How does Datafly Signal handle telco scale (millions of daily customers)?

The platform is single-tenant Kubernetes with horizontal scaling on every component. Real customer events queue through Kafka with backpressure handling, and Delivery Workers scale independently per vendor. Tier-1 telco deployments routinely run at 50,000+ events per second sustained, with peak handling capacity of 500,000+ events per second. Capacity planning is part of the deployment runbook.

How is web-to-app-to-store-to-contact-centre identity stitched?

A single anonymous_id flows from Datafly.js on the web, through the iOS / Android SDKs in the app, into POS systems via server-to-server events from store retail systems, and into contact-centre conversion events delivered from your CCaaS platform. Sign-in events on any surface link the anonymous_id to a user_id, after which all surfaces resolve to the same customer record. Cross-channel attribution becomes deterministic rather than probabilistic.

Can it support roaming and pre-paid top-up event flows?

Yes. Roaming session-start events from network OSS systems and pre-paid top-up events from your billing platform flow into the same pipeline as web and app events. The unified event schema means downstream segmentation, churn modelling, and offer targeting work across every event source. Telco-specific event taxonomies (data session, voice session, SMS pulse) are supported via custom blueprints.

How does this fit alongside our existing CDP or marketing automation platform?

Datafly Signal is the upstream first-party collection layer. Cleaner, higher-volume events flow into your CDP (Salesforce Data Cloud, Adobe Real-Time CDP, Treasure AI, Amperity) for identity resolution and audience activation. Most telco clients keep their existing CDP and add Signal underneath rather than replacing the CDP. The same upstream pipeline also delivers to ad vendors and the warehouse.

How is per-region data residency handled for multi-country operators?

Single-tenant deployment per region means UK customer data lives in UK infrastructure, EU in EU, US in US, with no shared data plane. Per-vendor delivery gates can further restrict downstream sharing. The same product team can manage every region through one admin console while the data itself stays compartmentalised by jurisdiction.

Telecommunications

Millions of events. Dozens of channels. One platform.

Telecoms providers run some of the highest-volume digital marketing operations in any industry. Customers research plans online, visit stores, call centres, and use apps — often over weeks. Signal captures every touchpoint and delivers complete data to every marketing platform in real time.

Millions

Events per day

High-volume event streaming built on enterprise message infrastructure

20-40%

Conversions hidden

Ad blockers and browser restrictions hide a third of plan signups

4+ weeks

Research cycle

Customers compare providers for weeks before committing to a contract

<50ms

End-to-end delivery

Events reach every vendor API in under 50 milliseconds

The multi-channel attribution problem

Telecoms customers interact across more touchpoints than almost any other industry. Client-side tracking breaks down when the journey spans web, app, store, and call centre.

Web-to-app-to-store journeys

A customer researches plans on the website, checks coverage on the app, and signs up in-store. Without persistent identity, each channel reports independently with no connection.

Massive ad budgets with poor visibility

Telecoms brands spend heavily on acquisition across Google, Meta, TikTok, and programmatic. When 30% of conversions are hidden, bidding algorithms optimise on incomplete data.

Long consideration, high lifetime value

Contract decisions take weeks. Customer lifetime value runs into thousands. Yet attribution data expires after 7 days in Safari — long before the customer commits.

How Signal solves telecoms attribution

First-party collection across every channel, 400-day persistent identity, and real-time server-side delivery to every ad platform — with attribution that covers the full customer lifecycle.

400-day attribution

Server-set first-party cookies that persist for the entire research and contract cycle. Attribute plan signups, upgrades, and renewals to the campaigns that drove them.

Survive the full consideration window
Track from first ad impression to contract signature
Attribute renewals and upgrades to original acquisition

Multi-source collection

Collect events from website, mobile app, server-side APIs, and webhooks — all flowing through one pipeline with unified identity and governance.

Web: 5.2KB JavaScript collector
Mobile: iOS and Android SDKs
Server: REST API for CRM and billing events
Webhooks: Ingest from any third-party system

Enterprise-scale delivery

Built on high-throughput event streaming infrastructure. Handle millions of events per day across dozens of vendor destinations with guaranteed delivery and zero data loss.

Event streaming decouples collection from delivery
Per-vendor rate limiting and exponential backoff
Dead letter queue for failed deliveries
Real-time debugger across all pipelines

Attribution across the full customer lifecycle

Telecoms customers interact across web, app, store, and call centre over weeks. Every ad platform claims credit for the same signup. Signal's configurable attribution models give you a single source of truth — from acquisition through upgrade to renewal.

