You shipped it, now what? The first week after a launch, for one person

The post goes up on a Tuesday morning. By lunchtime four hundred people have been on the site, which is four hundred more than yesterday, and you are refreshing the tab like it owes you money. By Tuesday evening there are eleven emails in your inbox. Nine of them ask the same two questions. You answer all eleven by hand, badly, because you are also fixing the sign-up bug that someone found in the first ten minutes.

Two things break in that afternoon, and neither of them is the thing you spent six weeks building.

Nine of the eleven emails are the same two questions

The questions people ask on launch day are not the interesting ones. On a product like yours they are almost always some version of “does this work with the thing I already use” and “how do I get my data back out”. You wrote both answers down already. They are on your site. The person emailing you did not find them, or did not trust them, or did not look.

You will answer these by hand for a while, and that is fine. What is not fine is that nobody is keeping score. By Friday you will remember that “a few people” asked about the export, not that six did, and not that most of those six were on the pricing page when they gave up and opened their email instead.

Keep a tally as they come in, even in a text file. One line per question, a mark each time it turns up again.

The traffic you are not keeping

Referrer data only exists while the visit is happening. If you install analytics on Thursday, the four hundred people who came on Tuesday are gone, and so is any record of where they came from. You will know your launch worked. You will not know whether it was the post, the comment someone left underneath it, the newsletter that picked it up on Wednesday morning, or the one person with sixty thousand followers who quietly shared it.

The next launch is a decision about where to spend your effort, and you have just thrown away the only evidence you had for making it.

This is the one job worth doing before you post rather than after, and it takes about a minute.

What actually deserves your Tuesday

Not the conversion funnel. Not A/B tests. Not a help centre with forty articles in it.

You need three facts in week one and they are all shallow ones. How many people came, where they came from, and which page most of them arrived on. That last one matters more than it sounds, because the page strangers land on is the page doing all the explaining, and you have almost certainly never read it as a stranger would.

Everything else in a modern analytics dashboard was built for a company with a growth team.

The smallest thing that covers both

Most founders end up with neither the tally nor the traffic data, because covering both normally means two accounts, two scripts, two bills and an afternoon of setup that launch day does not have in it.

Deacon covers both from a single script tag.

The support half puts a button in the corner of your pages. Someone types a question and gets an answer built from the help pages you gave it, with a link to the page it used. When your material runs out it hands the conversation to you with the visitor’s address attached, so nothing is lost while you are asleep. As the questions arrive they are grouped by what they mean, which is where the “six people asked about the export” fact comes from without you having to remember it.

The analytics half runs off the same tag, with no second script to install, and the counting sets no cookie of its own. You get visitors, visits, page views, how many left after one page and how long people stayed. Alongside those you get the visitor counts for each source, each landing page, each country and each kind of device. Mark up your own sign-up button and you can watch how many of the four hundred got that far.

What you can safely ignore until month two

Session recordings. Heatmaps. A proper help centre. A status page. A changelog.

None of those are bad ideas. They are all things you build once you know which parts of the product people struggle with, and in week one you do not know that yet. By the end of the month you will, because the topics and the unanswered questions will have told you, and then you can write the right three help pages instead of guessing at forty.

If you have not launched yet

Put the tag on before you post. It is one line and a minute, and it is the difference between knowing what happened on your launch day and guessing about it for the rest of the year.

The three ways to add a chat widget to a site covers the options if you would rather compare before choosing, and the install guide walks through the dashboard side once you have picked.

If you built it with an AI coding tool, Lovable, Bolt, v0, Replit and Base44 each have a page with the steps for that editor.