You Funded. Now the Real Risk Begins.

The press covers raises. It does not cover tooling rework six months later. So the narrative arc the public absorbs is: idea, campaign, success, product. The gap between "success" and "product" is where most of the actual work lives, and it is almost entirely invisible.

Crowdfunding platforms surface campaigns partly based on momentum, which means a successful campaign rewards the ability to tell a compelling story and generate early social proof. Those are real skills. They are not, however, the skills required to harden a design for volume production using ABS+PC injection molding with P20 tooling, negotiate supplier terms for 10,000 unit MOQs, or navigate FCC Part 15 regulatory testing. You can be world-class at one and completely unprepared for the other.

What the raise actually does is hand you obligations: to hundreds or thousands of backers who now expect a physical object, on a timeline you committed to publicly, built from a design that has usually only ever existed as a prototype or a render. We have seen this tension at close range, on campaigns that raised genuine capital and still had to rebuild entire subsystems after the close because the prototype had never been stress-tested against what a contract manufacturer actually needs. For example, a campaign raising $300K for a complex IoT device might find that their initial PCB design needs a complete overhaul to meet EMC standards, incurring an additional $20K-40K cost and 8-12 weeks delay. The raise didn't cause those problems. It just made them expensive.

1. DFM: The Prototype Lies to You

Design for Manufacturing (DFM) is the critical discipline of adapting a design so that it can be produced reliably, at volume, by actual machines, operated by actual humans, with actual tolerances. A prototype, especially one produced by a skilled hands-on builder, does not test this. It tests whether the concept is physically coherent. That is a different question.

The classic failure mode: a founder hands a beautifully crafted prototype to a contract manufacturer, who comes back with a list of changes required before the part can be molded or stamped or assembled at scale. The wall thickness is wrong for injection molding (e.g., less than 1.5mm in a critical area). The undercuts prevent the mold from releasing, requiring a redesign. The snap-fit geometry that worked in one hand-built unit will fail at rate due to material flow variation with PC/ABS V0. The surface finish requires a secondary operation nobody costed (pad printing or vapor polishing).

Each of those changes costs time and money. More importantly, each of them can cascade: changing a wall thickness changes the thermal behavior, which changes the warp, which changes the fit with an adjacent part, which invalidates the tooling for that part. A single DFM gap does not stay contained.

The cost difference between catching a DFM problem at the design stage versus after tooling is cut is not marginal. A design revision before tooling might cost a few hundred dollars in engineering hours. The same fix after an injection mold has been cut can run anywhere from several thousand to over fifty thousand dollars especially for steel molds or complex multi-cavity tools, depending on the geometry and the tool complexity. That is not a hypothetical range: it is the bracket that experienced tooling engineers cite when they explain why DFM review is not optional. A medium-complexity injection mold typically costs between $15,000 and $80,000 to cut. That is the asset you are putting at risk when DFM review is skipped or deferred until after the campaign closes. For insights into preparing designs for mass production, consider our manufacturing strategy services.

The honest version of DFM review happens before tooling is cut, before the campaign launches, before the render is the centerpiece of the pitch. A design that hasn't been through it is not a product. It is a hypothesis wearing a product's clothing.

2. Supplier MOQ: The Economics That Weren't in the Plan

A common challenge in hardware campaigns arises when founders encounter Minimum Order Quantities (MOQs), which exist because suppliers have setup costs and need a minimum run to amortize them. This is completely rational on their side. It is often a genuine shock on the founder side.

Say your campaign funded at a volume that felt like a lot: several hundred units. Your injection molder quotes tooling at a cost that assumes thousands of units to make sense over time, perhaps an MOQ of 5,000 units for a particular part, with P20 tooling costing $50,000-$100,000 with 8-12 weeks lead time. Your PCB assembly (PCBA) house has a minimum board run that may be two to three times the number of units you actually need for your first production batch (e.g., 2,000 unit MOQ for your flex PCB). Your specific color of polycarbonate has a minimum pellet order that commits you to more raw material than your first production run requires, forcing a 25kg bag purchase when only 5kg is needed. Each of these independently generates either a large unexpected cost or a negotiation that delays the build.

Now compound them. A typical consumer hardware product involves somewhere between six and fifteen distinct vendor relationships, each contributing to the Bill of Materials (BOM). If three of them hit you with MOQ realities that weren't in your original cost model, you are either burning contingency reserve (if you had any), cutting specifications, or slipping the timeline. Most founders budget for one of those. Reality delivers all three simultaneously.

The unit economics that looked solid at two hundred units can deteriorate by 30% to 50% once real supplier minimums are applied across a full bill of materials. That is not a disaster figure invented for effect: it is the range that founders encounter when they move from an internal cost model to actual vendor quotes for the first time. The most common mistake here is underestimating component lead times, especially for custom parts, which can be 16-24 weeks or more. For help proactively addressing these issues, robust mechanical engineering and electronics design are essential.

This is the part nobody puts in the campaign video.

3. Tolerance Stack-Up: Small Numbers, Big Failures

Another frequent cause of late-stage rework in hardware development is unmanaged tolerance stack-up, an issue stemming from the fact that every manufactured part has dimensional tolerances: the actual part will not be exactly the nominal dimension, but within some specified range. Tolerance stack-up is what happens when you assemble multiple parts and their individual variations combine in the worst case. A chain of six components, each held to a tolerance of a few tenths of a millimeter (e.g., ±0.15mm), can produce a worst-case assembly error that may exceed one millimeter in total. That may be enough to prevent a housing from closing, a seal from seating, or a battery door from latching, all of which generate visible failures at inspection.

