Build-to-Print vs Build-to-Spec: Which One Do You Actually Need?
A customer sent us a four-page drawing for a display harness. Page one said "build to print, no deviations." Page two never stated a wire gauge. Page three called out a connector that was discontinued in 2021. Page four was a signature block.
That drawing cost them eleven days.
Not because we were slow. Because "build to print" only works when the print is actually buildable, and this one was a spec wearing a print's clothes. Nobody had told them that, in four previous rounds of quoting, which tells you something about the industry.
Here is the distinction, what each one costs, who owns the failure in each, and how to tell which one your project actually is.
The sentence nobody says out loud
Build-to-print is not an order type. It is a liability split.
When you say "build to print," you are saying: the design is mine, the execution is yours, and if the harness is wrong in a way the drawing allowed, that is my problem. When you say "build to spec," you are saying: here is what it must do; you decide how; and the design failures are at least partly yours.
Most buyers say the first sentence and mean the second. That mismatch is fine until something fails in the field, at which point the four pages of drawings stop being a document and start being an argument.
| Build-to-print (BTP) | Build-to-spec (BTS) | |
|---|---|---|
| What you supply | A complete, buildable drawing package | Requirements: electrical, mechanical, environmental |
| Who designs | You | The supplier |
| Who owns a design failure | You | The supplier, in most cases |
| Typical NRE | $0-300, drawing review only | $400-1,500, real engineering time |
| First quote accuracy | High | Needs a round or two |
| Cost of a change | Formal ECO, $150-400 | Often free during prototype phase |
| Best for | Mature products, regulated programs, second sources | New designs, prototypes, startups, first builds |
Which one are you actually doing?
Three questions settle it, and the answers are usually uncomfortable.
Does your drawing state a wire gauge for every circuit? If any circuit lists a color but not a gauge, that is not a print. That is a spec with formatting.
Does your drawing specify pull force or reference the terminal datasheet? If it does not, the supplier is choosing your crimp quality. That is a design decision you just outsourced without paperwork.
Have you issued more than two ECOs to fix errors the drawing allowed? If yes, your "print" is still being designed, by your supplier, one rework at a time.
Two or more yes answers and you are buying build-to-spec. Which is fine — but price it and staff it as what it is.
Build-to-print: what a complete package actually contains
A print is complete when a stranger could build the harness without calling you. That is the test. Here is the checklist we apply before quoting a BTP job.
| # | Item | What "complete" looks like |
|---|---|---|
| 1 | Harness drawing | Dimensioned, with branch points and breakout locations |
| 2 | Wire list | Point-to-point, every circuit, no "same as" references without a table |
| 3 | Connector part numbers | Housing and terminal listed separately, with manufacturer and series |
| 4 | Wire specification | Gauge, insulation type, UL style number, temperature rating |
| 5 | Color code | Per circuit, with a stated convention if it deviates |
| 6 | Crimp requirements | Pull force or a referenced terminal datasheet, crimp height window |
| 7 | Shield termination | One end, both ends, or 360 degrees to the shell — not "grounded" |
| 8 | Labeling | Position, content, material, and orientation |
| 9 | Test requirements | Which tests, and the acceptance criteria, not just "tested" |
| 10 | Revision control | One revision level across all pages, and a named contact for questions |
Of the BTP packages we receive, roughly two in three are missing at least three of these ten. That is not a criticism of the buyers. It is a consequence of drawings being written by people who know the product and never handed them to someone who has never seen it.
The five gaps we find most often, and what each one costs
| Gap | What happens if nobody catches it | Cost when caught late |
|---|---|---|
| Wire gauge missing on one circuit | Supplier picks a default; the circuit runs warm or the connector housing does not accept the barrel | Full rework of the batch, 1-2 weeks |
| Discontinued connector | Quote assumes a substitution, buyer finds out at delivery | Re-quote, redesign of the mating side, 3-6 weeks |
| Shield termination unspecified | "Grounded" gets interpreted as one end; EMI complaint arrives from the field | Field retrofit program, unbounded |
| No pull force requirement | Crimp quality is whatever the line habit produces | Nothing measurable — which is the problem |
| Revision mismatch between pages | Half the batch built to Rev B, half to Rev C | Scrap or 100% re-inspection |
The discontinued connector deserves its own sentence. Component end-of-life is not an edge case. In a typical year a meaningful slice of any connector family goes NRND or hits long lead times. A drawing is a snapshot, and connectors age faster than drawings get revised.
Build-to-spec: what the supplier needs from you
BTS does not mean "you figure it out." The quality of what comes back is proportional to the quality of what goes in. Nine inputs cover almost every harness we design.
| # | Input | Why it changes the design |
|---|---|---|
| 1 | What it connects, both ends | Connector families, keying, and backshell space |
| 2 | Voltage and current per circuit | Gauge, and whether it is a signal or a power run |
| 3 | Environment | Temperature, oil, UV, washdown — each rules out insulation families |
| 4 | Motion | Static versus flexing changes wire stranding and routing entirely |
| 5 | Space and bend radius | Rules out overmolds, dictates connector orientation |
| 6 | Standards | UL, CE, industry-specific — drives wire cost 40-80% |
| 7 | Quantity and product life | Decides between a tooled overmold and a heat-shrink boot |
| 8 | Connector preference | Open connector choice is the cheapest lever you own |
| 9 | Budget target | Uncomfortable to state, most useful number on the list |
That last one surprises people. If you tell a supplier the target is $6.50 and they cannot get under $9, you learn in a day instead of after a first article. Hiding the budget does not protect you; it just delays the conversation you will have anyway.
