The Business Launch Checklist

Most launches do not fail on launch day. They fail three weeks later, when someone finally tries to buy, the payment form errors out on their phone, and there is no email address anywhere on the site for them to tell you about it. They just leave. You never find out it happened.

This is the list we work down before putting a business in front of customers. It is deliberately unglamorous. Nothing here is clever — it is the set of things that must be true so that a stranger can find you, understand what you sell, believe you enough to pay, and reach a human when something goes wrong. Work straight down it.

One assumption: you have already decided this is worth launching. If you have not spoken to anyone who would pay for it, stop and validate the idea first. A launch checklist executed perfectly on an idea nobody wants produces a very tidy failure.

What "ready to launch" actually means

Launch is the day the buy path is open and you start telling people. That is the whole definition. It is not the day the website is finished, because the website is never finished, and it is not the day you feel ready, because you will not.

Three tests. If all three pass, launch:

1. A stranger who has never met you can go from your homepage to a completed purchase without asking you a question.
2. You will find out within one hour that they did it, and so will they.
3. If something breaks in that path, they have an obvious way to reach a person, and that person answers.

Everything below exists to make those three sentences true.

30 days out

This is the phase for anything that depends on someone other than you. Banks, payment processors, domain registrars and insurers all move on their own schedule and none of them will hurry.

The offer and the price

• Write the offer in one sentence: who it is for, what they get, what it costs. If you cannot, the website cannot either.
• Set a real price. Not "contact us for pricing" unless you are genuinely custom-scoped, and not a placeholder you intend to change later — you will not.
• Decide what is included and what is not, in writing. Scope is what kills service businesses in month two.
• Write down the delivery promise: what happens after payment, and by when. "Within one business day" is a promise; "shortly" is not.
• Decide your refund policy and the exact wording of it, before your first unhappy customer rather than during.
• If it is a subscription: billing interval, trial or no trial, how cancellation works, what happens to their access afterwards.
• Sanity-check the unit economics. What does one cost you to deliver, including payment fees of roughly 3%, and what is left? If nothing is left, the price is wrong.

Legal, money and the boring machinery

• Register the entity, or confirm you are legitimately operating as a sole trader in your jurisdiction.
• Get the tax registration number (EIN in the US, UTR or company number in the UK, equivalent elsewhere).
• Open a business bank account. Never take customer money into a personal account with the intention of tidying it later.
• Open the payment processor account and complete identity verification. This takes days, not minutes, and is the most common cause of a delayed launch.
• Check whether you need sales tax, VAT or GST registration for what you sell and where. Rules for digital goods differ from physical goods and differ by state and country — look up your own case rather than assuming.
• Check licensing and insurance for your trade. Trades, food, health, finance and childcare all have requirements that are neither optional nor fast.
• Set up bookkeeping, even if it is a spreadsheet with three columns. Retrofitting a year of transactions costs real money.

Domain, hosting and identity

• Buy the domain and turn on auto-renew, with a registrar contact address you will still read in two years.
• Set up email on the domain. you@yourbusiness.com, not a personal Gmail on the contact page.
• Configure SPF, DKIM and DMARC. Skipping this is why your emails land in spam.
• Lock down DNS access and enable two-factor authentication on the registrar, host, payment processor and email.
• Settle name, logo, colours and typography enough to be consistent. Plain-and-consistent beats clever-and-inconsistent.
• Claim the social handles you intend to use, even if you will not post yet.

14 days out

Now the thing you built has to work. If you are still deciding what the product is, you are not two weeks from launch — you are somewhere earlier, and the honest version of that timeline is worth reading.

The website and its critical paths

There are only three paths that matter. Everything else on the site is optional.

• Homepage states what you sell, for whom, and what it costs, in the first screen. No mystery.
• Every page has one obvious next action.
• Pricing is visible. If you truly quote custom, show a range and what drives it.
• Mobile layout checked on a real phone, not a browser window resized to look like one.
• Page loads in under three seconds on mobile data. Compress the images — hero images are the usual culprit.
• Every link clicked once, by hand. Broken navigation on a new site is common and invisible to whoever built it.
• 404 page exists and links back to somewhere useful.
• Favicon set, so the browser tab reads as your business rather than "Untitled".
• Social preview image and title set, then tested by pasting the URL into a message to yourself.
• Spelling checked by someone who is not you.

