Start a Business guide
How to Start a Business
A commercial, evidence-led route to starting a business: find an opportunity, validate demand, price it properly, and get first customers before you overcommit.
The short answer
Starting a business commercially means establishing, with evidence, that someone will pay for what you intend to sell before you invest heavily in building it. That evidence comes from talking to customers, testing a minimum offer, and watching what people actually do, not what they say they might do. Registration and legal setup matter, but they come after you have something worth registering.
- Look for evidence of demand before building anything
- Test your assumptions cheaply before committing money or time
- Price and cost the business before you launch it, not after
- Get a small number of real customers before scaling up
- Treat the first months as a learning exercise, not a victory lap
Where does a sensible business idea come from?
Most workable business ideas come from a problem you have seen up close, either in your own work, your household, or a gap you have noticed in how a market is currently served. That is a more reliable starting point than casting around for something fashionable. If you already understand a problem, you are a long way ahead of someone who has to learn both the problem and the solution from scratch.
It helps to separate the idea from the romance of it. An opportunity is not just "I like the idea of this"; it is a specific gap between what people currently get and what they would prefer, where that gap is big enough that someone will pay to close it. Lots of ideas are interesting without being commercial. The test is not whether it sounds good, it is whether somebody with money would choose it over what they do now.
If you are not sure where to start, the free What Business Should I Start? tool is built to help narrow a wide field of possibilities down to a shortlist worth investigating properly, based on your situation rather than generic lists of "ideas". It is a starting point for direction, not a verdict — the investigation described in this guide still has to happen afterwards.
Who is actually going to buy this, and why?
Before you go further, you need a specific answer to who the customer is — not "small businesses" or "homeowners", but a description precise enough that you could picture ten real examples of them. The more specific the customer, the easier everything else becomes: pricing, marketing, and deciding what to build first.
Alongside that, work out why they would switch from whatever they currently do. Every customer already has a way of solving their problem, even if that way is doing nothing. Your job is to understand what that current option costs them in money, time, or hassle, and whether your alternative is different enough to be worth the disruption of changing supplier or habit.
This is where a lot of plans go wrong: people describe a product in detail and the customer only in passing. Spend real time here. Write down, in plain terms, who buys, what they buy instead today, and why today's option is not good enough for them.
How do you check that demand is real, not assumed?
Validating demand means getting outside your own head and finding out what people actually do, not what you imagine they would do. Conversations with potential customers are the cheapest way to start: ask about the problem, how they currently handle it, and what it costs them — before you mention your idea at all. If people struggle to describe the problem or shrug it off, that is useful information too.
Beyond conversation, look for evidence that already exists: how people search for solutions, what competitors or substitutes are already charging, and whether there is a visible, active market rather than a theoretical one. A market with no visible activity at all is not necessarily empty opportunity — it is often a sign that demand is thinner than it looks.
The point of this stage is to reduce how much you are guessing. You will never remove guesswork entirely, but there is a large difference between an idea that rests on a hunch and one that rests on a dozen real conversations and some visible market activity.
What assumptions do you need to test before committing?
Every business idea rests on assumptions: that people have the problem, that they will pay to fix it, that you can reach them affordably, and that you can deliver profitably. List these assumptions explicitly rather than letting them sit unexamined. The riskiest assumption — the one that, if wrong, kills the idea — should be tested first, not last.
Testing does not need to be elaborate. It can be as simple as offering the service to a handful of people before it is fully built, putting up a page describing the offer and seeing who responds, or asking for a deposit before you deliver anything. The aim is to get a real decision out of a real person — money committed, a booking made, a form filled in — rather than a polite opinion.
Be honest about what counts as evidence. Friends and family saying "that's a great idea" is not evidence. A stranger paying for something, or actively asking when it will be available, is.
What business model fits what you are selling?
The business model is how you package and sell what you do: one-off jobs, a subscription, a retainer, a marketplace taking a cut, products sold directly, or some mixture. The right model depends on how the customer buys and how often they need what you offer, not on which model you find most appealing. A service people need once a year suits a different model to one they need every week.
Think as well about how the work gets delivered — on your own time, through others you employ or subcontract, or through a product that scales without your direct involvement each time. Each route has different implications for how much of your time a given level of income requires, and that in turn affects what growth actually looks like for you.
There is no single right model, and switching later is often possible, but choosing with the customer's buying behaviour in mind from the outset avoids a lot of wasted effort.
What will this actually cost to run?
Before setting a price, you need an honest view of what it costs to deliver what you sell, including the costs that are easy to overlook: your own time, materials, software, insurance, transport, and any fees taken by a platform or payment provider. A business that looks profitable on revenue alone can be the opposite once true delivery cost is accounted for.
It helps to separate costs that scale with each sale (direct cost per unit or per job) from costs that exist regardless of how much you sell (fixed overhead). If a service costs £40 in materials and your time to deliver and you sell it for £100, that £60 gap has to cover your overheads and leave something over — and that arithmetic needs to work before you scale it up, not after.
This is also the stage to check obligations around tax, insurance, and any licences relevant to what you plan to do; GOV.UK and a qualified accountant are the right places for that detail, not general guides like this one.
How do you set a price that actually works?
