Home › Field notes › Modular Bag Programmes: Own-Design or Platform-Adaptation Route

Modular Bag Pro · 2026-10-10 · 2,673 words · 12 min read

A modular bag programme has one structural decision to make before anything else: define the attachment interface, the chassis geometry and the compliance file from scratch, or adapt a platform that already exists and change only what differentiates it. In industry shorthand those two paths are sometimes labelled with a pair of abbreviations, OEM for the first and ODM for the second, and this article uses own-design route and platform-adaptation route instead because those names describe what actually happens. The choice is not about prestige; it is about where development cost, iteration time and interface risk land. Both routes end at the same place: MOQ 500 per reference, sampling in 6-10 working days, and mass production across 35-50 days, released at AQL 2.5.

The two routes, described without jargon

On the own-design route, the buyer arrives with a requirement and leaves with an article that did not exist before. Geometry, materials, hardware, interface dimensions and the compliance file are all defined for this programme. Nothing is inherited, which means everything is controllable and everything takes time.

On the platform-adaptation route, the buyer arrives with a requirement and leaves with a variant of something that already works. The chassis, the interface geometry and the base compliance position exist; what changes is configuration, material, colourway, trim, branding and the module set. The buyer controls the differentiation and inherits the engineering.

In industry shorthand the first is sometimes called OEM and the second ODM, and those labels are worth recognising because they appear in supplier correspondence and in sourcing documentation. They are not used as the primary terms here, because they describe commercial arrangements rather than engineering content, and the engineering content is what determines whether a programme succeeds.

What neither label changes is the underlying work. Both routes require a brief, a technical pack, a sample, a wear trial, a compliance file and a lot release. The difference is how much of each has to be created rather than confirmed, and that is the only question worth arguing about.

It is also worth stating what this site does and does not do, since the terminology invites assumptions. This site works on modular load-carriage platforms and their attachment interfaces, and it supports both routes as development approaches. It does not present itself as a contract production service for unrelated bag categories, and nothing here should be read as an offer to manufacture an arbitrary article to order. The two routes describe how a modular programme is developed, not a claim about corporate identity.

Selection rule: Selection rule: Decide first who owns the interface drawing, because every later question about tooling, exclusivity and second sourcing is answered by that allocation of rights, and the answer changes the cost of a first run of 500 units.

What each route actually decides: interface ownership

The decision that matters is not stylistic, it is about who owns the interface. On the own-design route the buyer defines the webbing width, the row pitch and the column repeat, and owns the consequence of that definition forever: every pouch the brand ever makes has to fit it, and every future chassis has to carry it.

On the platform-adaptation route the interface already exists and is already proven across other programmes. The buyer inherits compatibility with an installed base of pouches and accessories, which is a genuine commercial asset, and gives up the ability to change the geometry if a requirement ever demands it. That is usually a good trade, because almost no civilian requirement genuinely needs a novel interface geometry.

The second-order consequence is about lock-in, and it cuts both ways. Owning an interface means no competitor can sell pouches that fit it, which protects accessory revenue and also caps it, because the installed base is only as large as the brand's own sales. Inheriting a shared interface means a larger installed base and more competition for the accessory sale. Neither is obviously better; the right answer depends on whether the brand expects to sell more chassis or more pouches.

A third consequence is documentation. An owned interface requires the brand to publish and maintain its geometry, or customers will not buy attachments in year three. An inherited interface already has published geometry maintained by whoever owns it, which removes a small but permanent obligation from the brand's plate.

Decision criteria by programme type

The route should be chosen from what the programme is trying to achieve, not from what sounds more ambitious. Six criteria cover nearly every case, and scoring them honestly usually produces an answer without argument.

Reading the table in practice: a brand with an engineering team, a funded development budget, a genuine geometric innovation and time before launch should take the own-design route. A brand with a merchandising team, a seasonal calendar and a differentiation that lives in colour, material and configuration should take the platform-adaptation route. Most programmes fall into the second group and pretend otherwise.

The most expensive outcome is a programme that takes the first route without the resources for it. Development spend is committed, the iteration budget runs out before the geometry is right, and the article launches late with an interface that has not been properly proven. That failure mode is entirely predictable from the table above and entirely avoidable.

A useful test before committing: ask what happens if the first three samples are all wrong. On the platform-adaptation route the answer is usually that the configuration is adjusted and the fourth round is close, because the underlying geometry was never in question. On the own-design route the answer is that a geometric assumption is wrong and every subsequent sample inherits it until someone stops and re-derives it. Programmes that cannot fund that contingency should not start there.

Bottom line: Bottom line: Choose the own-design route when the duty profile is new and the platform-adaptation route when the duty profile is already served, and write the choice into the brief before any quotation is requested, since platform adaptation returns a sample in 6-10 working days while a new chassis usually needs 12-15.

Cost structure, line by line

The two routes do not differ in unit cost so much as in where cost is incurred. The own-design route front-loads it into development; the platform-adaptation route moves it into inventory and gives up some control over the final unit price, because the chassis is shared.

The crossover is worth calculating rather than guessing. Divide total development spend by the difference in unit price between the two routes; the result is the volume at which the own-design route becomes cheaper per unit. If expected volume over the article's life is below that figure, the platform-adaptation route is the better financial decision regardless of how attractive the first route sounds.

One line is frequently omitted from this comparison and often dominates it: the cost of being wrong. On the own-design route a geometric error discovered after tooling is a tooling change. On the platform-adaptation route the same class of error was made years earlier by someone else and corrected before the platform reached the buyer. That is not a cost line, but it belongs in the decision.

