⚠ Preview build — connected to the LIVE database
Anything you save here is a real record on the production board, saved the moment you click Save. There is no separate test database yet — one is in development. Until then, write down every record, action and suggestion you create or change here so it can be undone.
⚠ Not saved — your last change is only in this browser
The database rejected the change and it could not be retried. Do not close this tab. Click below to try again; if it keeps failing, check your connection, then sign out and back in.
Someone else saved a newer version, and saving yours would have overwritten it. Nothing of theirs was changed. Note down anything of yours you still need — it is on screen until you load theirs — then load the latest version and make your change again.
RFT Board

RFT Board

Coryton Advanced Fuels · Right First Time
Access is restricted to approved investigators. Contact an administrator to be added.
Coryton Portal

Coryton Right First Time Board

Misblend investigation, root-cause tracking and countermeasure register
… Role Request more access
▶ Training walkthroughTen short sections, about ten minutes 📄 Quick guideTwo sides of A4 — the daily handful 📘 Full user guideEvery screen and every field 🎥 What the board is forNinety seconds, no sound — for a team meeting
Something in your way? Use 💡 Suggest an improvement.

Assigned to me

Everything on the board waiting on you: investigations you own, reports for you to review, reports that came back to you, and actions you own. Click a row to open it.

Triage inbox · records needing triage

New records from the ERP webhook and records held for lab / formulation response after a triage return-to-source. Click a record to open triage.

Sign-off queue · reports waiting on a person

Prepared reports waiting for a reviewer, and reports a reviewer has returned for rework. The board sends no notifications — this list is how you find out. Click a record to open it.

Repeating causes · IA investigation

Causes that have come up . Each needs an IA investigation; enter its reference in any of the records (IA investigation reference) to show it was investigated. Click a row to open the latest record. This list reflects the records loaded when the page opened.

Open investigations · by SLA

Records routed to an investigation path after triage and still being worked. Once a report is prepared it moves to the sign-off queue above. Target close-out is 14 days from open. Investigations older than 14 days need escalation.

Actions Register

Countermeasures across all misblend investigations. Corrective (fix root cause) and preventive (prevent recurrence). Click a row to open the source investigation.

WO / Ref Type Action Owner Priority Due Status

Misblend Log · Full history

Finished investigations — closed and verified. Set the status filter to All statuses to search every record on the system. Click a row to open the investigation.

0 records · sort: Newest first Total CoQ (filtered): £0
Date WO / Ref Blend Family Cause Status CoQ Actions

KPI Report · Monthly RFT by category

Rolling 12-month view. RFT (Monthly) reflects that month's total and misblends. RFT (YTD) is year-to-date cumulative. RFT (R12) is a 12-month rolling window ending in the row's month. Both denominators are maintained in Admin → Monthly Blends.

Against original blends

Blends only — the historic basis, and the one the dashboard, the charts and the wall board use.

Against all blending operations

Blends plus additisations, adjustments, reblends and intermediate blends. The same misblends set against every operation that could have gone wrong, so it reads higher than the table above — not because fewer things went wrong, but because more work is counted.

Cost of Quality · by month

Cost of Quality is what a misblend actually cost — material that was wasted, external analysis, and the labour to put it right. Material that could be reused, and material added to bring a blend back into specification, are not in it. Potential is what the same misblends would have cost had nothing been reusable — the same analysis and labour, with every excess or dropped-out volume counted as lost. When that volume goes back into stock it is luck, not process; the mistake was the same size either way. Recovered is the difference. Unconfirmed is the part of the cost that rests on imported records where the register never said whether the excess was wasted or reused; it is counted at the register's own figure until somebody confirms it. Hours is the rework time by role, unpriced. Potential is not on the dashboard; it is here for the people who need the size of the mistake rather than what it happened to cost.

Cost of Quality · the records behind a month

Pick a month and every record in it is listed with what its two figures are made of. Material at risk is the dropped-out, overblended or scrapped volume at cost; Material lost is the part of it that was wasted. Cost = lost + analysis + labour; Potential = at risk + analysis + labour. Click a row to open the record.

Misblends
0
 
Potential CoQ
£0
 
Misblends
0
 
RFT %
—
 
Rework hours
0
 
Potential CoQ
£0
 
Misblends
0
 
RFT %
—
 
Rework hours
0
 
Potential CoQ
£0
 

Recurring causes (frequency)

Actionable tertiary causes — excludes Non-Accountable · click a bar to see its records

Direction of travel vs prior period

Positive = worsening (more events). Negative = improving. Signals which controls are landing and where new attention is needed. Click a bar to see its records.

Ongoing investigations

Closed in the last 7 days

Live investigations

What the shift teams should be aware of — open work orders currently under investigation.

Recent outcomes (last 7 days)

Closed investigations — root cause and what to watch out for on future blends.

Admin

Reference data. Changes are saved immediately and applied to all subsequent investigations. Changes do not affect historical records.

Numeric cause codes with their Area / Sub-area / Specific cause and the evidence guide shown in the investigation modal. Historic records keep their existing code even if renamed here.

CodeAreaSub-areaSpecific causeEvidence guideActions
Add new cause code

Triage workflow phases and evidence items. T1 (QA verification) is a checklist an investigator ticks off before a record can proceed. T2 (Investigation path) is direct path selection — edit those in the Investigation Paths tab. Changes here affect every future triage; historic records keep the wording that was in place at the time of their triage.

Investigation path definitions. Each path drives the walk-through steps an investigator sees when a record is routed to it. Admin only — changes here affect every future investigation, so edit carefully. Steps store a snapshot of their definition on each walk entry, so historic reports render the wording that was in place at the time of the walk (not the current wording).

