The product development process, stage by stage

Strategy is the work of getting specific before you spend money on samples. Answer these on one page:

• Customer and occasion. A real person in a real moment. Behavior suggests product; vague taste doesn't.
• The problem. What is missing or annoying in what they buy today?
• Price position. Accessible, mid or premium, and a target retail range.
• The hero product. The statement piece the rest of the range sits beside.
• What you won't make. Saying no to adjacent products keeps the first range coherent.

Finish with a one-line value proposition: "[Product] for [customer] who [use case], so they can [outcome], unlike [the obvious alternative]." If you can't write it, you aren't ready to talk to factories.

Minimums, packaging and lead time are consequences of how you sell. Pick the first channel, then work backward:

• Your own site: you control price and story, and you pay for ads, returns and photography.
• Wholesale: buyers need margin (often half of retail or more), barcodes, case packs and delivery windows.
• Marketplaces: reach, with fees and lookalike competition squeezing margin.
• Events and clubs: small quantities, hard dates, tactile first impressions.
• Preorder: demand signal before you commit cash, if you are honest about timing.

Then sketch the price architecture: target retail, likely wholesale, a landed-cost ceiling and room for returns and selling costs. A product that works at $90 on your own site can fail at $45 wholesale if it lands at $28. If wholesale is in the plan, design for it first.

A brief is how you stop describing your product differently in every chat window. Two suppliers who get two different stories send two quotes you can't compare.

You don't need a 40-page tech pack on day one. You need a document that says what it is, who it's for, how it must look and feel, how many you want, what you can pay and when you need it. At minimum: the use, annotated reference images, dimensions or size range, materials and weights, colors, components and hardware, decoration placement, labels and packaging, quantity by size and color, target cost, delivery date and destination, testing needs, and open questions.

For apparel and other sewn goods, the brief grows into a tech pack once you start sampling. See what is a tech pack and the free tech pack template.

Version the brief. The revision that matches the approved sample is the one that goes on the purchase order.

First decide the route, then find the people.

Domestic wins when cash is modest, the launch is close, minimums must stay low or you need to fix fit in person. Overseas wins when the product needs construction or materials you can't get locally at the right cost, your volume supports the minimum and your calendar has room for sampling and freight. Hybrid is common: develop and wear-test domestically, then take a clean package overseas to scale.

Then build a longlist of about five suppliers, searching by construction and material rather than generic product names. Qualify them on capability, capacity and communication before price, and send three to five the same RFQ: the same brief, the same questions, the same quantity tiers. Our guide on how to find a manufacturer covers the search, including how to trace who makes for brands you admire.

Every sample should answer a question. Can this supplier make something good? Does our design work in hand? Is production ready to match a locked standard?

The usual stages are a stock or reference sample, a development sample, revised samples, a pre-production sample (PPS) and a sealed approval sample that bulk must match. Score each one from 1 to 5 on accuracy to the brief, materials, construction, fit or function, color and decoration, packaging and communication. Write change requests as measurements, keep a change log, and approve the PPS in writing.

Custom work almost always takes more than one round. Budget for it instead of treating revision as failure. The full method is in our product sampling guide.

Model margin from landed cost, not the quote. Landed cost is everything it takes to get a sellable unit to your shelf: product, samples, tooling, packaging, freight, duty, brokerage, inspection, receiving and a reserve for defects, divided by the good units you actually receive. Run it at three quantities and map when cash goes out and comes back. See how to calculate landed cost.

Then negotiate the whole deal, not just the unit price: quantity breaks, sample credit against bulk, tooling ownership, packaging inclusions, inspection rights, payment milestones, delivery dates and what happens if goods are late or wrong. A slightly higher price with sample credit and balance paid after inspection can be the better deal.

An approved sample without a purchase order is a memory. A PO without a sample reference is a hope. The PO should state:

• Final specification and the approved sample's revision, photos and measurements
• Quantity by size and color, and the acceptable variance
• Unit price and every setup charge
• Tolerances for critical measurements
• Production schedule and ship-ready date
• Inspection standard and defect tolerance
• Packaging, labeling and carton marks
• Shipping term (Incoterm) and named place or port
• Payment milestones
• Remedies for late delivery, defects or nonconformance
• Who owns molds, artwork and tooling

For a meaningful overseas order, budget a third-party pre-shipment inspection against the sealed sample and the PO, and agree in advance what happens if it fails: rework, sort, discount or cancel.

A launch date is not a production lead time. Work backward from the date inventory has to be in customers' hands, through receiving, customs, freight, inspection, production and sampling, and give every step an owner.

The typical critical path: brief locked, shortlist, RFQ, first sample, feedback, PPS, PO and deposit, materials, production, inspection, export booking, freight, customs, warehouse receiving, spot check, content and launch.

Ask what each lead time assumes. "Three weeks production" often assumes materials are already in the building. Put factory holidays on the calendar at the start; Chinese New Year shuts much of the supply chain for weeks. Price air freight before you promise a date that needs it, and ask for packed carton sizes early, because freight is charged on volume or weight, whichever is more. Add buffers you're willing to defend, and set a design freeze date after which only safety or compliance issues change the product.

The first order is a learning purchase. Launch in a way that produces evidence: seed people who live the use case, track returns by reason, keep defects separate from "changed my mind," and tag feedback by product, size, color and channel.

Write decision rules before launch: reorder if sell-through hits a set level in a set time with returns under a limit; redesign if the same defect repeats; pause if the margin doesn't clear after two honest tests.

A couple of weeks after the first deliveries, hold a post-production review. Record actual lead times, true landed cost, defect rates, best sizes and colors and the words customers used. That file is what makes the second run faster and more profitable than the first.

Realistic ranges, from brief to first inventory:

It runs longer if samples need extra rounds, production hits a factory holiday, or the cost model forces a redesign. Speed comes from clear decisions and clean documents, not from skipping sampling.