This is a calculation, and it must be done deliberately. It requires knowing the actual tolerances your specific suppliers and processes will hold, not the nominal tolerances quoted in a spec sheet (which can be optimistic by 20-30%). There is a difference, and it is not small. Injection-molded parts in standard ABS or polycarbonate typically hold tolerances in the range of ±0.1 to 0.3 millimeters depending on geometry and wall thickness. Over a five-part assembly, the worst-case stack can easily reach one to two millimeters, which is enough to fail a snap-fit or gap a seal.

Founders who skipped this analysis before locking tooling discover it when the first production samples from EVT (Engineering Validation Test) or DVT (Design Validation Test) come back and the assembly line can't hit acceptable yield, often below 90%. The costly oversight here is assuming standard COTS (Commercial Off-The-Shelf) parts will always mate perfectly without considering manufacturing variance, or that a prototype built with SLA (Stereolithography) or FDM (Fused Deposition Modeling) will reflect production tolerances. Fixing it at that stage means either reopening tooling, which is expensive and slow, or writing an engineering change that propagates through the design. Neither is free, and neither is fast. We have had this exact conversation with founders mid-campaign: the news is never welcome, and it is always better than the alternative of finding out after thousands of units are committed.

4. Regulatory Testing: The Budget Line Nobody Put In

A critical oversight that derails many hardware campaigns is the lack of a budget and timeline for mandatory regulatory testing. A consumer hardware product sold in most major markets needs certifications. In North America, you are likely looking at FCC Part 15 and IC markings for anything with wireless or digital electronics, UL or equivalent for electrical safety, possibly ASTM F963 or CPSC compliance depending on the product category. In Europe, CE marking covers a family of directives that may include EMC Directive (2214/30/EU), Low Voltage Directive (LVD), and RoHS, among others. If your product has a rechargeable battery, that adds UN 38.3 transport testing. If it is medical-adjacent, the list extends considerably.

Each of these requires testing at an accredited lab. A single test round at an accredited EMC facility typically runs between $5,000 and $20,000 depending on product complexity and the number of frequency bands tested, and that is before any retesting. For a product with Bluetooth or Wi-Fi, FCC Part 15 testing alone can take 3-6 weeks at a busy lab. If you fail a round, you revise the design and retest. Founders who budgeted for one pass and needed three find the regulatory phase destroying both their margin and their timeline simultaneously. The most common mistake is assuming "it will pass" or that a simple pre-scan is sufficient for final certification. Moreover, failing to account for the iteration time (e.g., 2-4 weeks per design revision) between test rounds is a major timeline killer. For example, a campaign raising $150K for a connected device might budget $10K for regulatory, but find themselves needing $30K-40K and 12-16 weeks instead.

The particularly painful version: you discover a regulatory issue after tooling is cut and first samples are in hand. The fix requires a PCB revision. You are now re-qualifying the board, re-testing, and re-validating the mechanical fit of any components that changed. The timeline you told backers was six months is now eleven. The polite update email is the least of your problems.

Here is the core point, and it is worth stating plainly: the money doesn't create these problems. It amplifies them by attaching real obligations to them.

A founder with a DFM-hardened design (using appropriate materials like Nylon 6/6 for structural components and considering gate locations for optimal mold flow), a supplier network that has been tested at small volume (e.g., a 500-unit pilot run), a tolerance analysis done before tooling (e.g., using GD&T for critical features), and a regulatory budget built into the cost model (accounting for UL 60950 and IEC 62368 standards): for that founder, the raise is genuinely an accelerant. It buys time and capacity to execute a plan that was already sound.

A founder with a beautiful prototype, a compelling story, and none of that infrastructure: the raise is a pressure multiplier applied to a structure that isn't ready for pressure. The clock starts running the moment the campaign closes. Backers don't care about the problems. They care about the date.

The concept was never the hard part. It was never even the interesting part. What the campaign funds is the thing that was already built before the campaign launched, or it funds a painful education in what "building a product" actually means. Those are different campaigns. Most people can't tell them apart from the outside, and some founders can't either.

The practical implication isn't "don't raise money." It's "don't confuse a campaign milestone with an engineering milestone."

A product that has been through serious DFM review, whose Bill of Materials (BOM) has been stress-tested against real supplier quotes (e.g., targeting a cost-per-part below $0.80 for a high-volume component), whose tolerance stack has been analyzed (considering ±0.2mm for critical fits), and whose regulatory pathway is understood (with confirmed CE and FCC compliance strategies): that product is ready to be funded. The raise then does what it is supposed to do.

A product that exists as a prototype and a story is not unraisable, but whoever funds it is funding the R&D, not the production run. That is a different contract, and the honest version of the campaign makes it explicit. Most don't. The campaign presents the prototype as representative of the final product. The backers believe them. The gap between those two things is where the late shipments live.

The founders who ship on time are not the ones who raised the most. They are the ones who resolved the execution problems before the money arrived. The raise is the reward for having done the work, not the starting pistol for doing it. If you want to understand what that preparation actually looks like in practice, the execution side of hardware development is worth reading before you open a campaign page.