Tech Pack vs Pattern: What Factories Actually Need

People mix these up constantly. A designer emails me a beautiful pattern and expects a finished garment back. Or they send eight pages of measurements with no pieces to cut. Both stall on day one. A custom tech pack runs $500 to $2,000 per style and takes 5 to 20 hours to build (Guru, 2022), so the confusion gets expensive fast. Here's the clean version: a tech pack is the specification, the description of what the finished garment must be. A pattern is the solution, the shaped pieces that achieve it. Different jobs, different authors, different moments. Get the order wrong and your sample line stops before it starts.

What's the real difference between a tech pack and a pattern?

The difference is spec versus solution. A tech pack documents every requirement of the finished piece: measurements, materials, construction, tolerances. A pattern is the physical or digital set of shaped pieces that, when cut and sewn, produce that piece. One tech pack typically drives one pattern, and the pattern only works because the tech pack told it what to hit.

Think of it like architecture. The tech pack is the blueprint and the engineering brief. The pattern is the cut timber and the joinery. You wouldn't hand a carpenter a stack of pre-cut wood and expect a house with no drawings. Yet brands do exactly that with garments, then wonder why the sample comes back wrong. If you want the full anatomy of the spec side, our guide on what a tech pack is breaks down every section.

Here's the machine-readable version, the one I wish every new client read first.

Specification vs construction instruction: the distinction that trips everyone

The cleanest way to hold the two apart is one phrase: a tech pack specifies, a pattern instructs construction. A specification tells you the target state. The finished chest must measure 54 cm, the seam must be a 5-thread overlock, the shell must be a 280 gsm French terry. None of that tells a cutter where to place their shears. A construction instruction does the opposite. A pattern piece says: cut this exact shape, on this grainline, with 1 cm of seam allowance on these edges and a notch 8 cm down from the shoulder.

Both describe the same garment. They just answer different questions. The tech pack answers "what must be true?" and the pattern answers "what do I cut and where do I join it?" A factory needs both answered before a sample can exist, but only one of them, the spec, has to come from you. The pattern is a derived artifact. It's the factory's working answer to your question, and it changes every time the answer gets refined.

Here's a test I give clients who still can't tell them apart. Ask yourself: could two skilled patternmakers, working from this alone, produce the same garment? A tech pack should get them within tolerance of each other even if their pattern pieces look nothing alike. A pattern gives you one fixed geometry with no target to check it against, so you'll never know if it drifted. That's the whole difference in one question. The spec is portable and verifiable; the pattern is specific and unverifiable on its own.

The two documents also overlap more than people expect, which is part of why they get confused. A few data points live in both places, and knowing which ones stops a lot of arguments on the sampling floor.

What is a tech pack, exactly?

A tech pack is the complete instruction set for one style. Techpacker defines it as a document built from eight core sections: flat sketches, a bill of materials, measurement specs, a cost sheet, size gradings, colorways, construction notes, and revision history (Techpacker). Nothing about it is decorative. Every field prevents a specific mistake on the floor.

The bill of materials is where most gaps hide. Sewport lists it as the components inventory: fabrics, threads, zippers, buttons, labels, every trim by supplier and placement (Sewport). Miss the interlining weight and your collar rolls. Miss the thread ticket and your topstitch puckers. If you're starting from zero, a structured tech pack template forces you to fill each field instead of guessing.

For a deeper look at how these sections get laid out on the page, the notes on what each page contains cover spacing, callouts and what a factory reads first.

What is a pattern, and what does it actually do?

A pattern is the set of shaped pieces used to cut fabric. Front, back, sleeves, facings, linings, each drafted with seam allowances, notches and grainlines so cutters and sewers know exactly what to do. The pattern turns a flat sketch and a spec sheet into geometry. It's the bridge between an idea and a stack of cut panels.

Here's a thing I've watched happen more than once. A brand sends a home-printed PDF pattern from a course they took. It has no seam allowance marked, no notches, no grainline. My patternmaker can't tell where the front ends and the facing begins. We rebuild it from scratch, which is slower than if they'd sent nothing but a clear tech pack. A pattern with missing marks isn't a head start. It's a cleanup job.

What does a pattern actually include?

A finished pattern is more than a set of outlines. Each piece carries five things a cutter and a sewer read without asking. Miss any one and the panel slows the line or gets cut wrong. Here's the checklist my patternmakers run before a pattern is called complete.