Legal and policy pages

• Privacy policy, matching what you actually collect. If you run analytics and an email list, say so.
• Terms of service or terms of sale.
• Refund and cancellation policy, matching the words you decided 30 days out.
• Cookie notice and consent if you serve the EU or UK, or use anything beyond essential cookies.
• Business name, contact method and address or registered office where your jurisdiction requires it.
• Accessibility basics: real text rather than text-in-images, alt text on meaningful images, contrast you can read outdoors, keyboard-navigable forms.

Payments, tested end to end

Not "the checkout is installed". Tested. With money.

• Run one real transaction with a real card, at the real price. Test mode passes when live mode fails.
• Then refund that transaction and confirm the refund lands.
• Confirm the money arrives in the business bank account, and note how long it took.
• Test on a phone, on mobile data, in the browser most of your customers use.
• Test a declined card. Confirm the customer sees an understandable message rather than a blank screen.
• Confirm the customer receives a receipt, from an address that looks like you.
• Confirm you receive a notification of the sale.
• Check what your confirmation page says. "Thank you" is not enough — say what happens next and when.
• If you offer subscriptions: test the renewal, the cancellation, and the failed-payment path.
• If you ship goods: test the shipping calculation, including the most expensive destination you accept.
• Check the currency, the tax line and the total. Confirm the number on the checkout matches the number on the pricing page exactly.

Analytics and tracking

• Install one analytics tool. One. Confirm it is recording, in real time, from a device that is not yours.
• Define the events that matter before launch: page view on the buy page, checkout started, purchase completed, form submitted.
• Verify each event fires by triggering it yourself and watching it arrive.
• Set up call tracking or at least a dedicated phone number if the phone is a real channel for you.
• Connect the site to search console and submit the sitemap.
• Confirm the site is not blocked by robots.txt or a stray no-index tag. Staging settings shipped to production are a routine and expensive mistake.
• Turn on uptime monitoring that emails you when the site goes down.

7 days out

Email and follow-up

• Enquiry form sends to an inbox a human checks, and you have tested it from an outside email address.
• Auto-reply confirms receipt and sets a response time you can hold.
• Purchase confirmation email is written, tested and readable on a phone.
• The next step after purchase is scheduled: onboarding email, booking link, delivery note, or a plain "here is what happens now".
• Email list set up with a signup form, if you intend to have one. Send yourself the welcome email and read it as a stranger.
• Send test emails to a Gmail address, an Outlook address and one other. Check the spam folder in each. If you land in spam, fix authentication before launch, not after.
• Draft the follow-up you will send to anyone who enquires and does not buy. Send it 48 hours later. This one email recovers more revenue than most launch tactics.

Customer support basics

• A contact method appears on every page. Email at minimum, phone if you will answer it.
• Response time stated, and it is one you can meet on a bad day.
• Someone is actually watching the inbox during launch week — with notifications on, not "checked occasionally".
• Write short answers to the six questions you already know people will ask. Keep them in a document; paste them until a FAQ page earns itself.
• Decide what you do when someone asks for a refund. Deciding in the moment produces inconsistent answers you will regret.

The launch announcement plan

A launch announcement is a list of specific people, not a broadcast into the void.

• Write the personal list: every individual who should hear this from you directly. Aim for 30 to 100 names — past colleagues, existing customers, people who gave you advice, anyone who said "tell me when it is live".
• Write the individual message. One or two sentences, personal, with a specific ask: buy, refer, or just look and reply.
• Write the public post, once, then adapt it per channel.
• Prepare the assets: one screenshot or photo, one link, one line of copy.
• Line up any partners, communities or newsletters you have permission to post in. Read their rules before launch day, not during.
• Decide the order and the times. Personal messages first, public posts second. The personal ones are where the first customers come from.
• If you are running ads, set the daily cap low, write two variants, and confirm the tracking works before spending.
• Do a full dry run: open the site in a private window on a phone, as if you were a name from your list, and buy something.