0 paths defined

Options shown in the Containment dropdown on the investigation modal.

Product family classifications used for filtering and cluster analysis.

Everybody with an account on the board, and their role.

Roles are set in the Access app and kept in step every 15 minutes. A change made here would be put back. Open the Access app
Loading users…
Name Email Role
Read only — can view records and reports; no editing. Investigator — can create and edit defects, run triage, walk investigation paths. Admin — investigator plus can edit admin settings and delete or withdraw records. Somebody who needs more than they have uses Request more access beside their role badge.

Standard hourly cost rates used to compute the labour components of Cost of Quality. Editable to reflect current fully-loaded rates.

How fast the Wall Display Mode board moves. Longer is calmer and easier to read from a distance; shorter shows more in a given time. Changes apply the next time the board is opened, or immediately if it is already running on this screen.

For a plant display that cannot sign in: open /wall on this site instead. It is this wall board, with no login, drawn from a snapshot published hourly between 08:00 and 18:00 — the tiles, the charts and the ticker of open and recently closed investigations. What the ticker shows is therefore visible to anyone with the address. Set it up by following The public wall board in docs/DEPLOYMENT.md.

Monthly totals used to calculate RFT %. Populate historical months from the ERP monthly report; update each new month at month-end. Only months with a value appear in the matching KPI table — the two tables below feed the two RFT tables on the KPI tab, and a month left blank in one still counts in the other.

Original blends

Blends only. This is the historic basis, and the one every dashboard tile, chart and wall-board figure uses.

All blending operations

Blends plus additisations, adjustments, reblends and intermediate blends — every operation that could go wrong. Should never be lower than the blend count for the same month; one that is gets flagged. Leave a month blank rather than guessing: a blank month has no operations rate, whereas a guess becomes a published figure.

Improvement suggestions raised from the board. Accepting one does not put it in the repository — copy the block it produces into GitHub, and the badge only turns green once the reference actually appears in the deployed roadmap index.

🔒 A colleague is currently editing this record.
You can view, but not edit, until they finish. Any changes they save will appear here within a few seconds.
You have read-only access, so you can look at this record but not change it. Request more access
1Define & Status
Use for repeat clusters and upstream/downstream WOs
For a repeating cause: the reference of the IA investigation raised, as proof it was investigated. It is a record of the reference only — the board is not connected to the IA system.
2Investigation Walk-through & Evidence Review (optional)
Not needed for most investigations. If you are not sure where to start, the walk-through takes you step by step through the evidence for the path chosen at triage.
Structured investigation walk-through with per-item evidence review. Walk each step even if a failure is found earlier — multiple contributory failures need capturing separately. Each failure adds candidate cause codes for the coding section below.
3Cause coding
Recorded on the report. Not counted in the KPIs.
Low confidence means further evidence is needed before the CAPA is finalised.
Sub-type (b) should trigger the pilot re-verification protocol before the next tweak-and-issue
45 Whys
5Containment
Manage this dropdown list in Admin → Containment Options
6Actions
7Root Cause & Sign-off

Cost of Quality Detail

Three figures from one set of inputs. Actual counts material only where it was wasted. Potential counts the excess or dropped-out volume whether or not it could be reused — the mistake was the same size either way. Hours is the rework time, unpriced. Material added to bring the blend back into specification is not a cost — it went into product that was sold — but it is what becomes the excess, so it counts in potential.
Blend basis
Full CoGs for the original blend
Recovery method
Cost of the extra product made by overblending — the volume above what was ordered, not the material added to correct it
Additional labour time (hours)
@ £–/hr
@ £–/hr
@ £–/hr
External analysis
Third-party testing to confirm the blend or the fix. Leave blank if none was needed.
Cost of Quality £0 Money actually lost. The reported figure.
Rework hours 0 Operator, formulator and lab time, unpriced. Reported.
Potential £0 Had nothing been reusable — every dropped-out or overblended volume counted as lost.

New Misblend Log

Capture the trigger data for a new misblend. Fields mirror the ERP misblend email so entering it manually is a straight copy. Once the webhook is live, this will populate automatically.
Required fields are marked with *. Other fields can be filled during the investigation. Cause code is not captured here — it is determined through the investigation walk-through.
The day the blend went wrong, not today — it decides which month’s figures this counts in.
Owner list drawn from authenticated users. Leave as “— Awaiting assignment” if not yet decided.
Created record will appear on the board with status NEW.

Investigation Walk-through

Step of

Triage

Complete the three sub-phases in order. Triage decides whether this record proceeds to investigation and, if so, on which path.
The day the blend went wrong — decides which month it counts in. Correct it here if it was logged later.
!

Confirm return

The record will stay open on the Triage tab in this returned state until the source has responded. Re-open triage at that point to confirm the outcome.

Cancel the walk-through?

It will no longer hold up sign-off.

The steps already answered stay on the record, and your name and the time are recorded.

You can pick it up again by unticking Stop using the walk-through in section 2.

Try another investigation path

The current path walk-through has not identified the root cause. Pick another path to investigate. The current attempt is preserved — you can review it any time from the walk-through summary.
Choose the next path to try

Edit path

Admin only
Path metadata drives the badge and picker labels. Steps drive the walk-through the investigator sees.

Steps

Action

Corrective actions fix the root cause of this specific event. Preventive actions stop recurrence — they land in the Actions Register for tracking to closure.
1 Where this came from
2 The action
Cancelling keeps the action on the record and stops it counting as outstanding. It is not deleted, and the reason is kept with it.
—
3 Completion and verification
Required before the action can be marked Complete. “Done” is not evidence.
Not yet complete
Data shared across all users of this board. Storage abstraction ready for Supabase swap-out.