Improving a Grade: The Redesign Playbook

Grade improvement is rarely a redesign of everything. Assessments concentrate deductions: typically one or two features - a sleeve system, an inseparable component, a material pairing - account for most of the distance between a unit's actual grade and its main material's potential. The playbook is therefore diagnostic before it is creative. Get the sensitivity list from the assessment. Rank the fixes by grade points recovered per euro of redesign and unit-cost change. Sequence the work onto commercial events that are already funded - relaunches, retooling, supplier changes. And re-grade the drawings at each step, because the cheapest grade is the one secured before anything is manufactured.

For each unit below target, list every deduction the criteria apply, its point value, and its cause class: material identity, separability, surface (labels, sleeves, print), chemistry in process (inks, adhesives), geometry (size, format). The classification matters because the fix owners differ - material identity belongs to packaging engineering, surface to artwork and print procurement, separability to component sourcing. Most deduction audits surface the same surprise: the expensive deductions are in the classes owned by the teams not in the room. Convene the audit with artwork, print and component suppliers represented, or expect the playbook to stall at the main material, which was never the problem.

Score each candidate fix on recovered points, one-off cost, recurring unit-cost effect, and risk (shelf-life validation, line trials, regulatory neighbours such as food-contact re-approval). The ranking disciplines enthusiasm: a mono-material conversion may recover twenty points at tooling cost, while a wash-off adhesive recovers six at a specification change - and six points at no tooling, across forty lines, is usually the better first programme. Boundaries matter more than points: a fix that crosses a grade threshold is worth its fee differential every year; a fix that moves within a grade is worth its claim value only. Rank boundary-crossers first, then cheap points, then the structural programme.

Redesigns attached to funded events - a relaunch's artwork budget, a mould's end-of-life, a contract renewal's specification window - cost a fraction of standalone projects and face a fraction of the internal resistance. The playbook therefore maps each fix to its carrier event and works backwards to a latest-start date per unit, with the 2030 grading date and the 2038 C-exit as the immovable frames. Units with no carrier event before their frame get one of two honest labels: funded standalone (the fee maths justifies it) or managed exit (it does not). The dishonest label is 'later', unattached to any event, which is how programmes arrive at 2029 with the whole queue intact.

Each executed fix closes with the assessment loop: re-score the unit from its new specification, test where the change touches borderline behaviour (a new adhesive is a testing trigger, a new label size usually is not), and version the file. Lock the gain contractually: the improved specification goes into supply agreements and artwork approvals as the controlled version, with the change gate from the formats station preventing drift back. Grade improvements decay silently through a thousand small procurement decisions unless the specification system holds them; the playbook's final step is the one that makes the first three permanent.

One beverage line, three candidate fixes. A wash-off label adhesive: four points, no tooling, a specification change - rank one, and it crosses the B boundary, so its value is the fee differential on the whole line's tonnage. A separable pump on the sister line: seven points, moderate tooling, carries a shelf-life validation - rank two, sequenced onto the pump supplier's contract renewal. A mono-material conversion of the pouch range: nineteen points, full tooling and barrier re-validation - rank three, attached to the 2028 retooling, without which it would not be funded this decade. The ranking took an hour; it ordered three years of work and priced each year. That is the playbook operating as intended.