Launch day

Short by design. If launch day is busy with building, you launched early.

• Confirm the site is live, on the real domain, over HTTPS.
• Buy something one more time, on the live site, and refund it.
• Send the personal messages. All of them. This is the actual work of launch day.
• Post publicly, once per channel.
• Update your profiles and signatures to point at the site.
• Watch the inbox and answer everything within the hour.
• Open analytics once, to confirm events are firing, then close it.
• Write down every question anyone asked. That list is your next month of website edits.

Do not change the site during launch day unless something is broken. Editing copy while people are looking at it is how a working page stops working.

The first week after

Now you are measuring, not celebrating. Four numbers, checked daily:

The gaps between the rows are the diagnosis. Visitors but no buy-page views is a message problem. Started checkouts but no completions is almost always technical — assume it is broken until you have proven otherwise on a real device.

Also in week one:

• Reply to every message the same day.
• Ask every buyer one question: what nearly stopped you? Their answer is the highest-value marketing input you will get all year. Ask the people who engaged and did not buy the same thing.
• Fix the top one or two friction points. Not ten.
• Confirm the first payout actually arrived in the bank.
• Check that receipts, confirmations and follow-ups all went out — automations fail quietly.
• Keep sending personal messages. Most founders send 40 on launch day and stop. The next 200 are where the business comes from.

What people forget

These are the items that break launches, and none of them are strategy.

Payment processor holds on new accounts. New merchant accounts commonly hold the first payout for a week or two, and high-risk categories longer. Your first customers may pay and you may see nothing in the bank for days. That is normal — but a sudden spike in volume on a brand-new account can also trigger a review that freezes funds. Verify the account fully before launch and know your payout schedule.

Domain email deliverability. A new domain has no sending reputation. Without SPF, DKIM and DMARC configured, a meaningful share of your launch emails go to spam — including purchase receipts. Set up authentication, then warm the domain with real messages to real people in the weeks before launch rather than blasting 500 addresses on day one.

Forms that were never tested on a phone. A form that works on a desktop can be unusable on mobile: a field hidden under the keyboard, a submit button below the fold, a date picker that will not open, an autofill that silently drops a required field. Test on an actual phone, on mobile data, with a thumb.

No way to contact a human. A contact form is not a contact method if nobody watches it. Put a real email address on the site and answer it. In the first month, being reachable converts better than any copy change you could make.

The confirmation page that says nothing. People need to know their money went somewhere and what happens next. A blank "thank you" generates support emails and refund requests you did not need to receive.

Nobody knows the site exists. The most common launch failure is not a bug — it is a founder who built for four months and told twelve people.

Expiry and access. The domain, the SSL certificate and the card on file for your host all expire — set calendar reminders now. And put every credential in a password manager someone you trust can reach.

A quiet launch is normal

Your launch day will probably be quiet. Some visitors, a few kind replies, possibly no sales. That is the ordinary outcome for a business with no existing audience, and it is not a verdict on the idea. Launch day is an announcement — it converts the people who already know you, which is a small number, and that is all it can do. The launches that look explosive almost always had an audience first, or spent money, or had been talked about for a year.

The real target is the first ten customers. Ten is enough to prove people pay, to hear the same objection three times, to find out what breaks at delivery, and to earn the testimonials that make the next ten easier. Getting there is mostly direct, unscalable outreach over several weeks — the path from idea to first customer is a run of individual conversations, not a launch-day event.

So plan for launch day to be small, judge the business on the first ten, and give yourself weeks to get there. If you would rather have a team run this checklist with you and take the build off your hands, that is what we do.

The short version

• Launch means the buy path is open and you have started telling people. Not that the site is finished.
• Start 30 days out with anything involving other people: entity, bank, payment processor verification, domain email authentication. They will not go faster.
• Test payments end to end with a real card, on a real phone, then refund it — and confirm the receipt, the notification and the payout.
• Three paths must work: buy, enquire, understand. Everything else on the site is optional.
• Launch day is a list of personal messages, not a broadcast. The public post is secondary.
• A quiet launch day is normal. The first ten customers are the goal, and they take weeks of direct outreach.