How to launch a small business in 30 days

Thirty days is enough to take paid work, on one condition: you keep the offer narrow. One service, one kind of customer, one area. Everything in this plan is ordered so that the expensive, slow, or irreversible steps come after the cheap tests.

Week 0 — the three decisions that make the rest cheap

Before anything gets built, answer these in writing. Half a page is enough, and it saves you a month.

1. Who is it for? Not "everyone who needs a kitchen". "Landlords with one to five rental flats on the north side of the city." Narrow customers describe themselves in your ad copy.
2. What is the single outcome? The thing the client can see afterwards: a kitchen that passes inspection, a dog that is calm on the table, a set of accounts filed on time.
3. What do you charge, and how does the money arrive? A number or a range, a deposit rule, and the payment method. Pricing is a decision, not a discovery; the market will correct it later, but it cannot correct silence.

Write the offer as one sentence: "I [do this] for [these people] in [this place], starting at [price]." If you cannot say it in one breath, the rest of the plan will keep stalling.

Week 1 — proof of demand and the boring paperwork

Two tracks, run at the same time, because one is slow and one is fast.

Fast track: ten real conversations

Find ten people who match your customer, and ask them what they currently do about the problem. No pitch. Three questions are enough: what did you do last time, what annoyed you about it, what would have made you choose someone else. What you hear becomes your website copy, word for word. This is the highest-value week of the launch, and it is free.

Slow track: registration, bank account, insurance

These gate everything else, so start them on day one. Register the business, open a business bank account, and check whether your trade or profession needs a license, certification, or insurance before you take money. Timelines vary by region, and being uninsured on your first job is a risk no launch calendar is worth.

Week 2 — build the buying path, not the brand

By the end of this week a stranger should be able to find you, understand you, check your price, and book you. That is the entire week's goal.

• Domain and site. A short domain in your own name, five pages, written in the words from week one. Get it live on a real domain early — bugs show up at launch, not before.
• Booking. Your real services, real durations, real hours. If you are not ready for a calendar, a form plus an email reply within two hours is an acceptable starting point, but decide which one, and say it on the page.
• Payments. Connect a card processor in your own name, and test one live transaction for a small amount, then refund it. Testing with a real card teaches you more than any documentation.
• Google Business Profile. Free, fast, and the highest-return listing a local service business has. Fill in every field, add real photos, link to your booking page.

Skip the logo commission, the branded van wrap, and the merchandise. All three are things to buy after the first month of revenue, not before it.

Week 3 — first clients, before the launch party

The most common launch mistake is waiting until everything is perfect to tell anyone. Instead, sell in week three and use the feedback to fix week two's work.

1. Start with people who already know you. Past colleagues, neighbours, the businesses you use yourself. Ask for one specific job, not "any work".
2. Ask for the referral, not the sale. "Who else has this problem?" is easier to answer than "do you need me?" and it reaches buyers who are already looking.
3. Post where your buyers already are. Two or three local groups or communities you can genuinely help in, not fifteen platforms. Show the work: before and after, with the price range.
4. Set a target of three to five paying clients. Enough to prove the offer, collect two testimonials, and notice which service people actually ask for.

Week 4 — launch, measure, and cut

Launch is not a party, it is a state change: the site is live, bookings work, and you are answering enquiries in hours rather than days.

• Run the full path yourself on a phone: search, land, read, book, pay, get confirmation, cancel.
• Check your confirmation and reminder messages. 24–48 hours before the appointment cuts no-shows dramatically.
• Write down where each of your first clients came from. In month two, spend time only on the channels that already produced customers.
• Cut one thing you built that nobody used. There is always one: a page, a package, a social account.

A 30-day checklist

Week	Must finish	Cost
0	Offer sentence, price, deposit rule, service area	$0
1	10 buyer conversations; registration, bank, insurance started	Registration and insurance fees
2	Domain and five pages live; booking; payments tested; Google profile	Domain, site build, processor setup
3	3–5 paying clients; 2 testimonials; one repeatable referral source	Time
4	End-to-end test, reminders on, source tracking, one thing cut	$0

What usually goes wrong, and the fix

The offer stays vague. Symptom: people say "interesting" and never book. Fix: name the customer and the outcome in your headline, and put a price on the page.

Booking is harder than messaging. Symptom: enquiries arrive on Instagram, not the site. Fix: put the same booking link in bio, in replies, and on the Google profile — meet people where they already are.

No deposit, so no-shows. Symptom: a full calendar that earns nothing. Fix: require a deposit for first appointments and state the cancellation window clearly.

Week two eats the month. Symptom: three weeks on the website, no clients. Fix: a site that answers price and takes a booking is finished. Everything else is decoration and can wait for month two.

What month two looks like

If the plan held, month two is not a rebuild. It is three small moves: add the service people actually asked for, raise the price that produced no objections, and publish answers to the questions you are still typing out by hand. That is how a launch turns into a business instead of a project.