The seven new product development process stages that move a product from a rough idea to a real launch are discovery, validation, concept and scoping, design and prototyping, engineering build, testing and compliance, then launch and iteration. Each stage ends with a decision, not a celebration: you either fund the next stage, change direction, or stop. Companies that lose money on new products almost never lose it because the idea was bad. They lose it because they skipped a stage, or because nobody was allowed to say stop.

This guide walks through all seven stages in plain language, with the timelines and budget ranges we see on real projects, and the question you should be able to answer before you spend anything on the stage that follows. It is written for owners and operators who are funding the work rather than building it.

Why the new product development process stages exist

The new product development process stages exist to make your risk cheap early and your spending expensive late. A bad assumption caught in week two costs a conversation. The same assumption caught after launch costs a rebuild, plus the customers who tried the product once and left. This staged logic has been standard practice in product organisations since Robert Cooper’s Stage-Gate framework formalised it in the 1980s, and it holds just as well for a five person company as it does for a manufacturer.

Two things are worth being honest about before you start. First, the stages are not a conveyor belt. You will loop backwards, usually from testing into design, and that is healthy. Second, the stages are not all the same size. Stages one to four typically consume 20 to 30 percent of the budget and remove 70 percent of the risk. If your spend profile looks like the opposite of that, something has gone wrong.

Stage 1: Idea capture and problem framing

Stage one turns a vague idea into a written problem statement with a named customer, a measurable pain, and a reason this is worth solving now. Expect this to take a few days of focused work, not months, and expect it to cost nothing but calendar time.

The output should fit on one page: who has the problem, what they do today instead, what it costs them, and what a win looks like in numbers. “Contractors waste four hours a week rekeying quotes into two systems” is a problem statement. “An app for contractors” is not. If you cannot write the sentence, the later new product development process stages will simply inherit the confusion and make it expensive.

The one question to answer before moving on

Would a specific, nameable customer change how they work to use this? If the honest answer is “they might be interested”, you have an idea, not a problem. Go and talk to eight of them first.

Stage 2: Discovery and feasibility

Stage two checks whether the problem is real, whether anyone will pay, and whether the thing can actually be built with today’s technology at a price that works. This is where technical feasibility and commercial feasibility get tested together instead of one after the other, which is the usual mistake.

On our projects a discovery sprint runs one to three weeks and lands between roughly CAD 5,000 and CAD 20,000 depending on how much technical unknown is in the mix. A web or mobile product with standard integrations sits at the low end. Anything involving hardware, computer vision, or data you do not yet own sits at the high end. That money buys you customer interviews, a competitive scan, a technical risk list, and a rough order of magnitude estimate for the full build.

Feasibility deserves a plain-language note. Engineers talk about readiness levels, a scale NASA developed to describe how proven a technology is, from a napkin concept up to flight-tested hardware. You can read the NASA technology readiness levels definitions directly. The practical version for a business owner: if a core piece of your product has never been done outside a research paper, budget for experiments, not for delivery dates.

Stage 3: Concept and scoping

Stage three decides what you are building first, what you are deliberately not building, and how the pieces fit together. It produces a scoped feature set, a system architecture, a delivery plan, and a real estimate with a range rather than a single number.

This is the cheapest stage to do well and the most common one to skip. Scoping is where you separate the version that proves the business from the version you eventually want. A useful discipline: list every feature, then ask which ones a customer would refuse to pay without. Everything else goes into a later release. We usually spend one to two weeks here, and the deliverable is specific enough that a different team could build from it.

How the new product development process stages apply to AI features

AI changes stage three more than any other stage, because AI features cannot be specified the way a login screen can. A login screen either works or it does not. A model gives you a distribution of outcomes, so the spec has to define an acceptable accuracy level, what happens on the failures, and who is accountable when the output is wrong. Before building, decide three things: which decision the model is actually making, what a human reviewer sees, and what data you already own to evaluate it against.

