Every inventor has heard “validate before you build.” Almost nobody explains what that means for a physical product.
Product feasibility analysis answers one question: can this be built, certified and sold at a profit — and what would have to be true for that to work? It happens before design, on paper and in spreadsheets, and it ends with a recommendation you can act on.
The reason it matters is timing. Almost everything that decides whether your product succeeds gets locked in early — the architecture, the components, the enclosure, the certification path. Once tooling is cut those decisions stop being cheap. Feasibility is where you buy down risk while it’s still measured in engineering hours instead of steel.
This guide covers what feasibility actually examines, the physics that no budget overcomes, why certification decides your architecture rather than following it, and how to price backwards from a shelf price to a parts budget.
What a product feasibility study actually covers
Six dimensions, each answering a different question and producing a different deliverable.
Technical feasibility. Can it physically work? Produces an architecture assessment, a power budget, and a list of the assumptions you’re betting on.
Manufacturing feasibility. Can it be made repeatedly? Produces a process assessment and the design constraints your enclosure will have to obey.
Regulatory feasibility. What certifications apply, what do they cost, how long do they take? Produces a compliance route with test costs and a timeline.
Economic feasibility. Can it be sold at a profit? Produces a bill of materials cost model at your target production volume.
Supply feasibility. Will the parts exist when you need them? Produces a component and lifecycle analysis.
Schedule feasibility. How long before you can ship? Produces a phased roadmap with the long-lead items identified.
A good feasibility study ends in a written engineering recommendation and a risk register — not a report that says “it’s promising.” As one UK electronics consultancy frames the central question: can the product be built the way you want it built?
Some studies add a proof-of-concept build, but only for the riskiest assumption — for instance, a battery-powered radio module with a real antenna, deployed in the environment it will actually live in, to verify range and signal quality. That’s not a prototype. It’s an experiment with one variable.
Technical feasibility: what physics will and won’t allow
The useful distinction here isn’t “possible versus impossible.” It’s physically impossible versus expensive but possible, and founders routinely misclassify.
Power budget. Battery life isn’t a specification you choose, it’s an output you calculate. Weight the current draw by time spent in each state — active, sleep, deep sleep, shutdown — to get an average, then divide capacity by that average.
Real numbers for a Bluetooth device: sleep with the real-time clock running is around 1.5 µA, deeper sleep without it under 200 nA, and peak transmit 3 to 6 mA on current parts. Older silicon can draw three times that or more. And datasheet sleep figures alone will mislead you, because wake-up time and pre-transmission processing are hidden power costs that don’t appear in the headline number.
Battery energy density is a hard ceiling. Commercial lithium-ion sits at roughly 150 to 250 Wh per kilogram and 300 to 700 Wh per litre. Lithium polymer is lower. Solid-state and lithium-sulfur promise more and are not commercially available.
That gives you a test no amount of money passes. Multiply the internal volume you can spare by about 500 Wh/L. If your power budget needs more than that, the battery isn’t the problem — the product is. It has to get bigger, do less, or do it less often.
Radios need physical space, and this is physics, not preference. A 2.4 GHz PCB antenna needs a ground plane extending at least 31 mm — a quarter wavelength — from the feed point, with no copper, pours or vias on any layer beneath the antenna element. A board narrower than that in the radiating direction forces efficiency compromises you cannot design around.
This is the classic feasibility collision: the industrial designer wants a tiny sealed puck, and the battery is exactly where the antenna keepout has to be. Better to discover that on paper.
Sensors have accuracy limits that marketing cannot move. Wrist-worn optical heart rate against ECG shows bias ranging from about −1.3 bpm on the best consumer devices to −9.3 bpm on others, and absolute error during activity runs around 30% higher than at rest. Accuracy depends more on wrist movement, fit and device position than on exercise intensity. There’s no universal acceptable-error threshold — it depends entirely on what you’re claiming.
