Canva unifies and activates data to power personalization for 260M+ users

Canva is one of the world’s fastest-growing visual communication platforms, empowering more than 260 million people across 200 countries to design, collaborate, and publish anything — from social media posts to enterprise-scale brand assets. With a mission to make design accessible to everyone, Canva has redefined how individuals and organizations communicate visually.

As its global user base expanded, Canva faced a growing challenge: personalizing outreach to its expanding user base, at scale. To sustain its rapid growth and deliver targeted, relevant experiences to millions of users, the company needed a unified, reliable way to turn vast amounts of data into actionable insights.

The challenge: Delivering personalization at scale

Canva offers users free trials of its platform — across Free, Pro, and Enterprise tiers — via its freemium model. Scaling this model to attract more people and grow the business depends on understanding user behavior across every touchpoint. Marketing, Sales, and Engagement teams needed a 360-degree view of the customer to optimize acquisition, retention, and lifetime value.

But data was scattered across dozens of systems — from advertising and engagement platforms like Google, Facebook, and Braze to internal analytics and product systems. Without a central source of truth, teams couldn’t easily compare performance, build consistent audience segments, or measure campaign impact.

Data engineers spent valuable time building and maintaining fragile, custom integrations to move data between systems. These pipelines often broke when APIs changed, creating reporting delays and data inconsistencies that slowed down decision-making. Instead of innovating or supporting new personalization initiatives, engineers were stuck managing manual data workflows.

As personalization needs expanded and data volumes grew, Canva recognized it needed a modern, automated solution to centralize and activate data across the business — one that would eliminate manual maintenance, ensure freshness, and provide every team direct access to reliable, analytics-ready data. 

Fivetran + dbt: A unified, scalable data foundation 

To achieve personalization at scale, Canva’s data team adopted a data foundation designed for speed, scalability, and accessibility — powered by Fivetran and dbt. Together, these technologies created a complete data lifecycle from ingestion to transformation and activation, helping not only centralize data from silos but ensure flexibility and adaptability across Canva’s data architecture.

Fivetran became the engine of data movement, seamlessly centralizing data from more than 60 marketing, sales, and engagement platforms into Snowflake, Canva’s cloud data platform. With reliable pipelines and an extensive connector library, Fivetran eliminated manual engineering and ensured every team had access to trusted, up-to-date data. With support for virtually any source, Fivetran allows Canva to easily test new tools and applications without needing to build or maintain manual APIs. This flexibility empowers Canva’s teams to innovate freely, choosing best-fit applications and scale without compromise.

“It was obvious Fivetran was the right choice,” says Frank Wu, Director of Data Platforms at Canva. “It makes us agile and more responsive to customer needs — without requiring engineering resources.”

Once the data was landed in Snowflake, dbt became the core of Canva’s transformation workflow. Canva’s data team used dbt to:

• Model and standardize core analytics logic that helped the team report on key outputs like sessions, marketing attribution, and search impressions
• Increase business team impact velocity through self-service analytics via governed GitHub contributions
• Automate downstream data products using dbt post-hooks to export modeled data to Amazon S3 and internal applications

By introducing incremental modeling and layered architecture, Canva reduced full data build times from hours to just minutes — accelerating analytics workflows, speeding up experimentation, and giving teams fresher insights faster.

But centralization and modeling were only half the story. Fivetran Activations turned that data into action. Using reverse ETL, Canva synced enriched data from Snowflake directly into operational tools like Braze and Salesforce, enabling real-time personalization, audience updates, and optimized engagement without engineering support.

By removing the need for custom APIs, Activations gave Canva’s marketing teams a self-service way to make data actionable across business applications. This interoperability lets teams easily deploy new pipelines, test emerging tools, and experiment with the latest machine learning innovations — all while maintaining consistency and trust across systems.

“Fivetran helps us recognize what’s working and optimize the opportunities we uncover,” adds Wu. “Fivetran Activations takes those insights and makes them actionable, allowing us to deliver more personalized experiences for our users.”

The impact: Faster insights, smarter personalization, measurable ROI 

Together, Fivetran and dbt transformed how Canva’s Data and Marketing teams collect, prepare, and activate data — driving measurable gains in efficiency, engagement, and performance:

• 360-degree customer visibility across marketing, sales, and product touchpoints
• Weeks of manual work are reduced to minutes, empowering analysts and marketers to self-serve
• Stronger governance and trust in data flowing across systems

By bringing together data movement, transformation, and activation through Fivetran and dbt, Canva has built a unified foundation — one that connects insights to action and powers meaningful engagement for over 260 million users. Canva’s journey proves that when modern data infrastructure is automated, scalable, and accessible, it becomes more than an analytics engine — it becomes a catalyst for innovation, empowering every team to move faster, experiment with confidence, and deliver experiences that make an impact.

[CTA_MODULE]