Engagement scoring, channel attribution and the lifecycle funnel

GuidesUpdated 2026-08-20

The CDP under Operations, CDP does three things that ordinary messaging tools do not: it builds one event timeline per person across every channel, it scores how engaged that person actually is, and it reports a lifecycle funnel you can act on.

One timeline per person

Every inbound and outbound message, delivery receipt, read receipt, form submission, Flow submission, QR scan, short-link click and order becomes an event on the contact's timeline.

Identity resolution runs underneath. Someone who filled a form with their email in January, scanned a QR code in March and replied on WhatsApp today is one contact, not three fragments. This is the part that makes the rest meaningful — you cannot score engagement you have split across three records.

Events sync on the background worker, which runs every minute, so a message sent seconds ago may take up to a minute to appear.

The engagement score

The score rewards recent two-way activity:

  • Replies weigh most. Someone answering you is the strongest signal there is.
  • Reads weigh next.
  • Deliveries weigh least — a delivered message only proves the number works.
  • Everything decays with age, so the score reflects the relationship now rather than a burst of activity two years ago.

That weighting is deliberate. Open rates on messaging channels are close to universal and therefore say almost nothing; replies discriminate.

Using it

Three segments extract most of the value:

BandWhat to doWhy
TopSend new offers here firstHighest response, lowest complaint risk — a cheap test before a full send
MiddleRun a re-engagement sequenceThese are recoverable; the cold tail usually is not
Cold tailSuppress from marketing sendsYou are paying per conversation to message people who never respond

Suppressing the cold tail is the fastest cost saving available in most instances, and it also protects your WhatsApp quality rating — repeatedly messaging non-responders is exactly what drives blocks and reports.

Channel attribution

Because the timeline is unified, you can see which channel a given contact actually responds on, not just which channel you happened to send on. Someone who ignores email and answers WhatsApp within minutes should not keep receiving email.

Use this to pick the channel per segment rather than per campaign. It costs nothing and it usually moves response rates more than rewriting the copy.

The lifecycle funnel

The funnel counts contacts by the furthest stage they reached, and never counts one twice:

  1. Reachable — you have a valid identifier and a lawful basis.
  2. Messaged — you have actually sent them something.
  3. Engaged — they responded.
  4. Converted — they completed an order or the goal you defined.

Read the shape, not the absolute numbers:

  • Wide top, narrow middle — an opt-in or content problem. You have permission to message people who do not want what you are sending.
  • Healthy middle, no conversions — a follow-up problem. People are engaging and then nobody closes.
  • Narrow top overall — an acquisition problem, and no amount of message tuning fixes it.

What it does not need

The Infobip People mirror is optional. It matters only if you run People-based journeys on Infobip; the CDP, scoring and funnel all work without it.

Steps at a glance

  1. Open the CDPGo to Operations, CDP. Every inbound and outbound message, delivery and read receipt, form submission, flow submission, QR scan, short-link click and order appears as an event.
  2. Read a contact timelineOpen a contact to see one merged timeline across every channel they have used, rather than separate histories per channel.
  3. Check the engagement scoreEach contact carries a score that rewards recent two-way activity: replies weigh most, reads next, deliveries least, and everything decays with age.
  4. Read the lifecycle funnelThe funnel counts contacts by the furthest stage they reached — reachable, messaged, engaged, converted — with no contact counted twice.
  5. Build segments from the scoreCreate dynamic segments for the bands that matter: re-engage the middle, suppress the cold tail, and treat the top band as the audience for new offers.
  6. Act on the shape of the funnelA wide top with a narrow middle is an opt-in or content problem. A healthy middle with no conversions is a follow-up problem. Fix the stage that is actually leaking.

Frequently asked

How is the engagement score calculated?

It rewards recent two-way activity. Replies carry the most weight, read receipts next, deliveries least, and every signal decays with age — so a contact who replied last week outranks one who read ten messages last year.

Does a contact who uses two channels count twice?

No. Identities across WhatsApp, SMS, Viber, RCS, email and social are resolved onto one contact, so a person who replies on two channels is a single record with a single merged timeline.

What are the lifecycle funnel stages?

Reachable, messaged, engaged and converted. Each contact is counted once, at the furthest stage they reached, so the stages sum to your total rather than double-counting activity.

What events feed the CDP?

Inbound and outbound messages, delivery and read receipts, form submissions, WhatsApp Flow submissions, QR scans, short-link clicks and orders.

Why has a message I just sent not appeared on the timeline?

Events sync on the background worker, which runs every minute, so a brand-new message can take up to a minute to appear. If it is still missing after a few minutes, that is a delivery or webhook problem rather than a CDP one.

Do I need the Infobip People mirror?

Only if you run People-based journeys on Infobip. The CDP works entirely without it — the mirror is an optional sync, not a dependency.

How should I use the score in practice?

Three segments cover most of the value: suppress the cold tail so you stop paying to message people who never respond, run a re-engagement sequence at the middle band, and send new offers to the top band first as a low-risk test.

Next step

Apply this to your own deployment

This guide describes decisions we make on live instances. Tell us your channels, systems and region and we will map it to an architecture outline, a provisioning plan and an indicative commercial model — usually within one business day.