← Back to Work

Data Management Platform

Building Enterprise Infrastructure for Audience Intelligence

TIMELINE 2016–2019
ROLE Product Manager
TEAM 15–20 People

Overview

After Email Machine, I joined a fast-growing Data Management Platform company as Product Manager. The DMP served major Czech and Central European media agencies, providing real-time audience segmentation, cross-device tracking, and programmatic advertising infrastructure.

During my tenure (2016–2019), I led product strategy for a platform handling 500+ million user events daily, serving 50+ enterprise media agency clients, and processing billions of dollars in programmatic ad spend. The platform was the backbone enabling agencies to target audiences with precision and transparency.

Scale Challenge: First deep dive into real big data challenges—processing 500M+ daily events while ensuring strict data quality and privacy compliance (GDPR).

The Problem: Data Intelligence Gap

"What We Solved"

By 2016, programmatic advertising was growing rapidly, but media agencies faced a critical gap: they had data, but couldn't leverage it effectively due to fragmented sources and slow, manual processes.

  • Fragmented data sources (App, Web, CRM)
  • Manual segmentation taking weeks
  • Lack of unified customer view
  • Growing privacy/compliance concerns

The opportunity was clear: build a platform that unifies data, enables real-time segmentation, and gives agencies programmatic control over their audience strategy.

Approach & Strategy

Three Core Pillars

1. Real-Time Data Ingestion & Unification

What it was: Event collection from all sources unified under a persistent user ID. Cross-device tracking matching users across touchpoints.

Why it mattered: Agencies could target users based on the complete journey, not isolated clicks. This drove campaign relevance and ROI.

2. Self-Service Audience Segmentation

What it was: Visual UI for non-technical marketers to build complex rule-based segments. Real-time query computation.

Why it mattered: Reduced segment creation from weeks to minutes. Marketers could test and iterate on audiences in real-time.

3. Privacy-First Architecture

What it was: First-party cookie strategy, GDPR compliance, transparent data handling, and audit trails.

Why it mattered: Privacy became a competitive advantage. Agencies could confidently use the platform as regulations tightened.

Platform Interface

Complex data made accessible through intuitive design.

Audience Segmentation Builder Campaign Dashboard Detailed Analytics

Technical Architecture

INGESTION
Apache Kafka Event Streaming
STORAGE
Vertica (Columnar DB) + Redis
PROCESSING
Spark Batch & Streaming Jobs
PRIVACY
SHA-256 Hashing & Audit Logs

Execution: Timeline

Phase 1 (2016 – 2017): MVP & Pilots

Built core ingestion and basic UI. Partnered with 3-5 pilot agencies. Validated value of unification.

Phase 2 (2017 – 2018): Enterprise Scale

Cross-device tracking, GDPR compliance features. Processing grew to 200M+ daily events. 30+ agency clients.

Phase 3 (2019 – Present): Leadership

Advanced ML features (lookalikes). 500M+ events daily. Consolidated market leadership in Central Europe.

Key Metrics & Impact

500M+

Daily Events

50+

Enterprise Clients

99.95%

SLA Uptime

90%+

Retention

Lessons Learned

1. Data Quality is Everything

Garbage in = garbage out. We moved from fast ingestion to strict validation because fixing bad data is harder than preventing it.

2. Privacy is a Feature

GDPR was an opportunity, not a constraint. Building privacy-first accelerated trust with enterprise clients.

3. Complexity Hides in Platforms

Engineers think queries are simple; marketers don't. We had to invest heavily in UX to make the complex infrastructure usable.

Building infrastructure for non-technical users?

The DMP taught me that scale requires relentless focus on UX alongside powerful backend engineering.

Let's Talk