Custom Dog Carrier Sizes: Building a Programme

Pet carrier production desk · Updated 2026-10-07 · 14 min read

Custom pet travel bag programmes are usually quoted with the airline dimension printed into the technical pack, so there is no argument at inspection.

Why the size decision comes before the design sketch

The instinctive way to develop a dog carrier range is to design one hero size, photograph it, sell it, and then add sizes when customers ask for them. That sequence feels efficient because it defers spending, and it is almost always more expensive. A size added late is a new pattern, a new baseboard, a new carton, a new size guide asset and, in most cases, a compromise on the original silhouette, because the hero proportion no longer holds at the extremes. Programmes that decide the ladder first spend their development budget once and reuse it every season.

What the size decision actually locks in is larger than the pattern. It determines how many cutting markers the cutting room will run, whether one baseboard blank can serve two rungs, whether the strap webbing can be bought in a single width, and how many distinct cartons the export plan has to manage. It determines the shape of the size guide the brand will publish, which in turn determines how many customers self-select correctly online. And it determines the photography set: a four-size ladder shot at a consistent scale is a merchandising asset, while four sizes photographed at four different scales read as four unrelated products.

There is also a purchasing consequence. Retailers do not buy a size, they buy an assortment. A buyer reviewing a three-rung ladder is reviewing a shelf block with a beginning, a middle and an end; a buyer reviewing a single size is reviewing a product and being asked to imagine the rest of the range. Suppliers who arrive with a completed ladder, a recommended size mix and a reasoned view of which rung sells fastest are having a different conversation from suppliers who arrive with a sample.

The counter-argument - that a brand should prove demand before building four patterns - is legitimate for a first drop. The resolution is to design the ladder but stage the tooling: develop all four patterns, cut the full set of prototypes, and put only two or three rungs into the first bulk order. The patterns exist, the grading rules exist, and the fourth rung can be activated in a later season without a fresh development cycle.

Practically, that means the first question in a brief should not be what the carrier looks like. It should be how heavy the heaviest dog in the target market is, and how the brand intends to describe that dog to a customer standing in front of a shelf or scrolling a size filter.

Weight bands hold up better than breed lists

Breed lists are a marketing instrument; weight bands are an engineering instrument. A breed chart reads beautifully on a hangtag and fails the moment a customer owns a mixed-breed dog, a rescue of unknown parentage, or a dog whose weight sits at the edge of the category. Weight is the only measurement a customer can actually produce at home, and it is the measurement that correlates best with the structural load the carrier has to carry.

Weight alone is not sufficient, and this is where most programmes under-specify. A 7 kg terrier and a 7 kg spaniel occupy different volumes: the terrier is compact with a deep chest and short back, the spaniel is longer and shallower. For carrier fitting, two secondary dimensions matter more than any other: the back length measured from the withers to the base of the tail, and the chest depth measured behind the front legs. A shell that clears both will accept almost any dog in the band; a shell dimensioned only by weight will fit the shallow dogs and squeeze the deep-chested ones.

The practical structure is four bands with deliberate overlap. Bands that abut exactly - 5 to 9, 9 to 18 - force a customer whose dog weighs 9.0 kg into an arbitrary decision, and generate the returns that follow. A one kilogram overlap at each boundary, combined with guidance that says choose the larger shell for deep-chested or long-backed dogs, resolves most of the boundary cases before they become a support ticket.

The floor dimensions in that ladder are working numbers, not laws, and every serious programme adjusts them. A brand selling into a market where small dogs dominate will find the XS and S rungs carry 70 percent of volume and may want a narrower XS floor. A brand selling into a market with a large-dog culture will find the M and L rungs drive the review and will want the L floor longer rather than taller. What should not change is the discipline of specifying floor area and height separately, because they fail in different ways: too little floor area and the dog refuses to settle, too little height and the dog cannot stand or turn.

Write the band into the tech pack as a load class, not as a suggestion. Once the band is a load class, every downstream decision - foam density, handle reinforcement, webbing width, carton strength - can be derived from it instead of argued about again on every sample round.

Grading a pattern is not scaling a pattern

The most common error in a size ladder is photocopy grading: take the small pattern, add 15 percent to every dimension, cut it, and call it a medium. It produces shells that look correct in a flat product photograph and behave wrong in three dimensions, because the proportions of a dog do not scale linearly and neither do the proportions of a bag around a dog.

Consider height. A Chihuahua standing inside a carrier needs roughly the same headroom as a Cocker Spaniel needs, because both animals clear their own shoulder height plus a small margin. Doubling the weight band does not double the required internal height. Length, by contrast, scales hard: moving from a 5 kg dog to a 20 kg dog adds substantial back length and very little chest depth relative to that gain. A correctly graded ladder therefore grows faster in length than in height, and grows floor width only enough to keep the dog from being compressed at the shoulder.

