User Story

MiQ Reinvents Programmatic Campaign Building with a Unified, AI-Driven Audiences Platform

Segment listing latency: ~2 second
acceleration on metric calculations: 40-80%
Query Caching Eliminated: Simpler

In the rapidly evolving world of programmatic advertising, MiQ has redefined what modern campaign building looks like. With the launch of its Audiences platform, MiQ has delivered a powerful solution that unifies audience discovery, segment management, and campaign activation across multiple demand-side platforms (DSPs)—all through a single, intelligent interface.

At the heart of the platform is a double engine of innovation: deeper data integration and applied AI. By consolidating previously fragmented data into one system and layering it with AI capabilities—from segment recommendations to persona generation—MiQ is giving its traders and sales teams smarter tools to drive better performance with less friction.

It’s a bold step in the industry’s ongoing shift from Mad Men to Math Men—where data is the fuel, and AI is the engine turning that data into insight, action, and advantage. The result is a future-forward platform that reflects MiQ’s leadership in bringing together technology and creativity to transform how programmatic campaigns are built and delivered.

The Challenge: Fragmentation, Friction, and the Real-Time Bottleneck

The idea of unifying data is not new—industries from finance to retail have long turned to data warehouses and centralized platforms to bring coherence to fragmented systems. But in the world of programmatic advertising, the challenge goes far beyond unification. Here, external users—including traders, sales teams, and increasingly, automated systems—issue a constant barrage of real-time queries across a fast-changing landscape of data sources and segment providers. Traditional approaches to data integration and retrieval simply couldn’t keep up, resulting in degraded performance and unacceptable activation delays.

Before the launch of Audiences, MiQ’s internal teams were burdened with manually searching across disparate DSP interfaces to locate and build campaign segments. Sales reps, meanwhile, relied on intelligence insights that were difficult to convert into actionable campaigns. This friction led to inconsistency in execution, time-consuming workflows, and lost opportunities.

As a platform-agnostic partner managing over 100,000 data segments across a range of providers, MiQ needed more than just a unified system—it needed one capable of handling high concurrency, sub-second query performance, and intuitive usability, all without sacrificing speed or scale. That system is now Audiences—a purpose-built solution that resolves long-standing industry bottlenecks and sets a new benchmark for operational efficiency and real-time responsiveness in programmatic advertising.

A Unified, Intelligent Workflow

At the heart of Audiences is a powerful search and comparison engine that allows users to discover and select segments using brand-based searches, pre-defined taxonomies, or even free-text prompts. Traders can compare reach and CPM across DSPs in real time, select from pre-approved campaign templates (via MiQ’s Automated Campaign Setup), and push directly to DSPs for execution.

Meanwhile, client-facing users can explore available audiences and build draft campaigns without access to sensitive pricing or activation tools. The result is a single, intuitive workflow that supports both internal operations and client collaboration.

"Audiences was built to solve a very real pain—our data was everywhere, and our traders were spending too much time stitching it together. Now, everything from discovery to activation happens in one place, and that’s a massive leap in productivity."

Stephanie Wong

Global Product Lead, MiQ

Infrastructure at the Speed of Search

Behind the scenes, Audiences is powered by a modern data stack designed for speed, scale, and flexibility. MiQ replaced Amazon Athena with StarTree, a real-time analytics engine built on Apache Pinot, to support the platform’s advanced search and indexing capabilities.

Pinot isn’t just any database—it’s the same technology trusted by Uber, DoorDash, Stripe, and emerging AI leaders like TogetherAI to deliver low-latency analytics on fresh data at extreme scale. These organizations rely on Pinot for mission-critical use cases like dynamic pricing, real-time fraud detection, personalized recommendations, and AI-powered customer interactions—all of which demand seamless integration with vector databases, model orchestration layers, and streaming data pipelines.

By adopting this proven infrastructure, MiQ ensures that Audiences isn’t just fast—it’s a forward-looking foundation built to fuel the next wave of innovation in AI-powered advertising.

The results were immediate and significant:

Segment listing latency dropped from 8–10 seconds to ~2 seconds
Complex metric calculations accelerated by 40–80%
Query caching eliminated entirely due to performance gains

StarTree’s columnar architecture enabled rapid filtering and retrieval of metrics like reach, impression volume, and CPM. This performance uplift was key to creating a responsive interface that traders could rely on in fast-paced campaign workflows.

Structured for Agility

MiQ’s taxonomy spans brands, audience categories, and segments, all mapped with the help of LLMs that generate descriptive metadata and improve discovery. The system’s vector indexing enables intelligent recommendations—even when users enter free-text queries—while ensuring all matches are grounded in actual, activatable segments.

Technically, the platform runs on containerized microservices within AWS EKS, with a React frontend and a Spring Boot backend. Data is stored in Amazon RDS and indexed in StarTree for ultra-fast analytics. This modular, cloud-native approach positions MiQ to continue scaling its capabilities as its audience data grows.

MiQ has already integrated AI into the Audiences experience—particularly through its persona-generation engine and metadata enrichment—but the roadmap doesn’t stop there. Plans are underway to incorporate more intelligent recommendations and tighter integration between the Audiences and Intelligence platforms.

By offering a more complete view of available segments, their characteristics, and their performance potential, MiQ is arming its teams with the insights they need to craft smarter campaigns, faster.

“Previously, slow segment listings in Audiences Manager limited user engagement. After integrating StarTree, performance leapt to sub-second response times, enabling multiple users to explore audience segments simultaneously—without any compromise in application performance. This boost not only drove higher usage but also encouraged users to discover and adopt richer features. And it all came with minimal architectural change: StarTree served purely as a query engine, seamlessly leveraging our existing S3 storage.”

Srijan Mehrotra

Principal Engineer at MIQ

Results and Impact

While the platform has only recently launched, early results indicate a clear step-change in how MiQ’s traders and sellers approach campaign design. Activation timelines have shortened, seller demo workflows are more impactful, and internal teams are reporting dramatically reduced effort in cross-DSP comparisons.

MiQ’s ability to unify segment data across platforms—and make it actionable through a single, intuitive tool—reinforces its role as a forward-thinking leader in the programmatic space.