Last Click

Attribute the plan signup to the final campaign. Clean baseline for migration.

First Click

Credit the channel that drove initial consideration — weeks before contract.

Time Decay

Weight credit toward recent touchpoints while acknowledging the research phase.

Position Based

40/20/40 — credit both brand awareness and the conversion driver.

Linear

Equal credit across the multi-week, multi-channel journey.

Custom Rules

Define logic per product — mobile, broadband, bundles, B2B vs consumer.

Pipeline as Code

One plan upgrade event, every platform

A single plan_upgrade event is captured once, enriched with plan details and contract value, then delivered simultaneously to Google Ads for conversion optimisation, Meta CAPI for audience building, and BigQuery for lifetime value analysis.

Plan name, value, contract length, and device inclusion captured
Customer PII hashed before leaving your infrastructure
Simultaneous delivery to ad platforms and data warehouse
Complete customer lifecycle data for LTV modelling

pipeline/plan-upgrade.ymlYAML

# Plan Upgrade Pipeline
event: plan_upgrade
source: web

org_data_layer:
  schema:
    - field: properties.plan_name
      type: string
      required: true
    - field: properties.plan_value
      type: number
      required: true
    - field: properties.contract_months
      type: number
    - field: properties.device_included
      type: boolean
  pii:
    - field: context.traits.email
      action: sha256
    - field: context.traits.phone
      action: sha256
  consent:
    required: [analytics, marketing]

pipelines:
  - name: google_ads
    integration: google-ads-enhanced-conversions
    event_name: purchase
    mapping:
      - source: context.traits.email_hash
        target: user_data.sha256_email_address
      - source: properties.plan_value
        target: conversion_value
      - source: properties.plan_name
        target: custom_variables.plan_name

  - name: meta_capi
    integration: meta-conversions-api
    event_name: Purchase
    mapping:
      - source: context.traits.email_hash
        target: user_data.em
      - source: properties.plan_value
        target: custom_data.value
      - source: properties.plan_name
        target: custom_data.content_name

  - name: bigquery
    integration: google-bigquery
    mapping:
      - source: properties
        target: event_properties
      - source: context.traits
        target: user_properties

Frequently asked questions

How does Datafly Signal handle telco scale (millions of daily customers)?: The platform is single-tenant Kubernetes with horizontal scaling on every component. Real customer events queue through Kafka with backpressure handling, and Delivery Workers scale independently per vendor. Tier-1 telco deployments routinely run at 50,000+ events per second sustained, with peak handling capacity of 500,000+ events per second. Capacity planning is part of the deployment runbook.
How is web-to-app-to-store-to-contact-centre identity stitched?: A single anonymous_id flows from Datafly.js on the web, through the iOS / Android SDKs in the app, into POS systems via server-to-server events from store retail systems, and into contact-centre conversion events delivered from your CCaaS platform. Sign-in events on any surface link the anonymous_id to a user_id, after which all surfaces resolve to the same customer record. Cross-channel attribution becomes deterministic rather than probabilistic.
Can it support roaming and pre-paid top-up event flows?: Yes. Roaming session-start events from network OSS systems and pre-paid top-up events from your billing platform flow into the same pipeline as web and app events. The unified event schema means downstream segmentation, churn modelling, and offer targeting work across every event source. Telco-specific event taxonomies (data session, voice session, SMS pulse) are supported via custom blueprints.
How does this fit alongside our existing CDP or marketing automation platform?: Datafly Signal is the upstream first-party collection layer. Cleaner, higher-volume events flow into your CDP (Salesforce Data Cloud, Adobe Real-Time CDP, Treasure AI, Amperity) for identity resolution and audience activation. Most telco clients keep their existing CDP and add Signal underneath rather than replacing the CDP. The same upstream pipeline also delivers to ad vendors and the warehouse.
How is per-region data residency handled for multi-country operators?: Single-tenant deployment per region means UK customer data lives in UK infrastructure, EU in EU, US in US, with no shared data plane. Per-vendor delivery gates can further restrict downstream sharing. The same product team can manage every region through one admin console while the data itself stays compartmentalised by jurisdiction.

Mobile SDK

iOS and Android SDKs for the telco app — same identity as web and store.

Identity

Cross-channel identity stitching from web to app to store to contact centre.

Data Warehouses

Real-time delivery to BigQuery, Snowflake, and Databricks at telco scale.

PII Handling

Customer data classifications and per-vendor transforms for compliance.

Complete attribution across every channel

See how Datafly Signal captures every touchpoint across web, app, and server — and delivers complete conversion data to every ad platform in real time.

Request a Demo View Platform