On Deck increases profile views by 70% with Snowflake, Hightouch, and Mixpanel
On Deck is a curated community platform designed to help founders increase their odds of building successful venture-backed companies. Founded in 2015, On Deck has helped found 1000+ companies. The platform offers a wide variety of resources through online-community, in-person experiences, and robust profiles to match founders with investors, early hires, and potential customers.
Results
- Increased community platform profile views by 70%
- Saved weeks of engineering time
- Enabled self-serve data access for every business team
- Built an entirely new data-driven process to ship a new category of features
- Enabled rapid experimentation and A/B testing of product features
The Challenge
On Deck’s entire business model relies on providing founder-to-founder support so members can more easily network and provide value to one another. Customer profiles are the building block for the entire On Deck experience. The problem was these profiles often were not updated frequently enough to stay fresh and engaging, and as a result, they were not attracting as much traffic and engagement as they could to introduce founders or encourage pier-to-pier support.
On Deck had rich data about founders and community members, but this data was heavily siloed and difficult to use. The company was also struggling to get the right data to internal teams so they could make decisions and drive product innovation. To better enrich community profiles with more data and empower internal teams to make better decisions, On Deck faced three core challenges:
- Simplifying the data architecture
- Syncing data to downstream operational tools
- Building data-driven products faster
This caused the company to implement Snowflake, Hightouch, and Mixpanel
The Solution
Snowflake
Struggling to get any value out of their existing data platform, On Deck’s data team turned to Snowflake to centralize their data. Since adopting Snowflake, all of On Deck’s data is available in one single source of truth–and since the platform is SQL native, there is no longer a technical gap preventing team members from accessing the data.
The data team can effortlessly scale workloads without worrying about query performance or underlying maintenance issues. This allows them to spend more time transforming and modeling the data to power both analytics and activation use cases.
With Snowflake, any engineer who knows SQL can work in our data stack. We never have to worry about query performance, and we can join together all of our data sources. Without Snowflake, we wouldn’t have a single source of truth.
Steven Schmatz
Head of Engineering
•
On Deck
Hightouch
While Snowflake gave On Deck the ability to centralize all of their customer data, the company still faced a key challenge in activating the rich insights living in Snowflake. This led the engineering team to implement Hightouch to automate data flows out of Snowflake to downstream destinations like Mixpanel, Airtable, and Postgres.
We were using another Reverse ETL tool, but it didn’t have any of the developer quality of life features that Hightouch provides, like version control, alerting, or support for other data tools.
Steven Schmatz
Head of Engineering
•
On Deck
Every business team at On Deck operates out of Airtable to manage day-to-day operations so ingesting Snowflake data into Airtable was extremely important, especially since this is where the operations team models user preferences and goals. Likewise, the product team operates extensively out of Mixpanel to analyze and understand behavioral data and Postgres (On Deck’s production database).
Before Hightouch, all of this data lived in disparate spreadsheets that users had to upload manually into Mixpanel, Postgres, and Airtable. Now, with Hightouch, all user data is ingested automatically and consistently.
Almost everyone in our company is extremely familiar with spreadsheet software, but not Snowflake. Hightouch allows us to create shared models so anyone in our company can self-serve and send data where they need it.
Steven Schmatz
Head of Engineering
•
On Deck
Having access to the rich insights in Snowflake has unlocked a number of use cases for On Deck, but the most notable is the new community search platform the data team was able to build in the platform.
Using Hightouch to sync Slack data from Snowflake to Postgres, the data team was able to build an entirely new feature that automatically aggregates, surfaces, and summarizes Slack threads to answer critical member questions.
Mixpanel
Understanding user behavior is vital to product development at On Deck. The problem is that most BI tools require a ton of legwork to set up, and building dashboards can take weeks or even months of engineering time. Mixpanel eliminates this problem allowing On Deck’s product team to create self-serve dashboards with very little work on the user’s part.
Most analytics tools force you to build your own dashboards. Mixpanel provides value immediately with default dashboards for engagement, cohort retention, flow analysis, and funnel metrics. Building these same dashboards in another BI tool would take weeks of engineering time.
Steven Schmatz
Head of Engineering
•
On Deck
Using Hightouch to merge historical data with event-stream data in Mixpanel, the product team can analyze engagement patterns, retention patterns, cohort analysis, flow analysis, funnel metrics, and engagement to inform on all product features.
This insight has allowed the product team to roll out several new product features to better serve profiles, including historical Slack interactions and advanced searching functions, which have increased founder profile views by 70 percent over the last year.
On Deck also defines experimental cohorts in Mixpanel and uses Hightouch to sync them to Launchdarkly to ensure the product team has access to accurate A/B testing statistics. This allows the product team to see exactly what percentage of a cohort is using a specific feature.
By analyzing Snowflake data through Mixpanel via Hightouch On Deck can surface actionable insights from every customer interaction. With that critical feedback, we can build experiences and product features that we know our users want.
Steven Schmatz
Head of Engineering
•
On Deck
What’s Next?
On Deck is focused entirely on helping members provide value to one another. Success for the company is directly linked to engagement metrics within the product and helping founders identify valuable profiles. People are the building blocks of the On Deck community, so On Deck’s focus is on continually creating a community of strong interconnectivity. This modern data stack combination of Snowflake, Hightouch, and Mixpanel helps On Deck unlock new insights and act on those insights immediately so On Deck can better match users based on their preferences.
Our profiles are critical to our business. If we can make our profiles better it makes our product better. We use engagement as a way to quantify the impact of our product.
Steven Schmatz
Head of Engineering
•
On Deck