Launch Day Is When Your Website Starts Working

We're fortunate to live in this era of computers, the internet, and AI.

Building a website has never been easier. And precisely because it's easy, many companies still think of it as "once it's built, we're done."

The last week of a website project usually looks like this: click through page by page hunting for typos, scroll through the mobile version, send a test message through the form to yourself. Sign off, wire the final payment, and everyone breathes a sigh of relief. Then the website is left alone, and it stays that way for three years. Open it three years later and the "Latest News" on the homepage is still stuck in the month it launched.

This piece is about an illusion of sequence. Launch day gets treated as the finish line, but that's the day the website starts working for you. So on that day, besides signing off, you should also make an organizational decision: from today on, who keeps it fed, and how often does someone look at its numbers.

01Projects close. Websites don't.

Treating a website as a project is a perfectly reasonable instinct. It has a quote, a timeline, a sign-off, and a clear completion date. When the project ends, the people on it go back to their regular jobs, and the budget closes along with it.

But once the website is delivered, the company doesn't stop. A service gets added, another gets dropped, prices are adjusted, the phone number changes, the person who handled inquiries leaves. Every change moves the version of you on the website a little further from who you are now.

An earlier piece, "Did You Adopt a Tool, or Another System Nobody Owns?", covered something very similar: an AI system with no designated maintainer becomes a liability three months in. The automation in that piece breaks quietly. It keeps producing output, it's just wrong, and nobody notices. A website is the exact opposite. It breaks in front of everyone: every customer who opens it sees an out-of-date version of you.

02Customers and Google are both checking the date

An SEO practitioner account put it bluntly: build a website and then ignore it, and Google will assume you've gone out of business. Their rule of thumb is that after more than three months without new content, a website looks to Google as if it has stopped operating, and it gets ranked further down.

That three-month line is this account's experience-based claim, and the exact number of months is open to question. Their analogy, though, is hard to argue with: if a shop had a "Closed" sign hanging on the door, would you still go in to buy something?

Customers don't wait for the algorithm. They make their own call. Someone hears your company's name for the first time, searches for it, and clicks through to your website. The latest news item is a photo from the year-end party two years ago, the copyright year in the footer is stuck at the year you launched, and the services list still includes work you stopped taking long ago. They won't write to ask whether you're still in business. They'll close the tab and move on to the next company.

The same write-up also points to two places that often get forgotten: the structured data on your pages, written for search engines and AI to read, and your frequently asked questions (FAQ). When the content changes and these two don't change with it, what machines read is old information. And the questions customers ask shift as the business changes, so the FAQ should change with them. The layer people can't see goes out of date too.

03Some things can't be made up later if you skip them on launch day

Old content can at least be fixed. The more troublesome kind is the kind you can't get back once you've missed it.

One checklist on preparing a new site for launch has a one-line subtitle: don't wait until after launch to fill these in. The last step on the list is installing Google Analytics 4 (GA4), and the reason is simple: without data, every adjustment after that is guesswork.

Another write-up on GA4 setup mistakes gets more specific. When you create a new GA4 property, event-level data retention usually defaults to just 2 months, after which the data is deleted automatically. Even if you change it to 14 months afterward, the old data that has already been deleted doesn't come back. The setting takes a few clicks in the admin panel. It gets missed because nobody was assigned to think about it on launch day, not because it's hard.

We're not going to add yet another maintenance checklist here; there are plenty out there already. What we want to point out is the shape of the problem: a sign-off checklist checks whether the website was built right, and not one line asks who takes over after launch. Leave that line blank, and everything after it lands on no one.

04On launch day, write down two answers

So we suggest that at the sign-off meeting, besides the signatures, you write down two more answers.

The first is a name. Who keeps it fed. It has to be one person's name; "the marketing department" doesn't count, and neither does "the vendor." This person doesn't need to know how to code, and doesn't need to write articles themselves. They need to know what the website currently says, be the first to think "the website needs to change too" whenever something changes at the company, and either have access to make the change or know who to ask.

What it gets fed mostly doesn't have to be produced from scratch. The question customers keep asking lately is the entry your FAQ should add; the pricing approach sales changed last week is the section your service page should update. The material has been inside the company all along. What's missing is someone to move it onto the website.

The second is a rhythm. How often someone looks at the numbers. Weekly, monthly, or quarterly all work, depending on how much your business relies on the website to bring in inquiries. A company that gets its work through referrals may be fine checking once a quarter; a company whose customers mainly come from search may need to check every week. The pace matters less than whether it's on the calendar rather than "we'll look when we have time."

What to look at doesn't need to be complicated either. Start with two questions: is anyone finding you through search, and once they find you, is anyone leaving their contact details? If you can answer those two, what to change next is usually clear.

Neither answer costs extra money, and neither has to be outsourced. They're organizational decisions. It's just that on launch day, nobody usually brings them up. The project team assumes the client will take over, the client assumes the project team is still looking after it, and the website falls into the gap between them.

The flip side needs saying too: not every company needs to post every week. Content squeezed out for the sake of it isn't much better than two years without an update. The three-month line is enough as a reminder that someone should be watching. Only when someone is watching do you know when it's time to act.

05A redesign is often paying off years of debt

When many companies reach the point of a redesign, the reason they give is that the website is old. Look closely, and what has gone old is mostly the content: three years untouched, and the website is no longer describing the company as it is today.

Rebuilding the whole thing at that point means paying off all the maintenance skipped over those years in one go. A redesign can bring the website back to the present, but if nobody writes down those two answers on the day the new website launches, the same bill will come due again three years later.

A handover, not a close-out

Our own website is set up the same way. On the 8th of every month, we run a fixed check of GA4 and Google Search Console. The check doesn't make any decisions for us, but "how often we look" now has a fixed time, and nobody has to remember it.

One team, from strategy to launch. The stretch after launch needs your own people.

If your website has been live for a while and you're not sure what it's saying for you right now, start with a website audit: lay out which pages are already out of date and which numbers aren't being tracked at all. Then decide whether to change anything. There's still time either way.

Rooted in craft. Built for the new wild.