Home › Field notes › Repair or Replace a Module? Cost, Remaining Life, Consistency and Bran

Modular Bag Pro · 2026-10-11 · 2,761 words · 13 min read

The repair-versus-replace decision for a module is made on four criteria - repair cost measured against replacement cost, remaining service life, visual consistency with the rest of the platform, and what the customer experiences while waiting - and cost on its own settles only the easiest cases. Commercial baselines sit at 500 units per reference, a 6-10 working day sampling stage and 35-50 days of bulk production, with repairability features fixed in the tech pack and lots inspected to AQL 2.5. The boundary is civilian load carriage for work tools, hiking, commuting and travel, and the figures below are planning illustrations for a service policy rather than any statement about what a given brand's customer promise covers.

What Repair Versus Replace Means at Module Level

Three outcomes sit on the table and they are often confused. Replace the module means swapping the whole unit for a new one. Repair the module means restoring the existing unit - restitching a seam, replacing a slider, re-terminating a webbing tail. Replace the product means retiring the bag, which is the outcome everyone wants to avoid and the one reached most often when no service path exists.

Module-level decisions are cheaper and quicker than product-level ones, and that speed is the argument for modularity in a service context. A pouch can be swapped by the user in under a minute; the same failure on a monolithic bag means a return, an assessment and a decision nobody enjoys making. But modularity introduces a criterion that monolithic products do not have: consistency. A repaired or replaced module sits alongside modules that have aged differently, and the customer sees all of them at once.

Failure mode is what sorts the cases before cost is even discussed. Localised failures - a split seam, a lost slider, a frayed strap end - are repairable by definition, because the rest of the unit is sound. Distributed failures - abrasion across a face panel, UV degradation, a hydrolysing coating - are not, because there is no sound area left to anchor a repair to. Structural failures at a load path require a judgement that no cost table can make for you.

The decision also has a time dimension that is easy to overlook. A repair that takes four weeks produces a customer who has been without the product for four weeks, and that absence is part of the cost. A replacement shipped the same day from regional stock often beats a technically cheaper repair, which is why service stock and repair capacity are planned together rather than separately. Module ranges and their attachment options are listed under modular everyday carry platforms.

Judgement: Sort every case by failure mode before looking at cost - localised, distributed or structural - because a distributed or structural failure is not a repair decision at all, and treating it as one produces a repair that fails again within weeks.

The Four Criteria: Cost, Remaining Life, Consistency and Brand Experience

Cost is the first criterion and the least decisive. Remaining life is second: repairing a unit with two years of service left is rational, repairing one with two months left is not, and the estimate comes from the failure record rather than from the customer's description. Consistency is third and the one most brands discover late: a replacement module in a fresh dye lot, or a repaired panel with visible stitching, changes how the customer feels about a product that still works perfectly.

Brand experience is fourth and it is not soft. It is the number of days the customer waits, the number of contacts required, and whether the outcome was explained. Two repairs of identical cost produce completely different customer outcomes depending on whether the customer was told what would happen and when. In practice this criterion decides more cases than cost does, because it is the only one the customer can actually observe.

Weight the criteria rather than counting them. A case that fails on load path is settled regardless of the other five, and a case where turnaround differs by a month is usually settled by brand experience even when repair is cheaper. The table is a way of making the weighting explicit so two people reach the same answer.

Selection rule: Apply the criteria in a fixed order - load path first, then remaining life, then turnaround, then cost - because a decision reached by cost alone replaces units that should be repaired and repairs units that should be replaced, and both errors cost more than the correct choice.

Which Failures Are Repairable at All: A Failure-Mode List

Repairability is a property of the failure, not of the brand's goodwill. A separated seam along a pouch mouth is repairable in 15-25 minutes on a standard machine, and the repair is as strong as the original if the stitch type and density are matched. A lost zipper slider is a two-minute fix with the correct slider and a pair of pliers. A frayed webbing tail is re-terminated in under ten minutes. These three account for the majority of service cases on most platforms, and all three are worth designing for.