Materials do not scale either, and this is where ungraded ladders fail in the field. Foam laminated at 6 mm in a small shell does not need to become 12 mm in a large shell, because the larger dog distributes its weight over a larger area; what the larger shell needs is a stiffer baseboard, not a softer wall. Zipper gauges step at thresholds rather than continuously: a #5 coil is correct through the S band and a #8 belongs on M and L, because the failure mode at size is zipper spread under load, not zipper length. Handle reinforcement follows the same logic - the same webbing is correct on XS and S, and the reinforcement plate roughly doubles at M.

Strap width is the clearest example of threshold grading. A 20 mm strap is comfortable on a 4 kg dog and inadequate on a 25 kg dog regardless of padding, because pressure is force divided by contact area and the force has gone up sixfold while the area has gone up perhaps twofold. Widening from 20 mm to 38 mm across the ladder costs cents and is the single most effective change a programme can make to perceived quality at the top rung.

The rule that keeps this manageable is to grade dimensions continuously but grade materials at thresholds. Write the continuous rules into the pattern, write the threshold table into the bill of materials, and review both at the same sample round so the two do not drift apart.

What to hold common across the ladder, and where sharing stops

A four-rung ladder with entirely bespoke components in every rung is a four-project development budget wearing the costume of one project. The purpose of commonality is to let a programme buy in volume, hold colour consistency, and reorder a single rung without reopening the whole specification. The target most successful programmes hit is 70-80 percent component commonality by count, with the remaining 20-30 percent deliberately size-specific.

Hold common: shell fabric and colourway, lining fabric, webbing colour and finish, zipper tape colour, hardware finish and plating spec, all branding components, thread, piping and the baseboard material family. These are the items where a variation is visible to a customer as inconsistency, and where consolidated buying produces real unit-cost movement. A single navy 600D across four sizes is one dye lot to control; four navies are four chances to ship a ladder that does not match on shelf.

Do not hold common: the baseboard blank and its thickness, the foam density under the floor, the strap width, the handle reinforcement plate, the panel dimensions themselves, and the carton. These are the items where holding common produces either an over-built small shell or an under-built large one, and both outcomes cost more than the saving. A baseboard is a cheap part and a load-bearing one; standardising it across a 2 kg and a 32 kg rung is a false economy that will surface as floor sag.

Cartons deserve their own note because they are usually forgotten until the freight quote arrives. A four-rung ladder shipped in four carton sizes creates a packing plan with four cube efficiencies, and the large size almost always dominates the freight cost per unit. Programmes that plan cartons during development - often by nesting two small shells per carton and specifying a flat-pack large - protect margin without touching the product. Export carton planning is worth reading alongside this section; see export carton planning for pet carrier orders for the cube arithmetic.

One more commonality lever is worth stating because it is frequently missed: the size guide asset itself. If all four rungs are photographed against the same reference dog, the same background and the same scale marker, the ecommerce team can build a comparison module that materially reduces mis-sizing. That asset is produced once and used for seasons.

Load cases differ by rung, and testing should follow the failure mode

A single test protocol applied identically to all four rungs will over-test the small end and under-test the large end, which is the worst possible allocation of a testing budget. The failure modes genuinely differ. Small shells fail at the seams and the zipper, because the animal's weight is concentrated over a short span and because small dogs move more, scratch at closures and push at mesh panels. Large shells fail at the baseboard and at the handle-to-shell junction, because the absolute load is higher and because a sagging floor is visible and immediately reported.

A sensible protocol starts from the band's rated weight and applies a safety factor rather than a fixed figure. Static load testing at four times the upper bound of the band, held for a defined period, is a defensible starting point for structure; seam peel and grab testing is then run at the panel junctions that carry the load. Mechanical methods referenced from bodies such as ASTM International give the production team and the brand's own reviewer a shared vocabulary, so a number on a spec sheet means the same thing on both sides. Independent perspectives on pet restraint and containment performance, including the testing programme published by the Center for Pet Safety, are useful context when deciding how a claim should be worded.

Drop testing is the case most often skipped and most often needed. A loaded shell dropped from waist height onto a hard floor is not an exotic event; it is what happens when a customer sets the bag down on a tile floor with a dog inside and the handle slips. Test the baseboard and the corner seams at the rated weight, not empty, and record corner deformation rather than pass/fail, because deformation tells the design team where to add structure.

Cycle testing belongs on the closure. Zippers are opened and closed hundreds of times in a product life, and the failure mode is usually slider spread or tape delamination, not tooth breakage. Cycling the closure at the rated weight with the shell loaded - rather than on an empty bench sample - catches the geometry problems that only appear under tension.

Document the results per rung, not per programme. A brand that can show a retailer the load case and the result for the L rung specifically is answering the buyer's actual question, which is whether the largest bag in the range is safe to sell.

