What Is a Tech Pack and When It Stops Working as Your Single Source of Truth

It is Tuesday morning and a designer at a $12M womenswear brand is on a call with a factory in Portugal. The factory is asking which version of the placket measurement is correct. There is a tech pack PDF dated three weeks ago, a follow-up email with a revised pocket placement, a WhatsApp photo of a sample with handwritten notes, and a costing sheet that references SKUs the merchandiser has since renamed. The production manager joins the call and pulls up a different file. The fabric BOM line lists a mill that the sourcing team stopped using last season. Nobody on the call knows which document is current. The order ships in eleven weeks.

What is a tech pack apparel teams actually use, and where does it break?

A tech pack, in apparel, is the engineering specification for a garment. It is the document a brand hands to a factory so the factory can produce the same garment, the same way, twice. A complete tech pack contains the technical sketch with callouts, the points of measure across a size run, the construction details and stitch types, the bill of materials covering shell, lining, trims, labels, and packaging, the grading rules, the artwork and print placements, the wash or finish instructions, and the labeling and care content for the destination market.

That is what a tech pack is supposed to be. What it actually becomes, in most apparel businesses between $5M and $100M in revenue, is a PDF that was true on the day it was exported and a series of unstructured corrections layered on top of it.

The question is not what a tech pack is. The question is when it stops working as the single source of truth, and what to do at that point.

What is inside a tech pack at a working level of detail?

Most templates cover seven sections. The technical flat with front, back, and detail callouts. Points of measure tied to a size run, with tolerances. Construction notes covering seam types, stitch counts per inch, and reinforcement points. The bill of materials with supplier, color, composition, weight or denier, consumption per unit, and price. Grading rules across sizes. Artwork files for prints, embroideries, or branding. Labeling, packaging, and folding instructions, including hangtag placement, polybag size, and carton specs.

A tech pack that is missing any of these forces the factory to make decisions on your behalf. Those decisions are usually wrong, and they show up in the first sample, the second sample, the production sample, and sometimes in production itself when it is too late.

A good tech pack is also written to remove ambiguity, not to demonstrate completeness. A points-of-measure table that lists thirty-two measurements but tolerances on only nine is worse than one with eighteen measurements and tolerances on all of them. The factory will not ask. They will guess.

When does the tech pack stop working as your single source of truth?

This is BP1 of the 6 Breakpoints framework, the point where product data starts fragmenting. It happens before inventory truth weakens, before order flow gets shaky, before warehouse execution drifts. It is the upstream breakpoint, and almost every downstream operations problem in an apparel business has a piece of its origin here.

From conversations with apparel founders and ops leaders, the breakpoint is rarely sudden. It looks like this. Your design team uses Illustrator and a tech pack template in Excel or a PLM tool. Your costing lives in a separate spreadsheet maintained by a merchandiser. Your factory works off the PDF you exported, plus whatever has been clarified over email since. Your sales team builds the wholesale linesheet in a different tool, often pulling product names and codes from a third sheet. Your DTC team uploads to Shopify with copy and imagery prepared by yet another person. Your 3PL receives goods against a packing list that may or may not match the SKU structure in any of those sources.

When I started Uphance, the pattern I saw repeatedly was that brands had a tech pack and they had data, but they did not have a connected system of record. The tech pack was current on the day of export. Costs were current in the spreadsheet. SKUs were current in Shopify. None of those three were current with each other.

That is when the tech pack stops being a source of truth. It becomes one of several competing sources, and the cost of reconciling them lands somewhere on your team.

What does that fragmentation actually cost?

For a $15M brand running wholesale, DTC, and a 3PL simultaneously, the operational tax is measurable. Six to nine hours a week of reconciliation across Shopify, the 3PL, and wholesale. A two to three percent oversell rate at peak. One full-time equivalent doing data plumbing instead of merchandising, planning, or supplier development.

Those numbers are downstream of BP1. Oversell at peak does not start in the warehouse. It starts when the SKU on the linesheet, the SKU in the WMS, and the SKU on the production order do not share a common parent record. Reconciliation hours do not start in finance. They start when the costing sheet says one landed cost and the ERP says another, because the BOM was updated in the tech pack but never flowed into costing.

If you are spending more than two hours a week answering the question “which version is current,” the tech pack is no longer functioning as a source of truth. It is functioning as a historical artifact.

Why does PLM alone not solve this?

