Developer Onboarding
SingleStore is a real-time distributed database used by major companies for high-throughput data ingestion and real-time analytics. Traditionally sales-led and enterprise-focused, its growth relied on long sales cycles and large deals. To create a more predictable and scalable model, the company aimed to expand into product-led growth (PLG), making it easier for developers and small teams to adopt the platform independently.
Role | Timeline |
|---|---|
Led research, strategy, and stakeholder alignment. | 3 months |
Problem
The complex onboarding experience for our cloud database was not optimized for self-serve developers, resulting in high drop-off rates and lost opportunities to convert users into paying customers. Simplifying this process was critical to improving user activation and driving product-led growth.
Analysis & Research
Beyond the initial analytics research that I found in Mixpanel and google analytics, I also wanted to explore the problem space more thoroughly.
Personas
Persona worksheet based on prior user research
Several personas signed up for SingleStore, but the core persona was the application developer. They had the ideal use case, making it easier to build a tailored solution, and could quickly recognize the value of paying for the service.
Goal of Application Developers
Test with sample data that match their industry and use case.
Load their own data and see how SingleStore performed against key queries.
User Flow Analysis
Analytics pulled from custom Mixpanel Dashboards
Detailed user flow of the onboarding process
Percent | Stage | Description |
|---|---|---|
~72% | Authentication | Inconsistent tracking meant this number was unreliable |
83% | Use Case Qualification | Although only 45% matched our ideal use cases. |
23% | Engagement | Write at least one query |
1% | Conversion | Enter a credit card, meaning very few made it to a paid tier. |
Competitive Research
Analysis of 9 competitors, including Snowflake, Databricks, and MongoDB
Stage | Findings |
|---|---|
Authentication | 5 competitors offered GitHub authentication. |
Most flows had ~4 form fields on a single page. | |
Qualification | Plan options were shown multiple times, helping users understand pricing and future costs. |
Engagement | Product tours were shown before instance spin-up giving the system time to initialize. |
Conversion | Most competitors had a clear CTA to upgrade, often with free credits as an incentive. |
Many platforms featured data import GUIs to streamline the onboarding process. |
Design
Beyond the initial analytics research that I found in Mixpanel and google analytics, I also wanted to explore the problem space more thoroughly.
Simplified flow
Updated flow with numerous steps removed
Several personas signed up for SingleStore, but the core persona was the application developer. They had the ideal use case, making it easier to build a tailored solution, and could quickly recognize the value of paying for the service.Condensed FlowFewer steps with integrated password & termsUnified tracking all in one system as opposed to 3
Rebranded Signup with Github Authentication
Updated sign up page
Github auth: aligned with developers' preferences and made it easier to integrate data loading in the later stages of onboarding.
Updated Branding: The forms were updated to be consistent with new marketing and product branding.
Qualifying use case
Simplified use case questions.
Condensed flow for faster progression.As most of the questions we asked here were pretty simple, it didn't make sense to break them into single tasks. We also randomized the possible answers to ensure that users were actually answering them and not just clicking through. Collected dev stack & coding language preferences to personalize onboarding in the next stages of the flow
Engaging users through tailored guides
Easy Setup
Easy setup. No configuration necessary.
Simplified configuration that eliminated the necessity to configure a workspace. Something the user didn't know about initially. Upfront pricing that highlighted how much the service could cost later. Spin up workspaces after this first step.
Animated educational tour
Simple tour that masked the cluster spin up time.
While the workspace was being spun out, we could walk them through a short tour about how the service was different from the databases they had used in the past.
Tour highlights:
Compute and storage were separate
AI can be used to optimize queries and configurations
Customized guides based on qualification questions
Customized notebook guide based on use case questions.
Using our python-based notebook application, we built out numerous guides so the user could try out the product and get set up in a couple of steps. These were very specific and much more likely to align with what they wanted to do. We also included a quick prompt to our AI tool where they could ask custom questions.
Conversion to paid
Simplified checklists and in app messaging.
Incentivized onboarding checklists
Add in checklists that take there first onboarding guide further into the product ecosystem. They can also get additional credits for doing these steps.
Better in app upgrade messaging
Add in key triggers that launch intercom automated messages that get the user in contact with a sales rep to walk them through pricing. As it can be somewhat complex.
Impact
This complex project took coordination across multiple teams. I had to meet with different eng managers and pms to figure out how we could seamlessly and easily fit these changes into their roadmap without affecting their velocity. Here are some of the key stats.
+10% increase in authenticated users.
+8% increase in qualified users.
+20% increase in users running their first query.
Future Improvements
Enhanced "Load Data" GUI to simplify data ingestion.
Integrations GUI to help users connect external tools.
Persistent upgrade options to increase conversion.
More refined sales messaging in chat & Intercom.
Key Take Aways
Big-picture strategy drives impact.
Taking a holistic approach to onboarding rather than solving individual UX issues in isolation led to significant improvements in engagement.Cross-functional alignment is crucial.
By involving engineering, PMs, and marketing from the start, we were able to execute faster and tackle systemic issues rather than just UI problems.