The second group is replaceable rather than repairable. A cracked buckle gate, a lost compression strap, a distorted ladderlock - all are parts rather than repairs, and all are resolved from service stock in minutes. This group is where a spare part programme earns its money, and it is why the parts list should be built from the failure list rather than from the bill of materials.

The third group is terminal. Abrasion across more than roughly a third of a face panel, a hydrolysing coating, a delaminated laminate, foam that has lost recovery, or UV degradation across the whole shell - none of these can be repaired, because the material itself has changed state. Offering a repair here is worse than offering nothing, since it delays a correct replacement and consumes goodwill twice.

The boundary case is the load path, and it never goes to repair. A shoulder strap anchor, a hip belt root, a harness attachment - any component whose failure drops the load - is replaced or the product is retired, regardless of cost. No service policy should leave that judgement to a cost comparison, and no technician should be in a position where the cheaper option looks available.

Takeaway: Publish a three-group failure list - repairable, replaceable from service stock, terminal - and never route a load-path failure into the repair group, because the first two groups absorb most service volume and the third is where a well-meant repair becomes a safety problem.

Designing a Module Worth Repairing

Repairability is decided at the drawing stage, and it is mostly about access. A seam that can be opened from one side without destroying the lining is repairable; the same seam buried under three layers is not, even though it is technically identical stitching. A webbing tail that is tucked and stitched closed takes ten minutes to expose; one terminated with a accessible bar tack takes two.

Hardware choice runs the same way. A machine screw releases a hardware plate in under three minutes; a rivet requires drilling and leaves a hole. A zipper fitted with a standard-size slider and an open end can be re-slidered; a sealed-end zipper with a proprietary slider usually cannot. None of these choices cost more at the unit level, and all of them are invisible to the customer until something breaks.

The pattern in that table is that repairability and recoverability point the same way. Fasteners that come apart, single-polymer panels, standard sliders and accessible seams make a module easier to fix and easier to recover at the end, because both outcomes depend on the same property: the ability to take the thing apart without destroying it. Where a repair changes a performance surface, the same standard methods confirm it has not regressed - seam strength to ASTM D5034, abrasion resistance to ASTM D3884 and bond strength of any laminate to ASTM D751.

Spec rule: Require accessible seams, standard-size sliders and mechanically fastened hardware on every module the brand intends to service, and write those three items into the tech pack as checked line items, because a swap agreed at the bench quietly removes repairability while leaving the approved sample looking identical.

Where the Repair Happens: Field, Workshop or Factory

Three locations are available, and each suits a different failure. Field repair is done by the user with a supplied kit: a slider, a cord, a patch, a printed card. It resolves perhaps a third of cases, takes 3-10 minutes, and requires a design that can be worked on without tools - which is another argument for standard sliders and accessible hardware.

Workshop repair covers everything needing a machine: seams, webbing, panel replacement. A regional repair partner with an industrial machine handles most of the remaining cases in 20-40 minutes of bench time, plus collection and return. This is the tier that keeps turnaround short, and a brand selling in more than one region should have at least one partner per region, because return freight across a continent adds two weeks to every case.

Factory repair is the slowest and the most informative. It suits distributed or structural failures where the unit needs to be assessed rather than fixed, and it is the only tier where the failed part can be examined by the people who specify it. Building that loop - a quarterly box of failed units sent back with their claim records - is the single most useful thing a service programme does for the next design cycle.

Freight decides the tier as often as capability does. Returning a module from another continent costs several times the value of the part, which is why regional service stock exists: a replacement sent locally by express in 3-5 days, or air in 5-8 days for a consolidated batch, beats a factory repair that is technically free. Cube planning for consolidated returns uses about 28 CBM for a 20GP and about 68 CBM for a 40HQ, and the mailer a customer uses to send a module back should be qualified to ISTA 3A, since a part damaged in transit turns one case into two.

