Pilot Loop
Limited lab group, focused test set, clear sample protocol, initial statistical reporting, and a decision on whether to scale.
Multi-lab asphalt QA program
Not a blame system. A precision system.
Multiple labs test the same asphalt binder every week — and KAT diagnoses why their numbers disagree.
The Quality Loop
When two labs grade the same binder differently, someone gets blamed and nobody gets better.
The Quality Loop replaces that. Participating labs test equivalent material on a repeating schedule, KAT analyzes every result, and the program tells you not just which lab is off but why — instrument, operator, method, sample prep, or the material itself. The goal isn't to rank labs. It's to make capable labs measurably more precise, so the decisions riding on their data hold up.
How it works
From repeated testing to targeted action.
A one-time round-robin gives you a snapshot. A weekly rhythm gives you a trend — and trends are what separate a real bias from a bad Tuesday. Every loop runs the same six steps under routine operating conditions, never a special "test-day" setup.
- 01Program setup — Agree the labs, tests, sample handling, data fields, schedule, confidentiality rules, and decision thresholds up front.
- 02Equivalent-sample testing — Labs test the same or equivalent binder on the agreed weekly schedule, under normal conditions.
- 03Statistical analysis — Results are read through RMSE, bias, variance, trend, repeatability, and reproducibility.
- 04Evidence review — The analysis separates likely instrument, operator, method, sample-prep, process, and material signals.
- 05Improvement options — Recommendations target the highest-value action: training, SOP review, troubleshooting, control charts, or deeper root-cause work.
- 06Follow-through — Progress is tracked over time, so you can see whether the corrective action actually moved precision.
Weekly cadence → trends, not snapshots.
One round-robin tells you labs differ. A trend tells you why.
The raw material of proof — the same material every lab in the loop is measuring, so a disagreement points to the measurement, not the sample.
Diagnostic capability
Analysis that points to the source.
Knowing a lab is different is easy. Knowing why — and what to do about it — is the whole value. The Loop is built to attribute variation, not just flag it, so corrective effort lands where it changes the result.
Instrument effectsWhether one instrument reads differently from the lab average, the network average, or its own history.
Operator effectsOperator-to-operator spread, separated from true material behavior and the method's own precision limits.
Method & SOP driftWhether conditioning, sample prep, loading, trimming, calibration, or reporting practice is driving the gap.
Production & releaseCleaner data feeding production control, release confidence, customer explanations, and fewer false alarms.
Operating philosophy
Built to make good labs stronger.
The Loop is designed for capable laboratories, not to catch out weak ones. Results are always read against method precision, practical impact, trend, and the decisions the data is meant to support — a lab that disagrees within the method's real reproducibility isn't "wrong," and the program says so.
Confidential by designLab identities, raw data, and corrective actions are governed by the agreed program rules.
Action-orientedReporting ends in the next useful technical step, not a leaderboard.
Technically groundedInterpretation draws on asphalt test precision, rheology, and how labs actually behave.
Improvement-focusedTraining, site visits, and troubleshooting are on the table when the analysis says they'll pay off.
“The goal isn't a ranking. It's a lab that trusts its own numbers — and can defend them.”
Who it's for
One network, or a mixed room with a shared stake.
The Loop can be scoped for a single company's lab network, independent commercial labs, agency labs, producer labs, or a mixed stakeholder group — anywhere the shared goal is higher confidence in asphalt binder data. Run it if:
- 01Lab results disagree enough to affect production or release decisions.
- 02Different operators or instruments seem to produce different answers.
- 03Borderline spec results are creating rework, delay, or customer concern.
- 04Management needs a clean view of lab quality without losing the technical truth.
- 05Existing proficiency testing isn't translating into real improvement.
Engagement options
Start small. Scale when the value is proven.
The first version can be a focused pilot — a few labs, a narrow test set, a clear sample protocol. As it earns its place, the same structure expands to more labs, more tests, richer reporting, and deeper improvement support.
Managed QA Loop
Ongoing repeated testing, scheduled analysis, periodic summary reports, and practical corrective-action guidance.
Evidence Deep Dive
Focused investigation of one lab, operator, instrument, method, or recurring variation pattern.
Training & SOP Review
Targeted training, method review, SOP comments, and practical decision rules for lab teams.
On-Site Troubleshooting
A site visit for when watching the actual workflow, instrument, or sample handling can change the outcome.
Network Leadership Support
Governance, communication, and technical stewardship for producers, agencies, and lab networks.
Where the method comes from
A diagnosis a lab can act on — not just a statistic that flags it.
Built from Pavel Kriz's work on asphalt laboratory QA programs, refinery-facing process improvement, Lean Six Sigma methods, and hands-on binder-test interpretation — the difference between a statistic that flags a lab and a diagnosis a lab can act on.
Start here
Better data is better production decisions.
A first conversation defines the labs, the tests, the sample logistics, the confidentiality rules, and the decisions the program needs to improve.
Every engagement is confidential.