Most businesses do not need a custom model. They need an existing model wired correctly into a workflow they already run, with guardrails and an audit trail, which is what our AI integration services cover. Training something bespoke is a stage three decision with a very different cost profile, and it is the right call far less often than the market suggests.

Stage 4: Design and prototyping

Stage four produces something people can react to before you write production code. For software that means a clickable prototype: real screens, real flows, fake data. For hardware it means a functional prototype that proves the mechanism, even if it is ugly and tethered to a bench power supply.

Clickable prototypes typically run CAD 8,000 to CAD 30,000 over two to five weeks. Hardware prototypes range far wider, roughly CAD 20,000 to CAD 150,000, because tooling, components, and iteration cycles dominate the cost. In both cases the point is the same: watch five real users try it. The Nielsen Norman Group’s well known finding on testing with five users is that five people surface the large majority of usability problems, so small tests run often beat one big test run late.

Prototyping is the stage where the new product development process stages earn their keep, because changing a prototype costs hours and changing a built system costs weeks. Our product engineering and prototyping service line exists for exactly this window, covering web and software products through to deep-tech hardware.

How the new product development process stages differ for hardware

Hardware adds two stages’ worth of work inside stage four and stage six: sourcing and certification. Lead times on components can be three to twenty weeks, which means your schedule is partly outside your control and your bill of materials needs a second supplier for anything critical. Software can ship a fix on a Tuesday afternoon. A hardware fix may mean new tooling, so the new product development process stages for physical products front-load far more testing before anything is committed to a mould.

Stage 5: Engineering build

Stage five is the production build of the first real version, shipped in increments you can see every two weeks rather than in one reveal at the end. If you go more than a month without using the actual product yourself, the project is drifting.

A focused first version of a web or mobile product generally runs CAD 40,000 to CAD 150,000 across eight to twenty weeks, with the spread driven by integrations, compliance requirements, and how much of your existing process has to be modelled. Connecting to an accounting system, a CRM, or a piece of shop floor equipment is usually where the hours hide, not in the screens themselves. If the product is primarily a customer facing application, our web and app development work covers this stage end to end.

Three things separate builds that hold up from builds that need replacing in eighteen months: automated tests on the parts that handle money or data, a deployment process that any engineer can run, and documentation good enough that your business is not hostage to one developer. Ask for all three in writing. They cost perhaps 10 percent more up front and they decide whether you own an asset or a liability.

Stage 6: Testing, validation and compliance

Stage six proves the product works under real conditions, with real data volumes, real users, and real regulatory obligations. It is not a final week of bug fixing. It runs alongside the build and then gets a dedicated two to four week window before launch.

Four kinds of testing matter to a business owner, and they answer different questions:

• Functional testing: does each feature do what the spec said, including the awkward cases like a refund on a partially shipped order.
• Usability validation: can a new user complete the main task without being trained or talked through it.
• Load and reliability testing: what happens at ten times today’s traffic, and what happens when a connected system goes offline mid transaction.
• Security review: an independent check against a recognised baseline such as the OWASP Top Ten list of common application risks, which is the minimum bar if you hold customer data.

Compliance is the item most often discovered too late. Canadian privacy obligations under PIPEDA, accessibility requirements under the AODA for Ontario organisations, payment handling rules, and sector certification for hardware all have to be designed in rather than bolted on. A compliance review in stage three costs a meeting. The same review after launch can cost a rebuild, which is why the later new product development process stages should never be the first time a lawyer or an auditor sees the plan.

Stage 7: Launch and post-launch iteration

Stage seven puts the product in front of paying customers and then improves it using what they actually do rather than what they said they would do. Plan for the first 90 days after launch as a funded stage with its own budget, typically 15 to 25 percent of the build cost, because that is when the real learning arrives.

Before you launch, make sure you can measure. Analytics and event tracking installed after launch leave you blind for the period you most need data. Decide in advance which three numbers define success, whether that is activation rate, time saved per job, or revenue per account, and make sure each one is instrumented. Demand for the product also needs its own plan, whether that comes from sales outreach, existing customer migration, or search visibility through SEO services.