Who owns what when it fails
This is the table to read before choosing a model, not after.
| Failure | Build-to-print | Build-to-spec |
|---|---|---|
| Wrong gauge selected, circuit overheats | Buyer — the print allowed it | Supplier |
| Connector does not mate with the PCB header | Buyer — the print specified it | Supplier, if they selected the connector |
| Crimp pulls apart in the field | Supplier — execution error | Supplier |
| Harness chafes through after 6 months of vibration | Buyer — routing was on the print | Shared, usually supplier — they knew the duty cycle |
| Connector goes EOL mid-program | Buyer owns the re-qualification | Supplier flags it and proposes alternates |
| Lead time slips on a branded part | Buyer — the part was specified | Supplier, if they chose it |
Notice what the table says about connectors. In BTP, every connector decision — availability, counterfeits, lifecycle — lands on the buyer. A supplier reviewing a print is not obligated to warn you that your terminal has been NRND for a year. Good ones do it anyway. Ours do. But the obligation is yours unless you contract it.
What each model costs
Using the same eight-circuit, one-meter harness from our other articles, at 50 pieces:
| Build-to-print | Build-to-spec | |
|---|---|---|
| NRE | $0-300 (review only) | $400-1,500 |
| Unit price, 50 pcs | $9.90 | $10.60-11.50 |
| First order lead time | 2-3 weeks | 4-6 weeks |
| Cost of a design change | $150-400 formal ECO | Often $0 during prototype phase |
| Cost of the third revision | Another $150-400, plus schedule | Usually absorbed — it is still our design |
| Break-even | Cheaper if the design never changes | Cheaper after the second design iteration |
Read the last two rows together. BTS costs more up front and less over a moving design. BTP costs less up front and more every time the drawing moves. If your product is early in its life, the "expensive" option is frequently the cheap one by the third order.
The hybrid most buyers actually need
There is a middle path and it has no standardized name. We call it design assist, and it is what we default to for first builds.
The buyer supplies the nine inputs. We produce a draft drawing, mark every assumption, and send it back for approval before anything is cut. From that point forward it is build-to-print — a real one, complete by the ten-point checklist above.
The buyer gets three things out of this:
A drawing package they can hand to any second source, because it is actually complete
Design decisions made by someone who builds harnesses daily, including the ones nobody thought about: service loops, strain relief, test points
A clean handoff from BTS to BTP at the moment the design freezes, with no re-qualification gap
The supplier gets one thing: a customer who can issue a change order without a meeting, because everyone knows what every decision was.
When to switch from one to the other
| Signal | Switch to | Why |
|---|---|---|
| Design frozen, volume above 500, second order placed | Build-to-print | The engineering is paid for; stop paying for it |
| You are qualifying a second supplier | Build-to-print | BTP is the only fair basis for comparing quotes |
| Three consecutive ECOs were all corrections | Build-to-spec | Your print is describing a design that does not exist yet |
| You cannot answer the nine inputs | Build-to-spec | Someone has to make those decisions; make it the party equipped to |
| A regulator or customer audit is coming | Build-to-print | Auditors want documents with revision levels, not email threads |
How we handle both
When a drawing arrives marked build-to-print, it goes to a process engineer before it goes to sales. If the package is incomplete, you get the gaps listed as questions, not a quote built on guesses. If the package is complete, you get a quote and a short DFM note — things we would change and things we cannot change.
When a spec arrives instead, we do design assist: draft drawing, assumptions marked, your approval before the first cut. The NRE is quoted as a separate line with the tooling ownership terms attached, the same way we handle it for low-volume orders — no engineering cost buried in a unit price where it can only cause a fight later.
And we will tell you which model your project actually is, even when it is not the answer that flatters the document you sent. Eleven days is cheaper than finding out after the first shipment.
Frequently asked questions
Can a supplier change my print without asking?
They should never make a change silently. What a good supplier does is flag deviations as written questions before quoting, and record approved substitutions on the drawing itself. If your harness arrives with unmarked substitutions, that supplier is not documenting their work, and the next failure will be unattributable.
What if I only have a sketch?
Then you are buying build-to-spec, whether you planned to or not. Send the sketch, the nine inputs, and your budget. A capable supplier will return a real drawing within a week. Do not pay for "build to print" pricing until a print exists.
Is build-to-spec more expensive overall?
On the first order, usually yes. By the third order, it depends entirely on how much your design has moved. Every design change under BTP is a formal ECO with a price; under BTS the early iterations are usually absorbed, because the supplier is iterating their own work.
Who pays if the supplier's design fails?
Under a proper BTS agreement, the supplier redesigns and rebuilds at their cost — that is what the NRE premium buys. Get the terms in writing before the first article: what counts as a design failure versus a spec change, and who owns the scrap. Suppliers who will not put that in writing are telling you how the first failure will go.
Can I switch suppliers with a build-to-print package?
That is exactly what BTP is for — but only if the package is complete by the ten-point checklist. An incomplete print means the new supplier inherits your old supplier's undocumented assumptions, and every gap becomes a negotiation about who pays for the discovery.
What is a DFM note?
A short list a supplier sends with a quote: what in your drawing is hard or expensive to build, what they would change, and what they cannot change without your sign-off. If you collect DFM notes from three suppliers and they overlap, the overlap is the truth about your drawing.
How do I know if my drawing is complete?
The stranger test: could someone who has never seen your product build this harness without calling you? If you are not sure, ask a supplier to review it against a checklist before you send it out for quote. Most do it free, because an incomplete package costs them more than it costs you.
Not sure which model your project is? Send the drawing or the spec.Our Email：sales5@xmkehan.com We will tell you which one you are actually buying, list any gaps, and quote accordingly — usually within 24 hours.