For the clearest example of a technical claim that hasn’t survived contact with reality: as of a February 2024 safety communication, the FDA has not authorized, cleared or approved any smartwatch or smart ring that measures blood glucose on its own, regardless of brand. Enormous capital has been spent trying. If your concept depends on a sensing capability nobody has shipped, that’s your riskiest assumption and it belongs in a proof of concept before anything else.
Component availability is the invisible killer. Consumer-grade integrated circuits now average about 4.7 years of production life, industrial parts 6.2 years — down from over a decade. When a part goes end-of-life you typically get a discontinuation notice, then a last-time-buy window of around six months, with total time from announcement to obsolescence in the 12 to 18 month range. Around 40% of components reach end of life earlier than expected. For startup-scale buyers, lead times of 6 to 18 months are common, and you queue behind customers buying millions a year.
Manufacturing constraints your CAD model probably ignores. Injection molding imposes wall thickness limits by material — ABS roughly 1.14 to 3.56 mm, polycarbonate 1.02 to 3.81 mm — plus draft angles of 1 to 2° for a smooth finish, 3° for light texture, 5° or more for heavy texture. Ribs should be 0.5 to 0.6 times the nominal wall. A model that ignores these isn’t a design yet, and making it manufacturable changes the shape of the product.
Regulatory feasibility: why certification decides your architecture
Founders treat certification as paperwork at the end. It isn’t. What your product is and does determines which certifications apply, and the answer changes the design.
If it has a radio. FCC authorization costs vary enormously by path: an unintentional radiator around $1,500 to $5,000 over two to four weeks; an intentional radiator using a pre-certified module $3,000 to $10,000; custom RF on a single band $8,000 to $20,000 over six to twelve weeks; cellular anywhere from $50,000 to $200,000 and up over six to nine months.
That gap between using a certified module and designing custom RF is an architecture decision made during feasibility — not a testing decision made later. With a module, the vendor’s grant covers the intentional emissions and you only test the host.
Add retest after a failure at $2,000 to $10,000, and a board respin at $1,000 to $15,000, and pre-compliance testing at $3,000 to $8,000 starts looking like insurance rather than expense.
In Europe, radios got harder. The Radio Equipment Directive’s cybersecurity requirements became mandatory in August 2025, with EN 18031 as the harmonized standard covering network protection, personal data safeguards and fraud prevention. The harmonization carries restrictions around password handling and access control, meaning conformity isn’t automatic and a Notified Body may be involved. This is a requirement that changes your firmware and account architecture, not your paperwork.
If it has a battery. UN/DOT 38.3 runs $5,000 to $7,000 over four to six weeks and consumes 16 packs. IEC 62133 runs $6,000 to $10,000 and consumes 33 sealed packs. UL 2054 or 1642 runs $15,000 to $20,000 over ten to twelve weeks and consumes 52 battery packs. At prototype pricing, the packs alone can cost more than the test.
If it touches skin. ISO 10993 biocompatibility commonly evaluates cytotoxicity, sensitization and irritation — though which endpoints apply depends on contact duration and the specific hazards. The feasibility consequence is bigger than the test: changing your pigment or masterbatch, cure cycle, mold release agent, cleaning process, or your molder can trigger a documented biological risk reassessment. Your material and your factory become part of your compliance file, so switching suppliers stops being a purchasing decision.
If it makes health claims. Under the FDA’s January 2026 general wellness guidance, the same sensor hardware is either an unregulated wellness product or a regulated device depending on what you claim. Outputting a reading in a fitness context and suggesting the user consult a professional is fine. Linking that reading to a diagnosis, using clinical thresholds, or offering ongoing monitoring and alerts for medical management crosses the line. The sentence on your box is an engineering decision.
If it’s for children. Products for ages 0 to 12 need a Children’s Product Certificate backed by testing at a CPSC-accepted lab, covering lead content, phthalates and ASTM F963 for toys, plus per-unit tracking labels. Cost scales with the number of distinct materials and colors — a five-color product is a five-times-tested product. That’s a feasibility finding that changes your product line, not your design.
