Home/Growth experiments/Analytics

Know which growth work brings you leads

We check your analytics for the visits and leads each growth experiment should count. Forms, booking steps and visits from agents or AI answers all need to be recorded correctly.

Book a call

What is marketing analytics?

Marketing analytics is the record of who came to your site, where they came from and what they did there, including forms and bookings.

It lets you judge an experiment by the leads it brought. We use Tin's native PostHog and Google Search Console connections.

Miscounted traffic hides which channels bring leads

When traffic changes but sales calls do not, we look at what the report has counted.

  • Tests become visitors. Our own agent's browser checks appeared in traffic figures.
  • Another domain supplies the report. We found Search Console connected to the wrong site.
  • A parked tab resembles demand. An open browser tab read as buyer interest.

A check along the visitor's path can uncover the reason for these misleading counts.

We check that every lead and visit gets counted

We take on the event checks behind your visit and lead totals.

Capture responses

We check events for forms, bookings and trial signups.

Separate automated visits

We distinguish agents from people following links in AI answers.

Identify companies

We name visiting companies when the available data permits it.

Find drop-offs

We locate the funnel steps where people leave.

We submit test leads before we trust the reports

We follow a submission from the page into its recorded event.

  1. First

    Define the lead

    We agree on a demo request, inquiry or trial for the experiment.

  2. Then

    Read the setup

    We inspect the property, domain and events feeding the report.

  3. Next

    Submit and trace

    We complete the path and confirm each event arrives, keeping our tests separate.

  4. Each week

    Choose the next action

    With you, we decide what to continue, change or stop: ship a change, repair measurement or wait.

Example: finding the visitors a cookie banner hid

After a UK client introduced a cookie banner on 16 September, analytics showed fewer visitors. Our engineer compared the periods: PostHog now counted only people who clicked Accept. It missed about 70-85% of visitors. The engineer found that in the first week.

See more results →

We treat a sudden traffic jump as a fault to rule out

We have seen one bot crawl reported as 28x growth on a Tin project. That is why we double-check what the system counted.

  • Use dashboard totals. A tool reports the visits and events it receives.
  • Read an agency report. A monthly report collects figures for the period.
  • Check before acting. We investigate the spike before it changes which experiment gets more work.

Talk to Tin

Let's discuss the numbers you use today, where they disagree with sales, and which buyer action should count as a lead.

Book a call

FAQ

Can this help us judge our paid ads?

Yes, we test the path from landing page to lead and read the counted action beside your paid campaign.

What can you check while traffic is still low?

We can test forms, booking steps and events before traffic grows. A few visits show where people stop, which gives us clues to follow up.

Does visitor identification name every visitor?

No, it identifies some companies where the data allows. We read that information beside what those visitors did.

Is a visit from an AI answer the same as agent traffic?

No, a person following an AI answer is a visitor; an agent reading or testing pages creates automated activity. We keep them apart.

Check which of your numbers you can trust

Meet the founders and engineers who will examine the experiments you are about to judge. We can start with what each test is meant to produce and which numbers would tell you it happened.

Book a call