Why We Built a Hybrid CDP on Snowflake Instead of Spending Millions on an Enterprise Platform.

When a leading QSR brand in Asia Pacific was preparing to invest millions of dollars in a traditional Customer Data Platform from a major enterprise vendor, we asked a different question: what if you could get better results, faster, with the infrastructure you're already building?

That question led to the RACE Platform, Rapida's AI Customer Engine, deployed on Snowflake with Braze as the activation layer. Within months, the client had a working proof of concept delivering real customer intelligence. Within a year, the platform powered a customer engagement program that won the Customer Impact and Growth Honour Award at the Braze Asia Awards, a first for any brand in the APAC region within this global QSR network.

No multi-million dollar CDP licence. No 18-month implementation timeline. Just a smarter architecture and a team that moved fast.

The Problem: Fragmented Data, Expensive Options, and Slow Execution

The client was sitting on a goldmine of customer data, POS transactions, delivery orders, app interactions, web behaviour, but none of it was connected. Marketing campaigns took months to plan and execute. Segmentation was manual and basic. Customer insights existed in PowerPoint decks that were outdated before they were presented.

The proposed solution from the market was a traditional enterprise CDP, with a long implementation runway, heavy licensing costs, and a dependency on external vendors to operate the intelligence layer. The brand's own customer knowledge would effectively be outsourced: every campaign cycle would require sharing data with third parties, generating point-in-time insights that got thrown away after each activation.

We saw a better path.

The Solution: A Hybrid CDP on Snowflake with RACE on Top

Rapida recommended building a Hybrid CDP directly on Snowflake, with our RACE Platform (AI Customer Engine) as the intelligence layer. The philosophy was simple: keep the data and the intelligence inside the client's own environment, build IP that compounds over time, and activate through Braze for real-time customer engagement.

Rather than a big-bang implementation, we started with a focused proof of concept. In a matter of months, the client was already seeing tangible value, not in a roadmap deck, but in production.

What RACE Actually Does

RACE is a progressive AI customer intelligence framework purpose-built for high-frequency consumer businesses. It's not a black box, it's a structured maturity journey that grows with the organisation.

The platform starts by understanding who your customers are, using algorithms that analyse purchase behaviour, frequency, recency, and spend patterns to create meaningful customer tiers and segments. This alone functions as a virtual rewards recognition program, grouping customers intelligently based on how they actually behave, without the overhead of traditional points-and-tiers loyalty systems.

From there, RACE layers on increasingly sophisticated intelligence. Advanced clustering algorithms incorporate dozens of behavioural signals, time of day, day of week, ticket size, product diversity, channel preference, to discover real customer segments that no amount of manual analysis would uncover: the weekend dinner couple, the weekday lunch solo diner, the shoulder-period group buyer.

The platform then scores every customer for churn risk and predicts lifetime value, creating a targeting matrix that tells the brand not just who is at risk of leaving, but how valuable they are and what kind of intervention makes sense. It even detects short bursts of purchasing activity, identifying when customers enter high-frequency phases so the brand can trigger personalised offers at exactly the right moment.

Critically, the entire system learns from itself. Campaign performance feeds back into model refinement. Segmentation evolves as customer behaviour shifts. The intelligence compounds with every cycle, creating a durable competitive asset that grows more valuable over time.

The Architecture

The platform follows Rapida's core philosophy: centralise business logic in the warehouse, build a scalable semantic layer, and enable self-service activation across teams.

At the foundation sits the Rapida Modern Data Layer, a Snowflake-based data lake paired with a curated semantic model. On top of this, the Rapida CDP provides a unified customer view with centralised metrics, customer journey mapping, and dynamic segmentation, all running inside Snowflake. No separate CDP licence. No data leaving the client's environment.

The RACE AI Customer Engine connects to the CDP layer and pushes intelligence directly into Braze for real-time activation across marketing channels. All of this is the client's IP. The intelligence compounds. The models improve. And unlike a traditional agency model where insights get presented in a slide deck and discarded, every campaign cycle makes the platform smarter.

The Phased Approach: POC to Production to Award-Winning

Rather than asking for a multi-year commitment and a seven-figure budget, we delivered in phases:

Phase 1, Proof of Concept: We built the Hybrid CDP on Snowflake and deployed the foundational segmentation model. Within months, the client had customer tiers running in production with personalised campaigns activating through Braze. The value was immediate and measurable.

Phase 2, Data Expansion: We progressively brought more data sources into Snowflake, delivery platforms, app behaviour, web interactions, additional POS feeds. Each new source enriched the customer profile and unlocked new segmentation possibilities. The ROI was remarkable: more data, better models, same infrastructure cost.

Phase 3, Advanced Intelligence and Race Advanced Segmentation: With a rich, unified data foundation in place, we rolled out the advanced models. The client launched it, a lightweight loyalty experience built on surprise and delight rather than traditional points and tiers. Personalised moments triggered by real-time customer behaviour, orchestrated through Braze, powered by RACE intelligence running on Snowflake.

Today, the client can perform deep customer analytics on any dimension through Snowflake and Power BI, and build and activate customer segments in Braze in hours, not months. The intelligence layer keeps getting smarter with every transaction.

The Recognition

The project won the Customer Impact and Growth Honour Award at the Braze Asia Awards, a landmark achievement and a first within the brand's global network in Asia Pacific. The award recognised what happens when you stop following the standard recipe: stronger customer engagement, more returning customers, and a measurable shift from transactional interactions to genuine relationships.

What This Proves

This engagement demonstrates three things we believe deeply at Rapida:

You don't need to buy an enterprise CDP to get enterprise-grade customer intelligence. The modern data stack, Snowflake as the foundation, a semantic layer for consistency, Braze for activation, and purpose-built AI models on top, delivers superior outcomes at a fraction of the cost.

Start with a POC, not a PowerPoint. The fastest path to value is a working proof of concept that demonstrates real results. Our client went from exploring a multi-million dollar platform investment to a production system delivering intelligence in months.

Build your own IP, don't rent someone else's. When you outsource your intelligence to an agency or a monolithic platform, the insights disappear when the contract ends. When you build the intelligence inside your own data environment, it compounds. Every campaign makes the models smarter. Every data source makes the segmentation richer. The competitive advantage deepens over time.

About Rapida Solutions

Rapida Solutions is a data and AI consultancy that helps enterprises turn fragmented data into scalable, production-ready platforms. We specialise in Snowflake data warehousing, Power BI semantic layers, Braze customer engagement, and AI-driven analytics, with deep domain expertise in QSR, retail, and high-frequency consumer businesses.

We work with brands across Australia, Thailand, and Brazil, delivering architecture that doesn't just report on the past, it activates the future.