Why This Document Saves You Money
Engineering firms quote against uncertainty. When the only input is a two-paragraph email saying "I have an idea for a smart water bottle," the firm has to price the risk that your idea is three times more complex than it sounds. That risk shows up as padding, as a large discovery phase billed hourly, or as a refusal to quote at all until you pay for a scoping study.
A clear idea brief removes most of that uncertainty. In practice it changes the first quote by 20 to 40 percent and turns a three-week back-and-forth into one call. It also filters: a firm that reads a good brief and still gives you a vague answer has told you something useful.
Two to four pages is right. Nobody reads twelve, and one page usually means you have not thought it through.
A Structure That Works
1. One-sentence summary. What it is, who it is for, what it does. "A countertop device that portions and dispenses dry pet food on a schedule, for owners of two or more cats." If you cannot write this sentence, the concept is not ready.
2. The problem. Who has it, how often, and what they currently do instead. Be concrete about the current alternative, because the alternative is your real competition. This is worth working through properly before writing anything else, using the discipline in validating a product idea.
3. How you imagine it working. The user's sequence of actions from unboxing to daily use. Where it sits, what it plugs into, who cleans it, what happens when it runs out. Sketches belong here, and they do not need to be good; sketching with no technical skills covers what a useful rough drawing contains.
4. Must-haves versus nice-to-haves. Two explicit lists. This is the single most valuable section, because it lets an engineer propose a cheaper architecture without feeling like they are cutting your product. If everything is a must-have, you will get one expensive option.
5. Known constraints. Size envelope, weight ceiling, target retail price, power source, operating environment, any regulatory context you know of, and languages or markets. Target price matters more than anything else in this list, because it determines whether the answer is injection molding or machining.
6. What exists already. Competing products with links, what you like and dislike about each, and any patents you have found. Saying "it is like product X but with Y" is efficient and honest, not embarrassing.
7. Your situation. Budget range, timeline, whether you have funding, what you have already built or spent, and what you can do yourself. Firms scope differently for someone with a working breadboard than for someone with a napkin.
8. What you are asking for. A quote for the whole program, a feasibility opinion, or one phase. Be specific: many projects sensibly start by hiring a firm for a single phase rather than committing to everything.
Describing It Without an Engineering Background
You do not need to specify components, and you should not try. Engineers translate intent into parts; that is the job you are paying for. What they need from you is function and behavior.
Say what it must do, not how: "it needs to hold cold water for eight hours in a hot car," not "it needs vacuum insulation with a 0.5 mm stainless wall." Use comparisons to things that exist: "about the size of a paperback," "quieter than a laptop fan," "the button feel of a car key fob." Use numbers wherever you honestly can, and flag guesses as guesses: "needs to run about a week per charge, though I have not tested that assumption."
Photographs of a rough physical mockup, even cardboard and tape, communicate more in one image than three paragraphs. So does a 30-second phone video of you miming the interaction.
Common Mistakes
- Specifying the solution instead of the need. "It uses a Raspberry Pi and a stepper motor" locks the engineer out of the cheaper answer. State the requirement and let them choose.
- No price target. The most common omission and the most expensive one. Without it, the design lands somewhere arbitrary and gets value-engineered later at full cost.
- Everything in version one. App, cloud dashboard, three colorways, and a subscription model on a product that has never been prototyped. If your list looks like this, cut it down to a first version before you send it.
- Hiding the budget. Withholding your number to "see what they say" wastes a round trip. A range is enough: firms need to know whether they are proposing a $40,000 program or a $400,000 one.
- Confusing the brief with a spec. The brief describes intent; the requirements document that follows it is a joint deliverable, and who writes the requirements spec explains why you should not attempt it alone.
- Excessive secrecy. Refusing to describe the invention until an NDA is signed, then describing it vaguely anyway. Reputable firms sign NDAs routinely, and the practical limits are in NDAs for inventors. If the concept is genuinely patentable, file a provisional first and then talk freely.
What Happens After You Send It
Expect a reply within a few business days: either clarifying questions, a call invitation, or a rough range with caveats. Good questions are a positive signal. A firm that quotes a precise number off a two-page brief with no questions is guessing, and the number will move.
The call that follows typically runs 45 to 90 minutes and covers feasibility concerns, the phases they would propose, and what they would need before firming the number. Send the same brief to three firms and compare, keeping in mind that quotes are rarely comparable line for line. Ask each one what they think the hardest part is; the answers tell you who actually read it.
Once you engage, the brief becomes the input to a formal product requirements document, written jointly, which is what the engineering is actually verified against. The brief is not wasted work; it is the first draft of that document.
Send Us Your Brief
Projects House reads every idea brief that comes in and answers with real feasibility questions, a phase breakdown, and a range you can plan against. Two pages is plenty. Send it through our contact form.