Ravensbourne Glass Case Study

Case study deck

Eleven slides about one customer, from Coalport, an invented shift handover product, and Ravensbourne Glass, an invented container glass plant. The usual case study order is challenge, solution, results and a quote. This one keeps that spine and adds the slide vendors normally cut: the internal argument in which the buyer nearly said no, with five reasons for and five against, and the eight week condition they attached to the deal.

Create a presentation with OneCraft11 slides, read in order

Every slide, in order

The whole deck as it renders, one slide after another. Read it the way the audience would.

CASE STUDY

Seven Minutes

Instead Of Twenty Two

How Ravensbourne Glass rebuilt shift handover across three furnaces

Coalport

Shift handover software for continuous process plants

Published

August 2026

Slide 1 · cover · Puts the headline result in the title rather than the customer logo, so the slide works as a link preview.

The Plant

Three furnaces that cannot be switched off, twelve forming lines, and a handover that happened on paper five times a day.

Scale

Three furnaces

Continuous melt. A furnace is lit for twelve years and never stopped.

Twelve lines

Container glass, forty two active moulds, changeovers every nine days.

620 people

Of whom 380 work a rotating five shift pattern around the clock.

Handover

Five a day

Every shift change on every line, plus the furnace desk and the cold end.

On paper

A carbon book per line, filled in by hand, filed in a drawer in the office.

22 minutes

Median time from the first handover word to the incoming crew starting.

What broke

Nothing written

Mould changes agreed verbally were not in the book the next morning.

No search

Finding the last time a fault appeared meant reading eleven weeks of paper.

No engineers

Maintenance saw the book at 8am, twelve hours after the fault was logged.

The cost

4.1 percent

Rejects at the cold end, against a 2.5 percent target set in 2019.

Repeat faults

Nineteen of the top twenty faults in 2024 had appeared before.

Overtime

Crews staying late to explain a shift the book could not carry.

Slide 2 · pillars · The customer and the problem on one slide, so a reader can decide in ten seconds whether this story is theirs.

What The Plant Manager Said

Every line keeps a carbon book. Nobody reads a book from a line they are not on, so a fault on line 4 is invisible on line 9.

The Book

Handover

Twenty two minutes of handover, five times a day, twelve lines. That is eleven shifts a week spent standing at a bench talking.

The Wait

Time

When a mould faults again, somebody reads back through paper to find the last fix. Most of the time they give up and guess.

The Search

Recall

Maintenance start at eight. A fault logged at nine in the evening waits eleven hours before an engineer has any idea it exists.

The Gap

Response

The customer audit asks for traceability on a batch. Producing it means a person with a box of carbon books and two days.

The Audit

Evidence

Slide 3 · checklist · The five failures in the customer own words, which is what makes another plant manager recognise themselves.

Why They Nearly Said No

The internal argument, in April 2025.

FOR

The reject number

A 1.5 point move on rejects is worth about £1.9m a year at this volume.

Searchable history

Every fault, every fix, every mould, findable in seconds rather than days.

Maintenance sees it live

An engineer gets the fault at 9pm, not at 8am the following morning.

Audit in the meeting

Batch traceability produced while the auditor is still in the room.

Crews asked for it

The shift reps raised handover twice in the 2024 works council minutes.

AGAINST

Screens on the floor

Hot end operators wear gloves. Nobody believed a tablet would survive.

A failed attempt in 2021

A previous system was abandoned after nine weeks and is still remembered.

Fourteen weeks

Training 380 people across a five shift rota with no production stoppage.

No signal at the furnace

The furnace desk had no usable network, and cabling it was not budgeted.

The number was a forecast

The £1.9m came from our model, not from their own plant, and they said so.

The Deal

Line 9 only, for eight weeks, with the paper book still running beside it. If rejects did not move, the contract ended and nothing was owed.

Slide 4 · comparison · The honest reason the deal nearly died, which is the most useful slide for a buyer at the same stage.