Pricing should start from the cost figures above and from what the market already pays for comparable alternatives, not from a round number that feels comfortable. If you price below what the arithmetic requires, more sales will make the problem worse, not better. If you price without reference to the market, you risk pricing yourself out of reach of the customer you identified earlier.
It is usually easier to start slightly higher and justify it with a clear proposition than to start low and try to raise prices once customers are used to a cheaper figure. Early customers are also a useful source of evidence on pricing — if nobody questions the price, that is itself information worth noting.
Pricing is rarely final. Treat your first price as a working hypothesis you will revisit once you have real sales evidence, rather than a figure fixed at launch.
What is your proposition, in one clear sentence?
A proposition is a plain statement of who you help, what you do for them, and why that is better than their current option. It is not a slogan or a mission statement — it is the sentence a customer should be able to repeat back after hearing it once. If you cannot state it simply, that usually means the customer or the problem has not been defined tightly enough yet.
Test the proposition on real people outside your own circle before relying on it. If it needs several minutes of explanation to land, it will struggle to do its job in an advert, a conversation, or a website — all places where you typically have seconds, not minutes.
A sharp proposition also makes every later decision easier: what to build first, what to say in marketing, and what to leave out.
What is the minimum you need to actually deliver it?
Resist the instinct to build everything before you sell anything. The minimum infrastructure is whatever lets you take a genuine order and deliver it properly — that might be a simple booking system, a basic website, a way to take payment, and the tools or stock to do the work, without the extras that can wait. Extras added before you have a paying customer are a cost with no evidence behind them yet.
This is also where Evans Business Builder is relevant for people who want structured support turning a validated idea into a working business over a defined period, rather than working it out alone — useful once you have evidence an idea is worth building properly, not before.
Keep a clear list of what is genuinely required to deliver versus what would be nice. Add from the second list only once real demand justifies the investment.
How will customers actually find and buy from you?
A sales route is the specific path by which a stranger becomes a paying customer: where they find out you exist, what convinces them, and how they hand over money. Many new businesses have a reasonable product and no clear route to a buyer — that is a common and avoidable failure, not a mysterious one. Decide on one or two realistic routes rather than trying to be visible everywhere at once.
Realistic routes usually involve either direct outreach (contacting likely buyers yourself), referral from people who already know and trust you, or a visible presence where your customer already looks for solutions — search, a relevant platform, or a local network. Each needs effort and usually money; pick the route that matches how your specific customer actually buys, not the route that is easiest to set up.
Work the chosen route before adding more. A second channel added before the first is working properly usually just dilutes effort rather than doubling reach.
What actually happens at launch?
Launch is rarely a single dramatic event; it is usually the point where you stop testing in private and start taking real orders in public, even in small numbers. Treat the first weeks as a continuation of the testing you have already been doing, not as the moment everything must be finished and perfect.
Keep the operation simple enough that you can watch what is happening closely: who buys, what they ask about, where they hesitate, and what goes wrong in delivery. This period generates more useful information than almost any amount of planning beforehand, provided you pay attention to it.
It is also sensible to have a modest, specific target for the first weeks — a number of enquiries or sales — so you have something concrete to compare actual results against, rather than a vague sense of how things are going.
How do you turn early interest into real customers?
Early interest — likes, enquiries, people saying they are interested — is not the same as a customer. The gap between interest and a completed, paid transaction is where most of the useful work happens: following up promptly, answering the actual objection rather than a generic one, and making it easy to say yes. Chase enquiries properly rather than assuming interested people will come back on their own.
Your first handful of customers are worth more than their individual revenue because they test your whole process end to end: pricing, delivery, communication, and whether the result matches what was promised. Treat them accordingly — ask what nearly stopped them buying, and what they would change.
Early customers are also often the best source of referrals and testimonials, provided the experience you gave them was genuinely good rather than merely adequate.
What should you do with the evidence you are gathering?
Everything from this point generates evidence: which enquiries convert, which don't, what people ask before buying, what they complain about afterwards. Build a habit of recording this rather than relying on memory or impression, even informally in a notebook or spreadsheet. Patterns across ten customers tell you far more than your strongest instinct about any one of them.
Be specific about what you measure: conversion from enquiry to sale, cost of getting each customer, repeat purchase or referral, and actual margin once all costs are included. These are the numbers that tell you whether the business works, rather than whether it feels busy.
Avoid the trap of generating lots of activity without reviewing what it tells you. Busyness is not the same as progress, and a business can be extremely active while quietly losing money on every sale.
How do you improve the business once it is running?
Improvement should follow directly from the evidence above: fix whatever the data shows is actually broken, not whatever feels most interesting to work on. If enquiries are plentiful but sales are low, the problem is probably pricing, proposition, or follow-up, not a lack of marketing. If sales are fine but margin is thin, the problem is cost or pricing, not volume.
Resist adding new products, services, or channels before the core offer is reliably working and profitable. Expansion before the basics are solid usually multiplies existing problems rather than solving them.
Once a business is established and generating reliable evidence of what works, tools such as Evans' Opportunity Engine or an Evans Launch website can help scale a proven offer further — but they are for amplifying something that already works, not for finding out whether it does.
This guide is general commercial information, not legal, tax, or accountancy advice; check current GOV.UK guidance or speak to a qualified accountant or solicitor before registering or trading.
Next step
Not sure which idea to pursue? Use the free tool. Already chosen? Explore Evans Business Builder.