Analytics Setup Before Launch Day

On launch day, you will want to know what is happening. How many people have visited? Where did they come from? Did they sign up? Did they do anything useful? If you wait until launch morning to set up your measurement, you will miss the first, most valuable hours, and some of what you miss cannot be recovered.

This article is a practical guide to setting up analytics before launch, in a way that is simple, trustworthy and respectful of the people you are measuring.

Decide what questions you need answered

Do not start with a tool. Start with questions.

• Where do visitors come from?
• Which pages do they look at?
• How many sign up, and at which step do they drop off?
• How many perform the key first action in the product?
• Which channels bring people who stay?
• How fast does the site feel to real visitors?
• What are people saying?

These map to a small set of measures: visits by source, page views, sign-ups, funnel steps, activation, retention and performance. If a number does not help answer a question, you probably do not need it.

Choose a tool, or two

Many analytics products exist, from simple, privacy-focused page counters to full product analytics suites. Whatever you choose, check that it:

• Reports visits by source and campaign
• Lets you define events, such as sign-up and first action
• Can exclude your own traffic
• Fits your privacy obligations
• Is understandable by the people who will use it

A single tool is enough for a first launch. Search Console is a useful, free addition for search visibility. Google describes it as a service that helps you monitor, maintain and troubleshoot your site's presence in search results: it can confirm that Google can find and crawl your site, help you fix indexing problems, show search traffic data and alert you to issues. If search is part of your plan, set it up and verify your site before launch.

Define your events

An event is something that happens that you want to count. Keep the list short and meaningful.

1. Landing page view. Usually automatic.
2. Sign-up started. The visitor began the form.
3. Sign-up completed. The account was created and confirmed.
4. First key action. The thing that shows the product worked for them: created a project, imported a file, sent a message.
5. Payment started and completed, if you take payments.
6. Return visit within seven days.
7. Support request sent.

Name events clearly and consistently, for example "signup_completed", and write the definitions where the team can find them.

Be careful about what data you attach to events. You rarely need names or email addresses in analytics. Counting is enough.

Tag your links

Tagged links tell you which channel produced each visit. The common method is to add short labels, often called UTM parameters, to the end of a link. They typically record the source (where the link was placed, such as the newsletter), the medium (the type, such as email or social) and the campaign (the name, such as launch week). You can add a fourth label for the specific link or content.

Example: a link in your launch email might carry the labels newsletter, email and launch-week. A link in a social post might carry the labels social, post and launch-week.

Keep a simple table of every tagged link you create, with where it is used. Use consistent, lowercase names. Test each link before you send it, to make sure it lands on the right page and the tags are recorded.

Test, test, test

Analytics fail silently. The only way to know they work is to try them.

1. Visit the page from a device that is not excluded and check the visit appears in real time.
2. Click a tagged link and confirm the source is recorded.
3. Complete a sign-up with a test account and check each event fires once, not twice or not at all.
4. Perform the key action and check the event.
5. Test on mobile and on at least two browsers.
6. Check what happens with privacy tools or blockers. Some visitors will block analytics, so numbers are never perfect.
7. Clean up. Mark or remove test data so it does not pollute your results.

Run these checks a week before launch, fix what you find and repeat the day before.

Exclude yourselves

Your team will visit the site constantly during launch week. Exclude your own traffic so it does not inflate numbers. Use filters on addresses, a cookie that marks internal browsers or a setting in the tool. Check it works. But remember to test from non-excluded devices, or you will never see what visitors see.

Respect privacy

Analytics involve data about people. Treat it carefully.

• Collect only what you need. Prefer aggregate counts to individual tracking.
• Say what you do. Your privacy notice should describe the tools you use and what they collect, in plain language.
• Follow the rules that apply. Depending on the tool and where your visitors are, you may need consent for certain cookies or identifiers. The Information Commissioner's Office provides advice for small organisations, and it is worth reading before you choose a setup. If you are unsure, take advice.
• Do not put personal data in URLs or event names.
• Set retention limits so data does not pile up forever.
• Give people a way to ask questions, such as a contact address.

A privacy-respecting setup is not only a legal matter. It is part of earning trust.

Watch performance as well as behaviour

How fast your page feels affects whether people stay. Google's Core Web Vitals give targets: Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less and Cumulative Layout Shift of 0.1 or less, at the 75th percentile of page loads. If your tools can report real-user performance, add it to your dashboard. If not, test on a mid-range phone and a normal connection before launch.

Build a simple dashboard

A dashboard should fit on one screen. Include:

• Visits today and by source
• Sign-ups today
• Conversion from visit to sign-up
• Activation rate
• Any errors or performance alerts
• Support messages waiting

Do not include everything. A cluttered dashboard hides the numbers that matter. Use words and symbols as well as colour so everyone can read it. The scribe on your team can update a manual version at set times if you prefer.

Capture the qualitative side

Numbers say what happened, not why. Pair them with human information.

• A question on sign-up: "How did you hear about us?" or "What are you hoping to do?"
• A feedback log: a shared document where every comment, email and message is recorded with the date and what you did about it.
• Short calls: a few conversations with early users in the following weeks.

These add colour that no chart can.

Agree what you will do with the data

Decide when you will look and what you will do. For example: check at nine, one and six; if sign-up conversion is below a threshold, review the landing page; if one source performs much better, share more there. Write the rules down before the day.

A worked example

A founder is launching an invoicing tool. She picks a privacy-friendly analytics tool, verifies the site in Search Console and defines five events: page view, sign-up completed, first invoice created, payment started and support request sent. She creates tagged links for her newsletter, two social posts and her listing, and records them in a table.

A week before launch, she tests: a visit from her phone appears in real time, a tagged link records the right source, but the sign-up event fires twice. She fixes the bug. She excludes the team's office address, then tests from her home connection to confirm that her own visits are still visible when she wants them. She adds a "how did you hear about us" field to sign-up and updates her privacy notice to describe the tool.

On launch day, she checks her dashboard at set times. At noon she sees that visits from the newsletter convert much better than visits from social posts, so she sends a second email in the evening. After the week, she has a clear table of sources, conversion and activation to bring to the retrospective.

Checklist

• Questions defined
• Tool chosen and installed
• Search Console verified
• Events defined, named and tested
• Tagged links created and recorded
• Internal traffic excluded
• Privacy notice updated
• Performance measured
• Dashboard built
• Feedback log ready
• Review times agreed

Remember to link your analytics plan to your listing: on this site the submit page lets you create it, the privacy page shows how this site treats information and the launches page shows what is scheduled.

Frequently asked questions

Can I rely on the numbers? They are estimates. Blockers and consent choices mean some visits are missing, so treat trends as more reliable than exact counts.

Should I pay for a tool? Not necessarily. Start with something simple and upgrade if you need to.

What if I discover a problem on launch day? Note it, fix it if you can safely and record what you could not measure so that you remember in the retrospective.