• Blocks. A block, or sloper, is the factory's tested base shape for a body type, with no seam allowance and no style detail. Patternmakers develop new styles by modifying a block they already trust, which is a big reason your factory prefers to draft its own pattern rather than adopt yours.
• Seam allowances. The extra margin added outside the stitch line, commonly 1 cm on most seams and more on hems. If it isn't marked, a cutter doesn't know whether the line they see is the finished edge or the cut edge. That single ambiguity is the most common reason I reject an incoming pattern.
• Notches. Small clips that tell a sewer which edges match, how a sleeve eases into an armhole, and where a dart or pleat starts. Notches are how two curved edges get joined without guesswork.
• Grainlines. The arrow that must sit parallel to the fabric's warp. Ignore it and the panel twists, the seam ropes, and the garment hangs off-square after the first wash.
• The marker and nesting. The marker is the full layout of every piece across the fabric width, nested to waste as little cloth as possible. The marker sets your fabric consumption, which flows straight into the cost sheet. Tight nesting is where a factory quietly saves you money per unit.

The marker is the part brands never think about and factories obsess over. On a recent knit order, retightening the nest dropped fabric consumption from 1.42 metres per unit to 1.31. On a 3,000-piece run that's 330 metres of cloth saved, and none of it changed a single measurement on the tech pack. That's the pattern side earning its keep in a place the spec never touches. The tech pack tells us the target; the pattern and its marker decide how cheaply we can hit it.

Digital vs paper patterns: does it matter?

Both work, but they travel differently. Paper patterns are physical, so they can't be emailed or version-controlled cleanly, and every re-grade means re-drawing by hand. Digital patterns, drafted in CAD, carry precise points and grade in seconds. Most factories I work with prefer digital because it plots cleanly, nests automatically, and stores with the style file. Paper still shows up, and it still gets the job done, but it slows the back-and-forth on every revision.

The software split matters when you exchange files. Pattern CAD systems like Gerber AccuMark, Optitex and Lectra dominate factory floors, while CLO 3D and Browzwear handle 3D sampling and virtual fit. The common handoff format between them is DXF, specifically the AAMA/ASTM DXF flavour that carries piece boundaries, grainlines, notches and grade rules as structured data rather than a flat drawing. If you ever do supply a digital pattern, send the AAMA DXF, not a PDF. A PDF is a picture of a pattern; the DXF is a pattern a machine can actually read, grade and plot. That difference alone can save a day of re-digitising at the factory.

Who writes the tech pack, and who drafts the pattern?

Two different people, usually. The designer or a technical designer writes the tech pack, because it captures design intent and measurement targets. The patternmaker drafts the pattern, and very often that patternmaker sits inside the factory. This split matters: the person who knows what the garment should look like is rarely the person best placed to draft the pieces that build it.

That handoff is where good factories earn their keep. Give us a clean spec and we'll draft the pattern, sew the first sample, and send it back with notes. You review, we revise. If you're sourcing production and want to understand who does what, the team at Portugal Clothing Factory can walk you through how a factory-drafted pattern fits into a first order.

Where do the tech pack and pattern sit in pre-production?

They sit at opposite ends of the same sequence. Pre-production runs from design intent to a cut-ready marker, and a custom tech pack that takes 5 to 20 hours to build (Guru, 2022) is the first hard deliverable. The pattern comes next, downstream of the spec, and everything after it, the sample, the grade, the marker, depends on the spec staying stable.

The order isn't a preference. It's a dependency chain. You can't draft a pattern without a spec to hit, you can't sew a sample without a pattern, and you can't measure a sample without tolerances to measure against. Each step feeds the next. The diagram below shows which artifact drives each stage, and why a late spec change is so expensive: it doesn't just edit one box, it re-triggers every box to its right.

Do I need a pattern before I contact a factory?

Usually not. Most clothing factories draft the pattern from your tech pack as part of the sampling process. What they need first is a clear specification, not a finished set of pieces. Send a strong tech pack and a factory can take it from sketch to sample. Send a pattern with no spec and they've got shapes with no rules.

This is the single most useful thing I can tell a new brand. Don't pay a freelancer to make a pattern before you've even chosen a factory. Your factory will likely redraft it to their own blocks anyway. Spend that money on a proper tech pack instead. A hoodie tech pack, for example, tells us far more than a hoodie pattern with no seam data ever could.

If your spec is solid but you don't have time to build the document, our tech pack service turns your sketches and measurements into a production-ready pack from $299, which is the cheapest insurance you'll buy against a rejected sample.

When does the factory make the pattern, and when should you supply one?