Naming the rungs so customers self-select

Size naming is a conversion decision disguised as a labelling decision. Three naming systems are in circulation and they perform differently. Weight-only naming (Small 5-10 kg) is the clearest and the safest, because it is the one measurement a customer can verify. Dimension-only naming (46 x 28 x 30 cm) is precise but requires the customer to go and measure a dog, which most will not do. Breed-example naming (fits Cocker Spaniel) converts well emotionally and misleads structurally, because owners of the same breed know their dog is bigger or smaller than the example.

The strongest commercial solution is a weight band as the primary label, with the two secondary body dimensions printed underneath as a tie-breaker, and the internal floor dimensions on the product page for the subset of customers who will measure. That structure lets 95 percent of customers decide from weight alone while giving the remaining 5 percent the data they need, and it gives customer service a single script for the boundary cases.

Consistency across the ladder matters more than cleverness within it. If the XS is labelled by weight, the L must also be labelled by weight. Mixed systems - a weight band on two rungs and a dimension on two - produce a size guide that reads as unfinished, and unfinished size guides are a documented driver of add-to-cart abandonment in this category.

Hangtags and retail labelling then have a second job: they must repeat the band in the same words used online. A customer who reads 5-10 kg on a product page and 5-10 kg on a shelf tag is reassured; a customer who reads a weight online and a breed name on the tag is being asked to reconcile two systems at the moment of purchase.

Finally, name the ladder in a way that survives extension. If the brand will eventually add an XL or a narrow-fit variant, leave naming room. Renaming a ladder after launch invalidates every published size guide and every retailer's planogram entry, which is a real cost that is never budgeted.

Splitting the minimum order across four rungs

The commercial constraint most programmes hit is not the total minimum, it is the per-size minimum. A colourway minimum of 500 units sounds generous until it is divided across four rungs and three colourways, at which point each individual size-colour combination is a very small production run with poor cutting efficiency and a high risk of shade variation between lots.

The workable approach is to treat the ladder as one production order and manage the split inside it. Cutting all four rungs from the same fabric lot in the same production window is what protects colour consistency, and it is the single strongest argument for planning the split before the order is placed rather than letting each rung be ordered independently by a different buyer at a different time. A split order also shares the set-up cost across rungs, which means the effective minimum per rung is lower than it would be standing alone.

A reasonable first-order split for a four-rung ladder in a general market is heavily weighted to the middle: roughly 20 percent XS, 35 percent S, 30 percent M and 15 percent L. That distribution reflects the fact that the small and medium rungs carry most unit volume while the large rung carries disproportionate revenue and disproportionate freight cost. Markets differ sharply - urban apartment markets push the split two rungs smaller, suburban and rural markets push it one rung larger - and a brand with twelve months of its own sales data should override the rule with its own numbers immediately.

Reordering is where the ladder either pays off or becomes a burden. Because the patterns exist and the grading rules are documented, a reorder can be a single-rung run against an existing specification, which is faster and cheaper than a new development. Programmes that document the ladder properly at the outset typically find their second-year reorder cycle materially shorter than their first-year development cycle, and the saving is entirely a consequence of having decided the size structure once.

The related discipline is colourway planning across the ladder: not every rung needs every colour. Concentrating a seasonal accent colour on the two rungs that sell fastest is a standard technique for keeping a colourway above its minimum without inflating total inventory, and it is covered in more depth in colour range planning for custom pet bags.

The development calendar for a four-rung ladder

A ladder takes longer to develop than a single shell, but not four times longer, because the grading rules do most of the work. The realistic sequence starts with a dimensional brief: the four weight bands, the floor and height targets for each, and the target breeds or body types the brand wants to serve. From that brief the design team produces a full set of paper patterns before cutting any fabric, and reviews them together, because the review that matters is the comparison between rungs rather than the assessment of any single rung.

The first physical round should be the complete ladder in a single neutral fabric with provisional baseboards. Reviewing a partial ladder is the most common source of rework: a small and medium that look correct together will still be wrong once the large arrives and the grading step turns out to be too aggressive. Fit testing at this stage uses weighted dummies rather than live animals, because a dummy gives repeatable measurements and because a live dog's tolerance for standing still during a fitting session is limited.

The second round introduces production-intent materials: the actual baseboard, the actual foam, the actual zipper gauge and the actual strap width per rung. This is the round in which the threshold decisions are validated, and it is the round most likely to generate a change to the bill of materials rather than the pattern. Brands that skip straight from a provisional round to bulk almost always discover the zipper gauge issue in the field.

Once the ladder is approved, bulk scheduling is a single window rather than four: bulk production 35-50 days from approval and deposit, with the rungs cut in the same production run to hold shade. Inspection against AQL 2.5 covers dimensional tolerance per rung, seam integrity at the load-bearing junctions, baseboard flatness and closure cycling, with the per-rung results recorded separately so a retailer can be shown the evidence for the size they are buying.

Working from a documented ladder, on a partner production base audited to BSCI and ISO 9001 with seven lines and a monthly capacity measured in hundreds of thousands of units, the constraint on a size programme is almost never capacity. It is clarity at the brief stage.