Substances, for anything sold in Europe. RoHS restricts ten substances measured per homogeneous material — a plating layer, cable insulation, a solder joint — at 0.1% by weight, cadmium at 0.01%. A single small component can fail compliance even when overall concentrations look fine. Documentation has to be retained for ten years. REACH runs in parallel with its own obligations.
Safety marks. UL isn’t legally required for most consumer products, but retailers demand it, which makes it practically required. Initial certification runs roughly $15,000 to $30,000 for a simple consumer device and $30,000 to $70,000 for mid-complexity electronics, over eight to twelve weeks at best and twenty-plus if redesign is needed. Then there’s ongoing maintenance and $2,000 to $10,000 per design change. Certification turns your design into a frozen asset with a change tax.
And the cheap one people miss. The EU’s General Product Safety Regulation requires a EU-based responsible person, traceability records, and labeling in the language of the country of sale. It doesn’t touch your electronics at all. It changes your packaging artwork and your label silkscreen — both of which get printed in bulk.
Economic feasibility: pricing backwards from the shelf
Most founders price forwards — add up the parts, add a margin, call that the price. That’s backwards.
Start at the shelf and work down. A worked example from a hardware manufacturing platform: $25 cost of goods sold, $50 net price to the retailer, $79.99 on the shelf. That’s a 50% gross margin for you and about 37% for the retailer. Consumer electronics companies generally need a gross margin above 50% to be healthy, and closer to 70% to self-fund growth.
Channel margins decide a lot of this. Online retail typically takes 15 to 20%, specialty retail 30 to 35%, big box can exceed 40%. Add product liability insurance, in-store display costs, and payment terms that are nominally 90 days and in practice much longer.
Here’s a genuine disagreement worth knowing about, because it’s how founders end up with an unsellable product. One widely cited rule says retail should be 2 to 4 times your COGS, ideally at least 3. Another says 2.5 to 4 times your bill of materials. Those are not the same rule. COGS includes assembly labor, test, packaging, freight, duty, scrap and warranty reserve on top of the BOM. A “4× multiplier” applied to the wrong denominator is the difference between a healthy margin and no margin at all.
The costs that never appear on a bill of materials. Tooling is the big one: aluminum tooling $1,500 to $8,000 with a two to three week lead time, suitable below about 10,000 units a year; pre-hardened steel $8,000 to $25,000; hardened multi-cavity steel $60,000 to $150,000 and up with a ten to fourteen week lead time. Then test fixtures, programming jigs, packaging, certification, and a scrap rate that starts around 5% and takes a few production cycles to come down.
And the one that surprises connected-product founders most: for an electronic product with an app, application software commonly runs two to four times the hardware development budget.
The mechanism behind most cost surprises isn’t bad arithmetic. It’s timing — waiting until the product is fully developed before calculating what it costs to manufacture, at which point the answer is fixed.
What a feasibility study costs, and what it saves
Very few firms publish a price, which is itself worth knowing — this work is almost always quoted per project. The one public anchor for an electronic product is a UK consultancy that lists feasibility at roughly £2,500 to £15,000, covering technical risk, component costs and radio range questions. Duration runs from a couple of weeks to a couple of months for a paper study, longer if a proof of concept gets built.
Judge that against what a single avoidable mistake costs:
Change the schematic during feasibility, and you’ve spent engineering time. Change it after PCB layout, and you’re paying $1,000 to $15,000 for a respin. Change it after FCC filing, and add $2,000 to $10,000 to retest. Change the enclosure after tooling, and you’re buying a new tool at $25,000 to $60,000 and waiting six to eight weeks. Change a UL-listed construction after listing, and it’s $2,000 to $10,000 per change — during which the listing isn’t valid.
You’ll also see a widely repeated claim that 80% of a product’s cost is locked in during design. Worth knowing that the figure has been challenged in the academic literature and its popular sources cite nothing. The direction isn’t in dispute — early decisions constrain later cost — but treat the specific number as folklore rather than data.