Ninety percent of the time, let the factory draft it. That's the default for a new brand, a new style, or any garment where the factory owns the sewing. Their patternmaker knows their machines, their blocks and their sewers' habits, so a pattern drafted in-house sews faster and truer than one imported cold. You supply the spec; they supply the geometry.

There are a few cases where supplying your own pattern earns its place. If you're moving an established, proven style to a new factory and you want to protect a fit customers already love, a validated pattern preserves it. If you hold a signature block that took seasons to perfect, you protect it as intellectual property. And if you're producing something so technical, structured tailoring, complex outerwear, that your own patternmaker has already solved it, handing over a clean AAMA DXF saves the new factory a costly re-development. Outside those cases, a pattern you commissioned before choosing a factory is usually money spent twice. The factory redrafts to its own blocks, and your freelance pattern becomes reference material at best.

How do the two work together across sampling?

They loop. The tech pack sets the target measurements, the patternmaker drafts pieces to hit them, we sew a sample, then we measure that sample against the tech pack's tolerances. Anything outside tolerance gets corrected in the pattern, logged in the tech pack's revision history, and resampled. Grading is the bridge that turns one graded-to-fit sample into a full size run.

Grading deserves its own note, because it's where tech pack and pattern meet most tightly. ASTM publishes body-measurement standards that patternmakers grade against. D5585-21 covers women's misses sizes 00 to 20 (ASTM), while D6960/D6960M adds plus sizes 14W to 40W (ASTM). These sit under ASTM Subcommittee D13.55 (ASTM). Your tech pack carries the grade rules; the pattern executes them.

How does inspection tie back to the spec?

Once garments are produced, they get inspected against a sampling standard. ISO 2859-1, mirrored by ANSI/ASQ Z1.4, defines how many units to pull and how many defects to accept. QIMA notes that AQL 2.5 is the most common general limit, 1.0 applies to higher-risk items, 4.0 covers minor cosmetic issues, and critical defects sit at zero (QIMA). Every one of those defects is judged against your tech pack, not against the pattern.

What goes wrong when brands send one without the other?

Plenty. The two failure modes are a pattern with no tech pack, and a tech pack with no realistic spec. The first gives a factory shapes it can't verify. The second gives it rules with nothing to cut. Both trigger the same result: a stalled sample and a defect that gets caught at inspection, measured against a spec that was never complete.

Here's the mechanism people miss. Rejection isn't a matter of opinion. An inspector pulls a sample size set by the standard and checks each unit against your documented tolerances. If your neck opening is meant to be 40 cm plus or minus 0.5 cm and the pattern was drafted with no such spec, the garment isn't wrong, it's undefined. Undefined garments still get rejected, because the buyer's own AQL check has nothing to certify them against. A factory can draft a pattern from a good tech pack all day. It cannot build anything trustworthy from a pattern that carries no tolerances, no BOM and no grading logic.

Which mistakes cost the most?

The expensive errors nearly always come from conflating the two documents. Over years of first samples, the same four keep landing in my inbox. Each one traces back to treating a pattern as if it were a spec, or a spec as if it drafted itself.

• Sending a pattern with no tolerances. The factory has shapes but no pass/fail rule, so the sample can't be verified and inspection has nothing to check against.
• Editing the spec after the pattern is drafted. A "small" measurement change forces a re-grade and a new marker, and every downstream cost gets recalculated.
• Supplying a PDF pattern instead of an AAMA DXF. The factory re-digitises by hand, which adds a day and quietly introduces its own drift.
• Assuming one pattern fits every factory. Blocks differ between factories, so a pattern that fit perfectly at one supplier can sew a full size off at another.

Notice that none of these are pattern-making failures. They're sequencing and document-hygiene failures, which means a proper tech pack and the right order prevent all four before a single piece gets cut.

So the order is simple. Spec first, pattern second, sample third, revise, repeat. Start with the specification, use a proper tech pack template to capture every field, and let the factory draft the pattern that solves it. That sequence has saved more of my clients' launch dates than any other single habit. If you want the full picture of how this fits together, the tech packs pillar links out to every piece of the process, from a first t-shirt spec to inspection.

FAQ

Sources

Every statistic and definition in this article traces to one of the following. Each is listed once here, and cited inline where it is used.

• Guru
• Techpacker
• Sewport
• images.pexels.com
• Portugal Clothing Factory
• otextilpt.sirv.com
• ASTM International
• ASTM International
• ASTM International
• QIMA