How Mattermost Achieved Visibility into the Entire Customer Journey and Empowered Their Product Managers

Mattermost is an open-source, self-hosted messaging platform for developers and IT teams. Their collaboration platform accelerates DevOps workflows in high-trust environments through secure, cross-platform messaging.

“Our previous vendor’s model kept our data in their playing field and the pricing meant we were limited in what we could afford to track. Mattermost was built as an open source solution, and we really wanted our data in our own warehouse, on our own playing field.”

— Alex Dovenmuehle, Head of Data Engineering, Mattermost
Mattermost Customer Data Stack

RudderStack Use Case
Mattermost uses RudderStack for comprehensive user event collection and routing, which drives powerful reporting through Snowflake and Looker.
Collection & Sync
RudderStack SDKs, Stitch Data, Heroku Connect
Data QA
RudderStack Data Governance API
Transformation & Enrichment
DBT
Warehouse & Data Lake
Snowflake
BI & Cloud Toolset
Looker
Job Orchestration
Airflow

To dig deeper, listen to Alex break down Mattermost's data stack in more detail in this episode of The Data Stack Show.


Mattermost is a fast-growing company building an open-source collaboration platform that rivals Slack and Microsoft Teams. For any high-growth product company, it is extremely critical to be able to make responsive, data-driven decisions about their products. Because their product is core to the daily workflow of large teams, understanding feature usage at scale is a key ingredient of optimization and new feature development.

Mattermost had initially invested in a well-known CDP vendor as a Customer Data Infrastructure solution but soon realized that there were several constraints in getting the most out of it. First, there was a limit on the amount of telemetry and the events that they could sync. In fact, because of the vendor's MTU-based pricing model, they were only able to sync 2% of the data to their warehouse, to stay within budget. Also, as a company built on open standards, Mattermost’s data engineering team wanted to build a non-proprietary stack, both because of their values and to hedge against vendor lock in.

Scalable Collection and Warehouse-first Architecture to the Rescue

Mattermost found the answer to their problems in RudderStack. The high-scale pricing model meant they could afford to collect all events and route them in real-time, not just a subset, and the ability to run the product in their own VPC solved significant security concerns related to their customers' high-trust environments. One of the most significant reasons for the switch, though, was the alignment with RudderStack on open standards, which is a core value at Mattermost.

"When we learned about RudderStack offering an open-source alternative to our original vendor, we thought, ‘Yes, this is exactly what we need.  Now we can remove restrictions to syncing events, and send all of the data we want to one central data warehouse.  We can analyze and act on all of that important customer data, and ultimately become a more data-driven business."

—Alex Dovenmuehle, Head of Data Engineering, Mattermost

Mattermost implemented RudderStack across all available platforms - including their server, as well as web and mobile apps. Previously, Google Analytics only gave them insights into  aggregate data, limiting their ability to join data sets, build complex cohorts and evaluate feature usage on a user level.  With RudderStack, Mattermost was able to get real-time data on custom user events across platforms, then join that with data from ZenDesk, Salesforce and Marketo. The result is a comprehensive customer data set, all synced seamlessly into their Snowflake data warehouse.

Empowering Product Managers with Feature Insights

As a specific use-case, Mattermost was looking at revamping the conversation threading feature on their collaboration platform.  With RudderStack, their engineering team can now give the product team specific, granular insights into how people are using threads, how many concurrent threads are running, and how many messages on average are posted under a specific thread.

Before, the product manager would rely on anecdotal feedback, intuition and incomplete data. Today, they work directly with the data engineering team to build a data-driven understanding of actual usage, which enables them to prioritize, plan and build better features.

Achieving Visibility into the Entire Customer Journey

Mattermost is also using RudderStack to get a much more cohesive view of the customer journey.  "For example, we can now see if a customer clicks on Google AdWords, lands on our landing page, purchases a license, and starts using our product,” Alex shared.“With RudderStack, we can now bridge all these gaps to have much better insight into our ad campaigns. We get to know how many people converted and how they used the product once they actually bought it."

Mattermost also uses RudderStack to further unlock the customer journey mapping across their offerings. For example, they have complete visibility on whether a customer filled out a form or visited certain content on their blog. They can then trace those cohorts to understand if customers have started using their product and how they are using it.

"We can give the marketing and sales teams timely data on how we need to think of our customers. We can now answer so many questions for the product managers - answers that we didn't have before, such as identifying the most relevant features that will have the biggest impact on our customer base."

—Alex Dovenmuehle, Head of Data Engineering, Mattermost
Discovering Hidden Insights into User Behavior

The Mattermost engineering team, as well as the products team, also discovered previously unknown use cases for their product.

For example, Alex discovered that Jitsi, an open-source alternative to Zoom, was an immensely popular plugin across the user base, something they didn’t learn about until they implemented comprehensive tracking with RudderStack.

Now they are working with Jitsi directly to better integrate the platforms.

Building for the Future with Better Collaboration

As head of engineering, Alex loves the product, but he also had high praise for the process and team. He described the relationship as a collaboration to discover new ways to unify and analyze customer data.

"With RudderStack, it feels less like we're a customer that bought a product,” he said. “Working with RudderStack is more like two partners building great things together.


Schedule a call with our sales engineering team—we'd love to answer your questions about where we fit in the stack

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.