Bottom line: Place the repair at the tier closest to the customer that can actually resolve it - field for sliders and straps, a regional workshop for seams and panels, factory only for assessment - because return freight across a continent adds roughly two weeks and often costs more than the part itself.

Cost Model: Putting a Number on the Decision

The decision reduces to a ratio. On the repair side: bench minutes multiplied by the loaded labour rate, plus the part, plus return freight, plus the administrative time of raising and closing the case. On the replacement side: the module cost, plus outbound freight, plus administrative time, minus any recovery value from the returned unit. The lower number wins unless one of the non-cost criteria overrides it.

Worked through on a mid-size pouch: a bench repair at 25 minutes with a small part and return freight lands at perhaps 40-60% of the module cost, which favours repair comfortably. Move the same case to a factory tier with return freight from another region and the ratio passes 100%, at which point replacement wins even before turnaround is considered. The ratio is sensitive to freight and to tier, not to the part.

Administrative time is the line consistently omitted. Raising a case, diagnosing it, corresponding, picking, packing and closing it consumes 8-15 minutes of paid time whether the outcome is a repair or a replacement, and that fixed cost is what makes small-value cases uneconomic to process individually. Pre-kitted envelopes and a self-service claim form cut it to roughly a third, which changes the ratio on precisely the cases that are otherwise marginal.

Verdict: Calculate the ratio with freight and 8-15 minutes of administrative time included on both sides, because those two lines decide most cases, and a comparison made on part cost alone will route cheap modules to repair and expensive freight to replacement.

How the Decision Feeds Back Into the Next Specification

Every repair or replacement is a data point, and the value lies in the coding rather than the volume. Four fields are enough: platform and generation, part, failure mode, and tier used. Coded consistently across a year, those four fields produce a Pareto of causes that is more reliable than any field test, because it reflects how the product was actually used rather than how it was tested.

The loop has to close on a schedule to work. A quarterly review that takes the top three causes and asks one question of each - is this a material, a geometry or a specification problem - produces one or two concrete changes per cycle. Without the schedule the data accumulates and nothing changes, and a service programme becomes a cost centre rather than an input to design.

Feedback also works downwards, into the module list. If one module accounts for a disproportionate share of claims, the honest response may be to retire it rather than to redesign it, and a platform with a stable interface can do that without affecting anything else. That is the quiet advantage of modular construction in a service context: a weak component can be withdrawn while the platform continues.

Service Stock, Procurement and Programme Timing for a Repair-First Policy

A repair-first policy still needs replacement stock, because a meaningful share of cases resolve that way and the ones that do are the visible ones. The stocking split follows the failure list: parts for the replaceable group, whole modules for the terminal group, consumables for the field-repair group. Sizing each against observed consumption takes one cycle; guessing all three at launch usually over-stocks parts and under-stocks modules.

On the production side, capacity stands at 200,000 units a month - 137 people, 149 machines, 7 production lines and 4,950 m² of SGS-verified floor - under a founder who has been in the trade since 2004 at an operation set up in 2014. Service batches begin at 500 units per reference; a sample takes 6-10 working days, or 12-15 where the part is complex, and the USD 50-150 sampling charge comes back on the order. Production then occupies 35-50 days, goods are checked against AQL 2.5 under the plan in ISO 2859-1, and they leave FOB Xiamen with payment on T/T 30/70; new tooling or screens add USD 300-2,500 where a part has none.

Timing is the practical constraint on all of it. A service batch ordered from nothing needs the sampling and production cycle plus sea freight at 25-35 days, so a policy that promises short turnarounds has to be backed by stock rather than by a fast production team. Where a brand wants both short turnaround and low inventory, the answer is a small regional buffer replenished on the annual production calendar, sized on measured consumption instead of on optimism.

Programme teams building a service policy from scratch should fix the repairability items - accessible seams, standard sliders, mechanical fasteners, documented failure coding - at the same time as the module list, since all four are specification items rather than service items. Workshop and inspection arrangements are described on the services page, work-platform module options under modular work backpack platforms, and the service planning desk on the contact page.