What Went In

Five parts. Four of them are ordinary, and the fourth is the one that made it work.

Ruggedised panels

Fourteen glove operable panels, one per line plus the furnace desk and the cold end. IP65, no keyboard, no mouse.

Structured handover

The handover is a form, not a blank page. Nine fields, five of them mandatory, and it cannot be submitted half done.

Fault history

Every entry is tagged to a line, a mould and a fault code, so the search that used to take two days takes four seconds.

The engineer alert

A fault marked urgent pages the on call engineer immediately. This is the part the crews say changed their evening.

Batch export

One click produces the traceability pack an auditor asks for, covering any date range across all twelve lines.

Slide 5 · pillars · What was actually deployed, in five parts, so the story is repeatable rather than magic.

FOURTEEN WEEKS

How It Was Rolled Out

Workstream	Weeks 1-3	Weeks 4-8	Weeks 9-11	Weeks 12-14
Hardware	Line 9 panel	Network to furnace desk	Eleven more panels	Cold end and spares
Handover form	Drafted with shift reps	Live on line 9	Revised after feedback	Standard on all lines
Parallel paper	Book unchanged	Book and system	Book and system	Book retired, week 14
Training	Six shift reps	48 operators, line 9	190 operators	136 operators
Decision gate	Baseline agreed	Rejects reviewed	Board approved rollout	Contract to 2029
The week 8 gate was the customer decision, measured on line 9 alone, with the paper book still running beside it.

Slide 6 · timeline · The programme as the customer experienced it, quarter by quarter, including the parallel run.

COLD END DATA

Reject Rate

4.1%

Before

2.6%

After

£1.9M

Annual value

Slide 7 · metrics · The trend that carries the story, measured on the plant own cold end data rather than on ours.

What Changed

Measured over the twelve months to April 2026, against the twelve months before the first panel went in.

BEFORE

AFTER

Handover took a median of 22 minutes, five times a day, on every one of the twelve lines.

Rejects at the cold end ran at 4.1 percent against a target of 2.5 percent set in 2019.

Nineteen of the top twenty recorded faults in 2024 had happened at least once before.

An urgent fault logged at 9pm reached a maintenance engineer at 8am the next morning.

Producing batch traceability for a customer audit took one person about two days.

Handover takes a median of 7 minutes, because the form is filled during the shift, not after it.

Rejects run at 2.6 percent. The plant credits about two thirds of that move to handover.

Repeat faults are down to nine of the top twenty, because the last fix is findable in seconds.

The same fault pages the on call engineer inside a minute, at any hour, on any line.

The traceability pack is exported in the audit meeting, covering any range across all lines.

Slide 8 · comparison · Five measured changes rather than one, because a single number invites the reader to disbelieve it.

£1.9M

Annual value, on the plant own numbers.

THE PLANT MANAGER

Twelve months after the paper book was retired.

I did not think a screen would survive the hot end and I said so at the time. What convinced me was line 9 in week eight. Same crew, same moulds, same furnace, and the rejects moved. The number in this deck is ours, not theirs.

Callum Rhodes, Plant Manager, Ravensbourne Glass

Slide 9 · testimonial · The customer voice attached to the one number the whole story rests on.

What They Would Tell You

From the closing review with the Ravensbourne project team, May 2026.

DO NOT

Pick your worst line for the pilot, not your best one.

Let the shift reps write the handover form themselves.

Keep the paper book running until the crews stop using it.

Agree the gate number and the date before anything is installed.

Train on the panel on the floor, never in a meeting room.

Turn maintenance alerts on late, once the data is trusted.

Do not roll out to twelve lines at once to save eight weeks.

Do not let the form grow past nine fields. It gets skipped.

Do not promise the vendor forecast to your own board as a target.

Do not run training on day shift only. Nights find the problems.

Do not skip the network survey at the furnace. It cost us three weeks.

