Cloud debut: Positioning an overlooked feature as the primary OSS to SaaS converter.

Cloud debut: Positioning an overlooked feature as the primary OSS to SaaS converter.

Project

Under a tight deadline, I pushed to reduce friction for Data Lineage, turning an under-treated feature into the primary conversion driver from OSS users to SaaS clients.

what i did

Project ownership

Product strategy

Research

Prototyping

UX & UI design

Team

Lead PM

1 Back-end

2 Front-ends

Project

Under a tight deadline, I pushed to reduce friction for Data Lineage, turning an under-treated feature into the primary conversion driver from OSS users to SaaS clients.

what i did

Project ownership

Product strategy

Research

Prototyping

UX & UI design

Team

Lead PM

1 Back-end

2 Front-ends

what i did

Project ownership

Product strategy

Research

Prototyping

UX & UI design

Team

Lead PM

1 Back-end

2 Front-ends

company

Elementary Data

Elementary Data

A B2B, open-source AI-driven data observability & governance platform that helps teams detect and resolve data-pipeline issues.

A B2B, open-source AI-driven data observability & governance platform that helps teams detect and resolve data-pipeline issues.

Context

Fastest path possible to top OSS to SaaS converting feature.

Fastest path possible to top OSS to SaaS converting feature.

The company was shifting all focus to align the major features for the upcoming SaaS release. I started with the Data Lineage: a visual node canvas that helps trace data issues and impact.

The company was shifting all focus to align the major features for the upcoming SaaS release. I started with the Data Lineage: a visual node canvas that helps trace data issues and impact.

impact

Conversion driver

Flagship feature

Positioned Data Lineage as the primary OSS to SaaS conversion driver, cited by 81% as the main reason.

69% faster

1.45x Reduction

Reduced cloud compute costs through optimizing Data Lineage core user flow.

Rapid adoption

Fiverr & Elementor

3 design partners converted to paid versions within weeks of private launch.

Community favorite

❤️

Repeatedly voted all-time favorite feature in monthly Slack polls.

the challenge

This is how the data lineage looked like when I joined.

This is how the data lineage looked like when I joined.

Adapting to a ~2 week timeline to deliver results and move on to the next feature, I prioritized speed, maintaining high-value yet focused strategy to deliver a unified solution.

Adapting to a ~2 week timeline to deliver results and move on to the next feature, I prioritized speed, maintaining high-value yet focused strategy to deliver a unified solution.

The default starting point was showing the entire data warehouse DAG, as is.

UX process

Unclear drop-off patterns and a very tight deadline.

Unclear drop-off patterns and a very tight deadline.

Event tracking showed high page-load volume alongside short session times, indicating drop-off patterns. The strategy: define success metrics → user research → hypotheses and solution.

Event tracking showed high page-load volume alongside short session times, indicating drop-off patterns. The strategy: define success metrics → user research → hypotheses and solution.

  1. Understanding the logic behind data-pipeline issue investigation (qual):

Interviewed ~12 data professionals (team leads, data engineers, analysts) at Elementor & Fiverr:

Most of the time, as an investigation starting point, they know which model they search for to begin with.

For exploration purposes, need easy navigation between assets.

Want to group models by type.

  1. Pinpointing the main pain points at high-scale, quickly (quant):

Surveyed our open-source user community on Slack (~5K members):

# data-lineage

Hi everyone 👋 We’re exploring improvements to the data lineage, how would you describe your current experience with it?

🙏🏻

1K+

❌

938

🚫

684

💔

473

🥲

341

Laggy, long loading times caused by rendering large DAGs.

Hard to navigate or locate nodes quickly.

Main learnings

The data lineage lacked an opinionated user experience.

The data lineage lacked an opinionated user experience.

Investigations are progressive

Usually, analysts trace paths from a known table to identify the root cause.

Table

Issue

Table

Root

Remaining within context

The current experience lacked tools to change the starting point during an investigation.

No for full visibility

Loading the full DAG by default is unhelpful and causes performance issues.

Table

Dashboard

Table

Table

Table

Dashboard

Table

Table

Table

Table

Table

Table

Table

Dashboard

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Dashboard

Table

  1. Understanding the logic behind data-pipeline issue investigation (qual):

Interviewed ~12 data professionals (team leads, data engineers, analysts) at Elementor & Fiverr:

Most of the time, as an investigation starting point, they know which model they search for to begin with.

For exploration purposes, need easy navigation between assets.

Want to group models by type.

  1. Nailing the main pain points at high-scale (quant):

Hi everyone 👋 We’re exploring improvements to the data lineage, how would you describe your current experience with it?

🙏🏻

1K+

❌

938

🚫

684

💔

473

🥲

341

# data-lineage

Laggy, long loading times caused by large DAGs.

Hard to navigate or locate nodes quickly.

Surveyed our open-source user community on Slack (we had ~5K members):

solution

Designing for progressive investigation, not full visibility.

Designing for progressive investigation, not full visibility.

1

Model-first journey

Model-first journey

Users select a model first, then see a focused DAG by default.

Users select a model first, then see a focused DAG by default.

2

File tree with search & filters

File tree with search & filters

Enables quick model location, matching the database structure.

Enables quick model location, matching the database structure.

3

Node filters, depth controller & direction stepper

Node filters, depth controller & direction stepper

Users can group by model type, choose levels to display, and toggle upstream/downstream.

Users can group by model type, choose levels to display, and toggle upstream/downstream.

4

Node-level actions

Node-level actions

Added contextual actions so users can trace issues without losing context.

Added contextual actions so users can trace issues without losing context.

5

Column-level lineage

Column-level lineage

Added visibility layer to follow column connections, a SaaS feature.

Added visibility layer to follow column connections, a SaaS feature.

Personal note.

Personal note.

Working on a true open-source product changed how I approach research. The Slack community gave unfiltered feedback in hours, not weeks. This direct access to 5K+ users shaped how I prioritize speed and specificity in user research.

Working on a true open-source product changed how I approach research. The Slack community gave unfiltered feedback in hours, not weeks. This direct access to 5K+ users shaped how I prioritize speed and specificity in user research.

Next project

Next project

© Adam Goddenyu. All Rights Reserved.

© Adam Goddenyu. All Rights Reserved.

© Adam Goddenyu. All Rights Reserved.