Planning a successful launch (or relaunch)

There's a difference between going live and launching well. Going live is a button. Launching is the sequence of decisions around it — what's ready, what's tested, who knows, and what happens on the day the first stranger arrives.

Most launches we've watched go wrong didn't fail on the technical side. They failed because something obvious was missed in the rush, or because the site went live to an audience of nobody and the owner concluded the site didn't work.

This is a checklist and a sequence. Use the parts that apply.

Four weeks out: content and decisions

The work at this stage is unglamorous and it's what determines whether the last week is calm or frantic.

Finish your content

Every page written, every photograph taken, every product described. Not “mostly” — finished. Content is always the critical path, and a launch date set before the writing is done is a launch date that moves.

Make the commercial decisions

These aren't website questions, but they'll block you if left until the end:

• What do you charge, and do you show prices publicly?
• Where do you ship, and what do you charge for it?
• What's your returns window and who pays return postage?
• Do you offer in-store pickup?
• Are you required to collect sales tax, and where?
• What are your lead times, honestly?

Sort out your domain

Register the name you want or locate the login for the one you own. This is the single most common cause of a delayed launch, because it depends on a registrar account somebody set up years ago and can't get into. Find it now, not on launch morning.

Two weeks out: build and configure

The pages

Home, About, whatever your offer page is, and Contact, at minimum. Plus your legal pages: generate your privacy policy, publish your terms, and write shipping and returns policies in your own words if you sell.

Delete every template section you can't fill honestly. This is the advice people most often ignore and most often regret.

If you're selling: the commerce setup

Work through it in this order, because each step depends on the last:

• Products. Load them by hand or by CSV import. Real photographs, honest descriptions, correct prices, variants where they exist, and accurate stock — or explicitly untracked for anything made to order.
• Categories. Group them so a stranger can find their way, and drag your best things to the top of each.
• Stripe. Connect your account. Do this earlier than you think you need to; account verification can take a little time and it's a genuinely blocking dependency.
• Shipping. Set your rates, so a shopper sees the real cost at checkout instead of a surprise.
• Sales tax. If you're required to collect it, switch it on and mark which products are taxable. It's calculated and collected automatically at checkout, with no added platform fee.
• Pickup, if you offer it. Decide per product, and say so on the product page rather than surprising people at checkout.

If you take bookings

Add your services, put a booking form on the right page, and set what you're actually available for. Then book yourself in, end to end, and read the confirmation as a customer would.

Everything else

• Per-page SEO. A real title and description on every page — see our SEO guide.
• A share image, so your link looks deliberate when someone posts it.
• A newsletter signup somewhere sensible.
• Two-factor authentication on your account.

One week out: test like a customer

This is the week that separates a good launch from a stressful one. Stop building. Start breaking.

Buy something from yourself

A real order, on a real card, all the way through. Then refund it. You are checking the whole chain: that the price is right, the shipping calculates, the tax appears, the confirmation email arrives, the order shows up in your dashboard, the stock decrements, and the refund works.

Nothing else finds a broken checkout. Do it on a phone as well as a laptop.

Submit every form

Contact form, booking form, newsletter signup. Check each one arrives somewhere you'll actually see it. A contact form quietly emailing an address nobody monitors is the most expensive bug on the internet.

Read the whole site on a phone

Every page, outdoors if you can, in daylight. You are looking for text you have to squint at, headlines that wrap into four cramped lines, buttons too close together, and anything that pushes the important thing below three screens of scrolling.

Get five people to look

People who'll be honest, not people who'll be kind. Ask each one: what do we do, and what would you do next? If they can't answer both quickly from the home page, fix the home page. The fix is nearly always fewer things, not more.

Proofread it properly

Read it aloud, which catches what your eye skips over. Then have someone else read it, because you cannot proofread your own writing — you read what you meant.

Every hour spent testing before launch saves you a day of apologising after it.

Launch: the actual sequence

1. Choose your plan

Starter if you're not selling online, Store if you are. Both are month-to-month with no setup fee, and both include hosting, SSL, daily backups and unlimited pages and bandwidth.

2. Connect your domain

Register a new one or point one you already own; the certificate switches on automatically and your mailbox comes with it. Do this a day or two before you announce, not an hour — domain changes propagate on their own schedule.

3. Publish

One click. Then load the site fresh, on a device you're not signed in on, and check it looks the way it did in the editor.

4. Soft launch first

Resist announcing immediately. Let a small group use it for a few days — existing customers, friends, your newsletter list if you have one. Real people find things test plans don't, and you'd far rather they were found by people who like you.

5. Then announce properly

Email your list. Post it. Tell your regulars in person. Update every profile that points at an old address — your Google Business listing, social bios, directory entries, your email signature. Announce more than once, spaced out; most people miss the first one, and a launch is not news to anyone but you.

If this is a relaunch

Replacing an existing site has one additional risk worth taking seriously: the search rankings and links you've already earned.

• Write down your existing URLs before anything changes, especially your best-performing pages.
• Keep the same addresses where you can. Changing a URL for cosmetic reasons throws away accumulated value.
• Redirect what has to move. Every old address that changes should point at its new equivalent, not at the home page.
• Expect a dip. Rankings usually wobble for a few weeks after a significant change and then recover. Don't panic and start undoing things in week two.

And use a sandbox version to build the new site while the old one keeps serving customers. That's what versions are for — build privately, look at it as long as you like, and switch when it's genuinely better.

The week after

Launch day is the start of the useful information, not the end of the project.

• Watch your dashboard. Where do people land, what do they do, where do they stop? A page everyone leaves from is telling you something.
• Read the enquiries. The questions you get in the first fortnight are the questions your site failed to answer. Each one is a free instruction for what to fix.
• Fix quickly and visibly. Small things, fast. It's much easier to change a live site now than to schedule a round of improvements you'll postpone.
• Ask your first customers for reviews. Start the habit immediately; a site with reviews at month three converts far better than one still waiting for social proof at month twelve.

---

Launch it before it's perfect

The most common failure isn't a botched launch. It's the launch that never happens, because there's always one more product to photograph and one more page to write.

A live site earning nothing yet is still infinitely more useful than a perfect draft nobody can see. It can be found, linked to, shared, and improved — and improved is what every good site actually is. Everything on it can be changed tomorrow.

Set a date. Work back from it. Launch on it.