I Have a Business Idea. What Do I Do Next?

There is a particular kind of restlessness that comes with a good idea. You think about it on the commute. You explain it to a friend and they say "actually, that could work." And then nothing happens for six months, because the gap between having the idea and doing something about it feels enormous.

It is not. The gap is about eight specific steps, most of which cost nothing. This guide walks through them in order: how to validate a business idea before you spend money, how to scope a minimum viable product, what MVP development costs in 2026, and how to brief the people who will build it.

I have a business idea. What do I do next?

Before you build anything, prove that someone has the problem, is currently paying to solve it badly, and will tell you so in a conversation. That takes two to four weeks and costs almost nothing. Building comes after that, not before.

Almost every failed first venture we hear about skipped this. The founder was not lazy or unintelligent. They were enthusiastic, and enthusiasm feels like evidence. It is not. Evidence is a stranger describing your problem back to you without being prompted.

How do I know if my business idea is worth building?

An idea is worth building when three things line up: a specific group of people have a recurring problem, they already spend time or money working around it, and you can reach them without an enormous marketing budget. Miss any one of those and the build is a gamble.

Score your idea honestly against these five questions.

1. Who exactly has this problem? If your answer is "everyone" or "small businesses", you do not have an answer yet. "Independent physiotherapy clinics with two to five staff" is an answer.
2. What are they doing today instead? A spreadsheet, a WhatsApp group, or a bad competitor product all count as proof the problem is real. "Nothing" is a warning sign, not a green field.
3. What does the problem cost them? In hours, in lost revenue, or in stress. If nobody can quantify it, nobody will pay to fix it.
4. Can you reach fifty of these people this month? Through a community, an association, a directory, or your own network. If you cannot, distribution is your real problem, not product.
5. Why you? Unfair advantage can be domain experience, an audience, or access. It does not have to be technical.

How do I validate a business idea without spending money?

Run cheap experiments that ask people to give up something real: time, an email address, a waitlist deposit, or a signature on a letter of intent. Opinions are free, so they are worthless. Small commitments are the cheapest reliable signal you can buy.

The four experiments worth running first

• Fifteen problem interviews. Ask about their last week, not your idea. "Walk me through how you handled X last time" tells you more than "would you use this?" ever will. Never pitch during these calls.
• A one page offer. Headline, problem, promise, price, and a button. Send it to your list or a relevant community and count how many people ask for details.
• The manual version. Deliver the outcome by hand for three to five customers, with no software at all. It is slow, it does not scale, and it is the single most useful thing you can do before writing code.
• A pre sale or waitlist deposit. Even a small refundable amount changes the conversation entirely. People who pay tell you the truth.

A 30 day validation plan

Days	Focus	What good looks like
1 to 7	Define the customer and the problem in one sentence each	You can name the segment and where they gather online
8 to 16	Fifteen problem interviews	At least eight describe the same pain unprompted
17 to 23	One page offer plus outreach	Five or more people ask a buying question
24 to 30	Deliver manually for two or three of them	Someone pays, or clearly commits to paying

If day 30 arrives and nobody paid, that is not failure. That is a saved budget and a sharper idea. Change the segment or the promise and run it again.

What is an MVP and how small should mine be?

A minimum viable product is the smallest thing you can put in front of a real customer that delivers the core outcome and teaches you something. It is not a cheap version of your full product. It is one complete journey, done well, with everything optional stripped out.

The exercise that makes scoping easy: write the single sentence that describes the value moment. "A clinic owner sees tomorrow's no show risk before they leave today." Now list every screen and step needed to reach that moment. That list is your MVP. Anything not on it is version two, including admin dashboards, settings pages, and the reporting suite you are already imagining.

Common things founders wrongly consider essential in an MVP: user roles and permissions, multi language support, a mobile app when a mobile web page would do, in app chat, custom analytics, and anything described as "we will need it later at scale". You will not be at scale. That is the point.

How much does MVP development cost in 2026?

A no code or low code MVP typically runs $3,000 to $12,000 and takes three to six weeks. A custom coded MVP with a real backend, payments, and authentication generally sits between $15,000 and $40,000 over eight to sixteen weeks. Complex, regulated, or AI heavy products start higher.

What actually drives the number:

• Number of distinct user journeys. One journey is cheap. Three user types each with their own screens is three products.
• Integrations. Every third party system you connect to adds testing, edge cases, and error handling.
• Compliance. Health, finance, and anything touching sensitive records adds real cost, and skipping it is not an option.
• Design ambition. A clean functional interface is fast. A bespoke animated experience is not.

You can get a specific range for your own idea in a few minutes with our cost estimator, or the app cost analyzer if you are thinking mobile first. Both give you a phase split rather than a single scary number.

Should I protect my idea before talking to anyone?

Protect the execution, not the concept. Ideas are far more common than the ability to ship them, and the people you need feedback from are rarely the people who would steal it. Keep a written record, use a straightforward mutual non disclosure agreement with suppliers and contractors, and make sure any development contract assigns intellectual property and code ownership to you in writing.

What genuinely matters legally, in order: your company registration, IP assignment in every contractor agreement, ownership of domains and accounts in your own name, and a trademark on the name once you are trading. A patent is rarely the right first spend for a software idea.

How do I brief a development partner properly?

Give them the problem, the customer, the value moment, your budget range, and your deadline. Do not give them a feature list and ask for a price. A good partner will design a cheaper solution than the one you imagined, but only if you tell them what you are actually trying to achieve.

A brief that gets useful quotes contains:

1. One paragraph on who the customer is and what they do today.
2. The single value moment your product delivers.
3. Your evidence so far, including interviews, pre sales, or manual deliveries.
4. A budget range. Withholding it wastes everyone's time and usually costs you more.
5. Your real deadline and what depends on it.
6. What you already have, such as branding, a domain, or an existing customer list.

Send the same brief to three teams. The quotes will differ, and the differences will teach you more about your own scope than the numbers will.

Frequently asked questions

How long does it take to go from idea to launch?

Four to six months is realistic for a focused first version, split roughly into one month of validation, two to three months of building, and a few weeks of testing and launch preparation. Shorter is possible with no code tools. Much longer usually means the scope was never trimmed.

Do I need a technical co-founder to start?

No, but you need a technical decision maker you trust, whether that is a co-founder, a fractional CTO, or a development partner who will tell you no. What you cannot do is outsource judgment entirely and hope the invoices work out.

Can I build my idea with no code tools first?

Often yes, and it is usually the smart first move for workflow, marketplace, and booking style products. You will outgrow it if your value depends on custom logic, heavy data, or performance, so treat it as a validation instrument rather than a permanent home.

What if someone is already doing what I want to do?

That is usually good news. It means the market exists and someone else has paid to educate it. Look at their reviews, find the complaints that repeat, and build for the segment they are underserving rather than trying to beat them on features.

How much money should I have before starting?

Enough to run validation, which is mostly your time, plus the cost of a first build with a reserve of around 30 percent for the changes you will inevitably want after real users touch it. Starting a build with exactly the quoted amount and nothing spare is the most common way projects stall halfway.

Your next step

If your idea passed the five questions at the top of this guide, do the thirty day validation plan before anything else. If you have already done it and people are asking when they can buy, you are ready to scope a build.

Once you are trading and wondering what to invest in next, what a startup should actually build first picks up exactly where this guide ends. And if you would rather just talk it through, send us the idea. We will tell you honestly whether it needs software yet, and what the smallest useful version would cost to build.