Timeline and the iteration budget

Calendar is where the two routes separate most visibly. Sampling requires 6-10 working days for a soft-goods item and 12-15 where a moulded shell or laminated panel is involved, and that figure is per round rather than per programme. The own-design route needs more rounds, because geometry is being discovered; the platform-adaptation route needs fewer, because geometry is being confirmed.

An iteration budget should be agreed before the first sample, expressed as a number of rounds with a decision at each. Two rounds is realistic for a configuration change; four to six is typical for a geometry that has not existed before. Programmes without an explicit budget tend to iterate until the calendar runs out, which produces a sample approved under time pressure rather than on merit.

Production timing matches on both routes: 35-50 days once the sample is approved and materials confirmed. The difference is what has to be confirmed. On the own-design route materials may include items that have never been purchased before, with their own lead times; on the platform-adaptation route the bill of materials is largely known and largely stocked, which is why the range of 35 to 50 days usually compresses toward the lower end.

Verdict: Verdict: Budget two sampling rounds rather than one whenever tooling is involved, and add the second round to the schedule up front, because a 35-50 day production slot cannot be compressed to absorb a late tool change.

Risk allocation and how each route fails

Each route has a characteristic failure, and knowing them in advance is most of the risk management. The own-design route fails by running out of iteration budget before the geometry is right. The symptoms are a late launch, an interface that was never wear-tested properly, and a first production lot that reveals what the samples should have.

The platform-adaptation route fails by drifting. Each small configuration change seems reasonable alone; together they move the article outside the assumptions the original compliance file and transit validation were built on. The result is an article that looks like a variant and has quietly become a new one, without the testing a new one would have received.

The control for the second failure is simple and worth writing into the brief: a change register. Log every deviation from the base specification, material, dimension, hardware, trim, and review the accumulated list before production release. If the list is long enough that the original validation no longer obviously applies, re-run it. That single discipline prevents most platform-adaptation failures.

Both routes carry the same production risks and are controlled the same way: AQL 2.5 lot release against ISO 2859-1, with the acceptance limit fixed in writing before the lot is put forward. MIL-STD-105 preceded those tables and holds on in dated documents; anything new should cite ISO.

Takeaway: Takeaway: Allocate risk in writing before the deposit, name which party carries tooling cost, which carries a failed compliance test and which carries a late artwork change, then release payment under T/T 30/70 against those named events.

Documentation package for each route

The deliverables differ more than buyers expect. The own-design route produces a documentation set; the platform-adaptation route consumes one and adds a delta. Knowing which documents are required avoids the common situation where a programme reaches production without the file a retailer later asks for.

Compliance deserves emphasis because it is the deliverable most often assumed rather than produced. REACH (EC 1907/2006) declarations, California Prop 65 position, CPSIA where the article goes to children, EN 1811 for nickel release from metal parts and OEKO-TEX Standard 100 for the textile stack: each is a document with a date, and a retailer can ask for any of them at listing. On the own-design route they are generated for this programme; on the platform-adaptation route they are inherited and then extended for every material substitution the configuration introduced.

Quality gates are identical on both routes

Whichever route a programme takes, the quality gates should not move. Material verification covers tensile under ASTM D5034, abrasion under ASTM D3884 or ISO 12947, water resistance under AATCC 127 and coating bond under ASTM D751. Article verification covers the assembled interface, a wear trial on a fully loaded chassis, and a transit sequence under ISTA 3A where the carton is new.

Lot verification is the gate most often skipped and the one that matters most commercially. A production run is a different object from an approved sample, and the only thing standing between them and the customer is a statistical inspection at AQL 2.5 with a written report. Both routes need it and neither route earns an exemption.

Hardware and chemical checks apply equally. ASTM B117 screens buckles, sliders and D-rings for corrosion; EN 1811 covers nickel release on metal parts in skin contact; AATCC 8 covers colour transfer from dark linings. Our SGS-verified production base runs all of these as standard, and the certificate set is available through the company profile.

Hybrid sequencing and switching routes

The routes are not mutually exclusive and the best programmes often use both in sequence. Start on the platform-adaptation route to establish a range, generate cash and learn what customers actually buy. Then, once a specific requirement has been demonstrated by real demand, take the own-design route for the one article that justifies it.

That sequence inverts the usual instinct, which is to develop the flagship first. Developing the flagship first spends the iteration budget on the article with the least information behind it, and it does so before any revenue exists. Establishing the range first means the flagship, when it comes, is developed against measured demand rather than against a forecast.

Switching in the other direction is also legitimate and often wise. A programme that took the own-design route for a chassis can move to the platform-adaptation route for its modules, because pouch geometry does not need to be novel and the iteration budget is better spent on the chassis. That mixed approach is what most mature ranges actually look like, whatever label they started under.

A final practical note on briefing. Whichever route is chosen, the brief should state the duty profile, the capacity band, the depth limit, the compliance markets and the target unit price in that order. Programmes that brief in the opposite order, price first and duty last, tend to discover late that the price required a material the duty profile cannot accept, and the correction costs a sampling round at minimum.

Neither route is limited by output: 137 people working 149 machines across 7 production lines on SGS-verified floor space of 4,950 m² deliver 200,000 units per month. What sets the schedule is how many references the range commits, since each carries a 500-unit floor. Indicative quotations are issued FOB Xiamen at the 500-unit minimum; settlement is T/T 30/70 on both routes, with 30 percent advancing material purchasing and the 70 percent remainder waiting for inspection release; transit adds 25-35 days by sea, 5-8 by air and 3-5 by courier. Route options for a custom modular backpack programme and full terms on the services page are set out elsewhere on this site.

Spec rule: Spec rule: Start on a platform for the first 500 units if the interface is unresolved, switch to an own-design chassis once attach rate is measured, and keep AQL 2.5 release identical on both routes.