Most brands in the $10M to $20M band, which is the predictable breakpoint zone, eventually buy a PLM tool. PLM solves part of the problem. It versions the tech pack, it gives the design team a single place to maintain construction and BOM data, and it usually has a comment thread that beats email.

PLM does not solve the rest. The BOM in PLM has to flow into costing. The SKU structure has to match what shows up on the wholesale order, the production order, and the WMS. The artwork and copy have to be available to the people building the linesheet and the Shopify product. If your PLM is connected only to itself, you have moved the breakpoint, not removed it.

The stand to take here is direct. PLM as a standalone tool does not fix BP1. It defers it. The tech pack only becomes a true single source of truth when the product record, including specs, BOM, costing, SKU structure, and assets, is connected to production, inventory, and orders inside one system.

This is what brands are actually buying when they replace three to five tools and a stack of spreadsheets. They are not buying a better tech pack. They are buying the connection between the tech pack and everything downstream of it.

How do you know your tech pack process is the breakpoint?

There are a handful of operational signals. The factory asks which version is current more than once a season. Sample iterations exceed three rounds on more than a third of styles. Costing changes after the BOM is finalized because trims were missed. The wholesale linesheet ships with a SKU that does not exist in your inventory system. A retailer chargeback hits because the carton label does not match the ASN, and the carton label was generated from a packing spec the warehouse maintains separately from the tech pack.

Any one of these is a signal. Two or more is the breakpoint.

The team-level signal is also clear. Your designer is doing data entry. Your merchandiser is reconciling spreadsheets. Your production manager is answering version questions instead of managing capacity. Your ops lead is the human integration layer between five tools.

What does a working product data backbone look like?

A working backbone has a single product record per style. The record carries the technical specs, the BOM with current pricing from current suppliers, the size and color matrix, the SKU structure, the artwork and copy assets, the costing rollup, and the season and collection metadata. When any of those changes, every downstream artifact reflects the change without rekeying.

The linesheet is generated from the record. The production order is generated from the record. The Shopify product is generated from the record. The 3PL receives a SKU and master pack structure that came from the same record. The tech pack PDF, when the factory needs one, is also generated from the record, with a version stamp.

In that architecture, the tech pack is not the single source of truth. The product record is. The tech pack is one of several outputs. That is the correct architecture. Asking a document to be a system of record is asking the wrong thing of a document.

This is what the 6 Breakpoints framework calls fixing BP1 at the architecture level rather than at the template level. Template-level fixes, better tech pack templates, more disciplined file naming, shared drives with stricter folder structures, work for a season or two and then collapse under volume.

What changes for the factory relationship?

The factory relationship gets cleaner in two specific ways. First, the version question goes away. The factory works against a record that has a clear current state and a clear history. Second, costing conversations get more honest. When the BOM is connected to costing in real time, you can have a conversation about why the landed cost moved, with the line items in front of both parties. You stop having the conversation where the factory quotes against one BOM and you cost against another.

The sample round count also drops. Not because the factory got better, but because the spec they are working from is internally consistent. Most extra sample rounds are not driven by quality. They are driven by ambiguity in the spec or contradiction between the spec and the comments.

Where does this connect to the rest of the operation?

It connects everywhere. A clean product record is the precondition for OTB planning that actually reflects committed inventory. It is the precondition for a wholesale order flow that does not require manual SKU mapping. It is the precondition for a 3PL handoff where the master pack on the carton matches the master pack in the WMS. It is the precondition for returns posting back to the right SKU in days, not weeks.

The tech pack, on its own, cannot deliver any of that. The product record can.

What this means for an apparel operations team

If your team is debating tech pack templates, you are working on the wrong problem. The template is not the breakpoint. The disconnection between the tech pack, the costing sheet, the linesheet, the production order, and the inventory record is the breakpoint. A better template will not close that gap. A connected product record will.

The practical move, in the $10M to $20M zone, is to stop treating the tech pack as the system of record and start treating it as an output. That requires a product data backbone that the design, production, sales, finance, and warehouse teams all read from and write to in their own contexts.

The diagnostic question to ask this quarter is simple. If a SKU specification changes today, how many systems and humans does it have to touch before the change is reflected in your wholesale linesheet, your Shopify catalog, your production order, and your WMS? If the answer is more than one system and more than one human, BP1 is open, and the tech pack is carrying weight it was never built to carry.