← All posts

Analytics deep dive: consent, sessions and CRM sync

Second Brain OSAug 22, 2026
Analytics deep dive: consent, sessions and CRM sync

The day-one analytics post covered the numbers. This one covers the plumbing underneath them, because the plumbing is where most analytics setups either get creepy or get useless.

This one is neither, and the reason is three ideas working together:

  • Consent.
  • Declared identity.
  • Session joining.

Consent first

Every visitor to your site starts anonymous. The cookie banner is not decoration; it asks before any tracking cookie is set, and the answer is honoured.

Decline, and you are a nameless page view.

The analytics still counts you (a visit happened, a page was read), it just cannot follow you from one visit to the next. Traffic numbers stay honest without anyone being shadowed.

You and I have both clicked through banners that clearly changed nothing.

This one is the actual switch.

Declared identity

Some visitors go further and tell you who they are:

  • They fill a form.
  • Start a chat with the receptionist.
  • Book a call.

That act is what we call a declared identity: the person volunteered a name, an email, a phone number.

Here the segregation rule from earlier in the series applies with full force. Personal details go to one place only: your CRM.

The analytics layer never stores a name or an email. Even where the two systems must coordinate, identifiers are hashed on the analytics side.

Numbers over here, people over there.

Session joining

Now the part that makes the whole series click together.

A session is the unglamorous name for one person's visit: the pages they read, in order, and where they came from. When a visitor has both given cookie consent and declared an identity, their session can be joined to their contact.

From then on, the trail flows into the CRM as activity events on that person's record:

  • The ad they clicked.
  • The blog post they read.
  • The form they filled.
  • The chat they started.
  • The appointment they booked.

Even later visits attribute back to them, because the site can recognise a consenting, known person when they return.

The joining runs in both directions, too. Someone who chatted anonymously and only later shared an email gets their earlier trail merged onto the same contact.

One person, one record, assembled from whichever door they used first.

What you get

Open a contact in your CRM and read the journey top to bottom: source, pages, questions, booking.

Sort your contacts by their lead score, which grew with each of those events, and the top of the list is who to call.

And the boundary stays clean the whole way. Anonymous visitors remain a count. Known leads become complete stories.

Consent, not cleverness, decides which is which.

That is a system you can put in front of any customer, and any regulator, without flinching.

Every piece you have read connects to this record: the receptionist fills it, the forms feed it, the events score it, the emails answer it.

One question remains, though. All of this measured your website. What about the lead who lives entirely in a WhatsApp thread and never browses at all?

Next up, the last post in the series: trackable links, which carry all of this into every other channel.

That was Part 15 of 17.

Next: Trackable links: know the moment a lead opens what you sent