Can you skip feasibility analysis?
Being honest about this matters more than selling the service.
Don’t skip it because the product looks simple. Inventornest always prefers to begin with feasibility, whether the idea is simple or complex. Even familiar products can contain uncertain requirements, integration challenges or component limitations that become expensive if discovered late.
Validate demand separately from physics. If you don’t yet know whether anyone wants this, a landing page test, a small pilot or genuine pre-orders will teach you more than any analysis. Feasibility answers “can we build it,” not “should we.”
Scale it to the decision. A feasibility study can be lean and focused rather than a long report. The rule that matters: concentrate on genuine risks rather than well-understood patterns. If a question has a known answer, don’t pay to have it researched.
The counterweight, though: feasibility matters most when you’re outsourcing the design work, because external teams tend to prioritize producing a prototype over assessing risk. A prototype makes everyone feel good. It’s also exactly what a feasibility study is supposed to tell you whether you should build.
How Inventornest runs feasibility analysis
We do the paper work before the build work, because that’s the only point where the answers are still cheap.
That means modeling the power budget against a real duty cycle rather than a datasheet headline. Checking component lifecycle and lead time before a part gets designed in. Establishing the certification route early enough that the module-versus-custom-radio decision is made deliberately.
Where an assumption is genuinely unproven, we build the smallest possible experiment to test it — one variable, not a product.
We’ll also tell you when your idea has a physics problem rather than a budget problem. That’s an uncomfortable conversation and a much cheaper one than the alternative.
If you have an idea and want to know what would have to be true to build it, book a consultation. If feasibility is settled and you’re deciding what to build next, read the difference between a prototype and an MVP. And when you’re ready for a factory conversation, here’s how to find the right manufacturer for a hardware product.
Product feasibility analysis: frequently asked questions
What is a product feasibility study?
An assessment, done before design work, of whether a product can be built, certified and sold profitably. For hardware it covers technical, manufacturing, regulatory, economic, supply and schedule feasibility, and it produces a written engineering recommendation with a risk register rather than a general opinion.
How much does a feasibility study cost?
Few firms publish prices because the work is scoped per project. The one public anchor for an electronic product is roughly £2,500 to £15,000. Compare that with a single PCB respin at $1,000 to $15,000, or a replacement injection mold at $25,000 to $60,000.
How long does feasibility analysis take?
A paper study typically runs from a couple of weeks to a couple of months depending on scope. Adding a proof-of-concept build to test a specific risky assumption extends that.
What is the difference between a feasibility study and a prototype?
A feasibility study asks whether the product can be built the way you want it built, and answers on paper and in cost models. A prototype demonstrates that a specific build works. Feasibility tells you whether it’s worth building the prototype, and what the prototype needs to prove.
How do I know if my product idea is technically possible?
Test the physics first. Calculate the power budget against a real duty cycle and check it against the energy your available volume can store — commercial lithium-ion tops out around 300 to 700 Wh per litre. Check whether a radio has room for its ground plane. Check whether the sensor accuracy you need has ever been achieved in your form factor. Those are hard limits. Everything else is a cost question.
When should I think about certification?
During feasibility, before architecture is frozen. Certification requirements change the design — whether you use a pre-certified radio module or custom RF is an eight-thousand-dollar decision at minimum, and choosing wrongly is discovered at testing when changes cost most.
How do I work out what my product should cost to make?
Backwards from the shelf price. Retail price, minus the retailer’s margin, gives your wholesale price. Your target gross margin then gives your cost of goods sold ceiling, and the bill of materials has to fit inside that alongside assembly, test, packaging, freight and scrap. Be careful with rules of thumb — “4× BOM” and “4× COGS” are very different budgets.
Do I need a feasibility study for a simple product?
Yes. Inventornest always prefers to begin with feasibility, whether the idea is simple or complex. Even familiar products can contain uncertain requirements, integration challenges or component limitations that become expensive if discovered late.