About UsBlogContact
Discuss your project

Guide · Planning a website

Planning website analytics before launch

Analytics installed after launch starts with a gap: the first weeks of data are missing, and the setup often reflects whatever was easiest rather than what the business needs to know. Planning analytics together with the website — as part of the structure and templates — produces data you can use from day one.

Updated · 3 min read · DealPro

[ 01 ]

In short

In short

  1. 01

    Start from decisions: what will you change depending on the numbers?

  2. 02

    Define key events: enquiries, calls, messenger clicks, purchases, bookings, sign-ups.

  3. 03

    Agree on UTM rules for every campaign before the first link is published.

  4. 04

    Connect website events to the CRM, so leads and sales can be traced to their source.

  5. 05

    Respect consent and never send personal data to analytics tools.

[ 02 ]

Guide

Step 1: questions and decisions

Write down the questions the business will ask:

  • Which channels bring enquiries and customers?
  • Which pages lead to enquiries, and where do visitors leave?
  • Which services or products get interest but no enquiries?
  • How do visitors from different languages and devices behave?

For each, note the decision it informs — budget allocation, page improvements, new content. This keeps the setup focused.

Step 2: key events

Typical events for a business website:

  • form submitted successfully (not just the button clicked);
  • phone number clicked;
  • WhatsApp or Telegram link clicked;
  • email address clicked;
  • booking started and completed;
  • purchase and its value;
  • file downloads, such as price lists or catalogs;
  • key interactions: calculator used, video watched.

Mark the events that represent business results as key events (conversions). Details in GA4 events plan.

Step 3: implementation in the templates

Developers add a data layer and event hooks in the site's templates, so tracking survives design changes. Forms send events only after successful submission; e-commerce pages send product and order data.

Step 4: campaign tagging

Agree on UTM conventions — source, medium and campaign names — and use them for all ads, emails, messenger links and social posts. Inconsistent tags make channel reports unreliable. See UTM tags guide.

Step 5: connecting to sales

Website analytics shows enquiries; the CRM shows which enquiries became customers. Pass the source and campaign of each enquiry into the CRM, and send qualified leads and sales back to analytics and ad platforms. This is the basis of end-to-end analytics.

  • Describe analytics in the privacy policy.
  • Use a consent mechanism where required and make sure tags respect it.
  • Do not send names, phone numbers or emails to analytics as event parameters.

See privacy policy for websites in Azerbaijan.

Step 7: reports and owners

Decide who looks at which report and how often. A simple dashboard with a few key numbers — traffic by channel, enquiries by channel and page, conversion rate, sales by source — is more useful than access to every report in GA4. See dashboards.

Step 8: test before launch

On the staging site, trigger every event and check that it arrives with the right parameters. After launch, compare analytics events with actual form submissions and orders for the first weeks.

Tools

  • GA4 for website and app analytics. See GA4 setup.
  • Google Tag Manager to manage tags.
  • Search Console for search performance.
  • Session recordings or heatmaps for qualitative insight, used within privacy rules.
  • CRM for lead and sales data.

[ 03 ]

FAQ

Frequently asked questions

Can we add analytics later?

You can, but you lose data from the launch period and often end up with an ad hoc setup. Planning it with the site costs little extra.

Do we need both GA4 and another analytics tool?

GA4 covers most needs. Some businesses add session recording or heatmap tools for design insights. More tools mean more scripts, so add them deliberately.

How do we know the data is correct?

Compare analytics with reality: forms received, calls logged, orders in the system. Small differences are normal; large gaps indicate tracking errors.

Related reading

[ 04 ]

Next step

Have a project in mind?

Describe the task in a few lines. We reply with questions and the next step — no commitment and no templated proposal.