Launch week is the one week a small product gets more attention than it can read. Hacker News sends a wave on Tuesday, Product Hunt another on Thursday, and by Friday you have a chart that looks like a mountain range and no idea which peak paid for anything. This post is the plan we would follow: what to set up weeks before, what to look at on the day, and how to read the week afterwards in money. The numbers come from the Vero demo, which tells exactly this story for a fictional SaaS.
Weeks before: install early and let a baseline form
A spike only means something next to a normal day, and a normal day needs history. Vero compares the last 24 hours with a typical day: the mean of the same 24 hours one to four weeks back, counting only weeks the site was already recording. Install a week before launch and you get a comparison with one week behind it. Install four weeks before and the baseline is as steady as it gets.
Use those weeks to check the plumbing, not to read trends. Confirm pageviews arrive with the three-step install, send the events that mark progress (signup, trial_started, purchase) with the custom events guide, and connect your payment provider and watch one real sale arrive with its source. A launch is a bad day to discover the webhook URL was pasted into the wrong environment.
Tag every link you control
You do not control the link on a Hacker News submission, and you should not dress it up with tracking parameters. The referrer does that job: news.ycombinator.com arrives on every click from the site. What you do control is everything else: the Product Hunt listing, your posts on X and LinkedIn, the email to your list, the message in a community you belong to. Tag each of those with a utm_source, a utm_medium and a utm_campaign.
Keep the names boring and consistent. Lowercase, one name per place (producthunt, never ProductHunt and product-hunt on different days), the medium as the type of link (social, email, listing) and one campaign name for the whole launch. Email is the case that needs tags most, because email clients often send no referrer at all: without a utm_source, your launch email looks like Direct. A UTM builder saves you from typos at eleven at night, and Vero has one in the dashboard.
What should you watch on launch day?
The realtime view is fun and mostly useless after the first hour. The useful question on launch day is how far you are from a typical day, and whether revenue moves with traffic. On the demo’s Sunday, days after both front pages, visitors were 1,484 against a typical 475, up 212.4%. Revenue that same day was €2,640 against a typical €3,500, down 24.6%. Traffic tripled and money fell. That is not a disaster, it is a launch: the crowd that arrives from a front page is curious, not ready to buy.
Anomaly alerts compare each day with its own baseline and email owners and admins when a number falls. Rises are marked as spikes in green, and they are emailed only if you switch spike emails on. For launch week, switch them on: a spike email at a time you were not looking tells you a link is circulating somewhere you did not post it.
After: read the week by revenue, not visitors
Here is the demo launch week. Visitors were 14,724, up 112% on the week before. Revenue was €18,423, up 34%. Traffic doubled and revenue grew by a third, which is the normal shape of a launch and the reason a traffic chart alone will mislead you.
Now split it by source. Hacker News, by referrer, sent 3,420 visitors and €7,124. Product Hunt, tagged producthunt, sent 1,632 visitors and €1,845 from 19 sales, a conversion of 1.16%. The site as a whole converted 1.27% (187 sales from 14,724 visitors), so Product Hunt brought attention but sold a little below the average. The posts tagged twitter sent fewer people, 1,245, and converted 3.9%: 49 sales and €4,821. And reddit, where the demo ran ads, sent 421 visitors and no sale at all.
| Source | Visitors | Sales | Revenue |
|---|---|---|---|
| news.ycombinator.com | 3,420 | 72 | €7,124 |
| 1,245 | 49 | €4,821 | |
| producthunt | 1,632 | 19 | €1,845 |
| 421 | 0 | €0 |
In the UTM report, Product Hunt is the biggest tagged source of the week. In money, it comes behind twitter, which sent fewer visitors. That is the whole argument for tying sales to sources before you launch, and it is why the next decision (where to post next time, where to spend) should come from the revenue table. If you paid for any of that traffic, subtract the cost first: revenue by source is not profit.
How do you tell the spike from what stays?
A launch has two effects and you want to know both. The spike is the burst of visitors in the days you were on a front page. What stays is the higher floor afterwards: more direct visits from people who remembered the name, more branded search, a few new referrers from blogs that wrote about you. Compare the second week after launch with the weeks before it, not with launch week itself, and watch Direct and Search in particular.
For paying customers, revenue cohorts group buyers by the week of their first payment. The launch cohort is the one to follow: if its renewals hold over the following months, the launch brought customers and not only curious visitors. Renewals keep the source of the customer’s first attributed payment, so the cohort still tells you which launch channel it came from.
A launch-week checklist
Two to four weeks before: tracker on every page, the events that mark progress firing, the payment webhook connected and one real sale seen with its source.
The day before: UTM-tagged links for every post you will publish, one campaign name, spike emails switched on.
Launch days: check the day against a typical day, not the live counter. Note what you posted and when, so you can line it up with the chart later.
The week after: revenue by source and by referrer, conversion per source against the site average, and the sources that sent visitors and no sale. Then the second week, to see what stayed. If you are a solo founder doing all of this alone, the indie hacker use case shows the smallest setup that covers it.
Unsure of a term? Read the glossary

Vero is built by one person in Spain. I read every waitlist entry and pick the beta accounts myself.