The seven stages at a glance

The table below summarises typical effort, cost and the decision each stage is supposed to produce. Treat the ranges as planning figures for a small or mid sized Canadian company, not quotes.

Stage	Typical duration	Typical cost (CAD)	Decision at the gate
1. Idea and problem framing	2 to 5 days	Internal time	Is the problem specific and worth solving
2. Discovery and feasibility	1 to 3 weeks	5,000 to 20,000	Will anyone pay, and can it be built
3. Concept and scoping	1 to 2 weeks	5,000 to 15,000	What exactly is version one
4. Design and prototyping	2 to 5 weeks software, longer for hardware	8,000 to 30,000 software, 20,000 to 150,000 hardware	Do real users succeed with it
5. Engineering build	8 to 20 weeks	40,000 to 150,000	Is it production ready and maintainable
6. Testing and compliance	2 to 4 weeks dedicated	10 to 20 percent of build	Is it safe, legal and reliable
7. Launch and iteration	First 90 days	15 to 25 percent of build	Do the numbers justify scaling

Where companies lose money in the process

The expensive failures follow a small number of repeatable patterns, and all of them are avoidable with a single decision made earlier.

• Building version three first. The feature list grows to cover every customer type, the budget triples, and launch slips past the window that made the product relevant.
• No gate authority. If nobody in the room can stop the project, the gates are ceremony and the new product development process stages become a schedule rather than a set of decisions.
• Outsourcing the thinking. A vendor can own the engineering. Only you can own the problem definition and the commercial case.
• Treating launch as the finish line. Products that win are usually on their fourth meaningful revision within a year of launch.
• Hiring a full team for a six month question. Fixed in house capacity forces you to keep building even after the evidence says stop.

How to run this if you are not technical

You do not need to become an engineer. You need to own the problem, the budget and the gate decisions, and you need a technical partner who will tell you when an answer is no. Practically, that means asking for a written scope before a build, a named decision at the end of every stage, and a demo you can use yourself every two weeks.

A good partner should be able to move with you across all seven stages instead of handing you off. That is how Prototype Toronto is set up, with three service lines covering prototyping and product engineering from software through to deep-tech hardware, AI development and integration, and digitalisation of the processes you already run. You can see the detail of the build work on our product engineering services page.

Run the new product development process stages in order and start with the cheapest one you have not completed honestly. For most companies that is stage two, a short discovery sprint that tells you within three weeks whether to commit real money. If you have an idea and want a direct read on feasibility, cost and timeline before you spend anything, book a free consultation and bring whatever you have, even if it is one page.

Frequently Asked Questions

What are the seven new product development process stages?

Idea generation, idea screening, concept development and testing, business case, design and prototyping, pilot or test marketing, and commercial launch. Each stage ends in a go or no-go decision, so weak ideas stop early and the budget concentrates on the ideas that survive review.

How long does it take to move a product from idea to launch?

For a straightforward software or light hardware product, six to twelve months is realistic. Hardware needing tooling, certification or regulated components often runs eighteen months to three years. The early stages are fast and cheap. Prototyping, testing and manufacturing setup consume most of the calendar, so build the schedule around those.

What does each stage actually cost?

Costs climb sharply as you move through the new product development process stages. Screening and concept testing are mostly staff time. A working prototype runs from a few thousand dollars for a simple proof of concept into six figures for connected hardware. Budget the largest amounts for tooling, certification and the first production run.

Can we skip stages to launch faster?

You can compress them, but skipping the business case or the pilot is where most failures start. Teams that skip validation usually pay for it later in redesign, recalls or unsold inventory, which costs far more than the stage they avoided. Run the stage quickly on a smaller sample instead of dropping it.

How do I decide whether to continue or stop after each stage?

Set the criteria before the stage starts, not after the results arrive. Common gates are evidence that customers will pay the target price, a technical path your team can actually build, a margin that holds at realistic volumes, and a launch date that still fits the market window. Failing a gate means stop or rework.