Scaling Consumer Intelligence for Fortune 500 Brands with a Modern Analytics Platform

InMarket is an omnichannel marketing platform that helps Fortune 500 brands find new customers and increase sales. The company turns first-party location data into consumer intelligence. Billions of events flow through the platform every day. They power audience targeting and attribution for brands like Anheuser-Busch and Dunkin’.

01 The Challenge

A growing data business that needed a platform to match its ambitions

The location-based advertising market reached $107 billion in 2024. It is projected to grow to $356 billion by 2033. For platforms like InMarket, the ability to process location data at speed and volume is the product. The insights are only as good as the pipeline that delivers them.

InMarket was processing over 5 billion events per day: location signals, real-time bidding logs, and consumer behavior data. The volume was growing. The team saw the opportunity to build a platform that could grow with it. Yet, data processing jobs completed about 40% of the time; the data team spent time on re-runs instead of building new products. Getting a new pipeline from notebook to production was estimated to take up to 12 months.

The company wanted to move faster. A modern analytics platform would raise job completion rates and shorten deployment from months to days. It would also free the data team to focus on product work that drives revenue. InMarket partnered with Provectus, an AI-first systems integrator and solutions provider, to build it.

02 The Approach

Design for 5 billion events per day, then organize the platform around four workload clusters

Provectus started with InMarket’s data architecture and processing patterns. The team mapped event data flows and identified where the architecture needed to evolve.

The approach was to rebuild the platform on AWS using a microservices architecture. Instead of one processing layer handling everything, Provectus split the workload into four clusters. Each handles a specific job type. Ingestion and filtering. Real-time event processing. Campaign analytics. Custom research and partner data delivery. Each cluster scales and fails independently.

Provectus chose managed AWS services for data ingestion, stream processing, and data preparation. A streaming layer handles real-time data flow. A cloud data warehouse supports BI and analytics downstream. The architecture cuts pipeline handoff from months to days.

03 The Build

Four workload clusters, a streaming data layer, and a modern development environment

The platform separates InMarket’s workloads into four clusters. The first handles ingestion and filtering: raw events arrive, get filtered by location and time, and move downstream. The second processes real-time events from bidding logs and location signals. The third runs large-scale campaign analytics. The fourth serves custom research and partner data delivery.

A streaming layer sits between ingestion and the data warehouse. It reshapes and aggregates data on the fly: filtering, enriching, and routing before storage. The data warehouse holds processed output for BI queries and analytics reporting.

The development environment changed how InMarket’s team works. Data scientists build, test, and deploy predictive algorithms through a managed pipeline. Model deployment that took months now takes days. The platform monitors job execution and surfaces failures immediately.

04 The Results

From 40% to 99% job completion, with $1M in annual ROI and $100K in monthly savings

Job completion rates went from 40% to 99%. Data scientists shifted from managing re-runs to building new data products.

$1M

Annual ROI increase

With $100K in monthly infrastructure savings

The data team’s productivity increased 50%. Reliable processing, faster deployment, and a modern development environment gave the team capacity it did not have before. The 10X increase in cluster throughput let InMarket optimize its infrastructure spend, saving up to $100K per month.

Pipeline deployment dropped from up to 12 months to days. Data products move from science to production without waiting for engineering sprints. For Fortune 500 clients, that means faster access to new consumer intelligence products.

05 What’s Next

A data platform built for the next wave of location-based marketing products

InMarket now processes over 5 billion events per day with 99% reliability. The platform gives the company room to grow its data product portfolio without rearchitecting. Provectus works with InMarket on extending the platform as the company grows.