.

01

Start with one clear number

The whole network's score for that index, right now — not last month's, not a modelled estimate.

02

Narrow to what matters

By plant, region, product, vendor, customer, or time — however the question in front of you is actually framed.

03

Follow the outlier

A ranked view shows immediately where the number is weakest, without anyone having to go looking for it.

04

See exactly why

Drill down to the order, the heat, or the event behind it, and the plain reason it counted the way it did.

Getting started

Built to grow with the plant, not ahead of it

Starts with data you already have

Built from the records your plants already keep — orders, heats, downtime, vendor and inventory logs. No new sensors or systems needed to begin.

Grows one index at a time

Start with the two or three numbers that matter most today, and add the rest as trust in the framework builds.

Ready for what comes next

Once a number can be trusted and explained, it becomes something a decision can safely be built on — the foundation for whatever comes after sensing.

Want to see your own network inside this framework?

We'll walk through the indices most relevant to your plants and show what the shared number would actually look like, using the kind of data you already collect.