Do not retire the paper book before the crews ask you to.

Slide 10 · content · What the customer would tell the next plant, which is more useful to a reader than another benefit list.

CONCLUSION

The handover was the constraint.

Ravensbourne had already bought vision systems, a new annealing lehr and a mould shop upgrade. None of them moved the reject rate, because the thing losing information was a carbon book on a bench at six in the morning. The eight week pilot on line 9 cost about £41,000 and settled an argument the plant had been having since 2019.

Coalport, with the Ravensbourne Glass project team

Slide 11 · closing · States the transferable conclusion rather than thanking the reader, so the page ends on an argument.

The structure

What each slide is doing, so you can reuse the order even with different content.

Slide 1cover: Puts the headline result in the title rather than the customer logo, so the slide works as a link preview.
Slide 2pillars: The customer and the problem on one slide, so a reader can decide in ten seconds whether this story is theirs.
Slide 3checklist: The five failures in the customer own words, which is what makes another plant manager recognise themselves.
Slide 4comparison: The honest reason the deal nearly died, which is the most useful slide for a buyer at the same stage.
Slide 5pillars: What was actually deployed, in five parts, so the story is repeatable rather than magic.
Slide 6timeline: The programme as the customer experienced it, quarter by quarter, including the parallel run.
Slide 7metrics: The trend that carries the story, measured on the plant own cold end data rather than on ours.
Slide 8comparison: Five measured changes rather than one, because a single number invites the reader to disbelieve it.
Slide 9testimonial: The customer voice attached to the one number the whole story rests on.
Slide 10content: What the customer would tell the next plant, which is more useful to a reader than another benefit list.
Slide 11closing: States the transferable conclusion rather than thanking the reader, so the page ends on an argument.

What makes this deck work

The hesitation slide is the useful one

Five arguments for and five against, including that the headline value was a vendor forecast rather than the plant own figure. A reader at the same stage of the same decision recognises their own meeting, which no benefit list achieves.

The deal condition is written down

One line only, eight weeks, paper running in parallel, and nothing owed if the rejects did not move. Publishing the terms of the pilot tells the next buyer exactly what to ask for, and it costs the vendor nothing to say.

Five changes, not one

Handover time, reject rate, repeat faults, engineer response and audit preparation, each measured over twelve months against the twelve before. A single number invites disbelief; five that move together are much harder to argue with.

Questions people ask

What slides go in a case study presentation?

Cover with the headline result, the customer at a glance, the challenge in their words, why they chose you, the solution, the implementation, quantified results, a named quote and a next step. Ten to twelve slides is the working range.

How long should a case study presentation be?

Up to about thirty minutes including questions, with one idea per slide. If it is being read rather than presented, assume a few minutes and put the result in the title so the cover works as a link preview.

What makes a case study credible rather than promotional?

Named people, the customer own measurement rather than yours, at least one number that did not move as hoped, and the conditions the customer attached before they bought. This example carries all four.

How is a case study deck different from a written case study?

The written version is read alone and can carry detail. The deck is performed or skimmed, so it leads with the outcome, keeps one idea per slide, and ends on a transferable conclusion rather than on a summary of the story.

How do I get a customer to approve a case study deck?

By measuring the story with their numbers, not yours. Every figure here comes off the plant's own cold end data, the timeline matches what they lived through including the parallel run, and the condition they attached to the deal is written down. A customer signs off a version they recognise far faster than a flattering one.

Should a case study show the implementation timeline?

Yes, because the buyer reading it is deciding how much work they are about to take on. Slide six runs the programme quarter by quarter as the customer experienced it, including the weeks the old process ran in parallel. A results slide with no timeline behind it reads as a claim rather than a project.

Build your own in about a minute

The button below opens the generator with this use case already described. Change the wording to match your own, generate, then edit anything you like.

Make one like this

Sources

Written and checked by the OneCraft team. Last checked 22 August 2026.