Yusp data-flow management

We transformed Gravity's script-based recommendation engine into a visual drag-and-drop editor, cutting integrations from months to minutes.

services
arrow_forward
services
arrow_forward
services
arrow_forward

/ The Pain

Gravity had a strong recommendation engine, but delivering it was slow and costly. Enterprise integrations could take more than a year, while smaller clients were never worth the engineering hours.

Data teams had to wire everything manually with scripts and SQL, turning the system into a “black box” that few could understand. This not only frustrated clients but also limited Gravity’s ability to scale its product.

/ Co-Creation

We began with interviews and workshops that involved Gravity’s business leaders, analysts, developers, data engineers, and consultants. Together, we mapped how recommendations were integrated and broke down the jobs to be done.

We created a shared design language, dictionary, information architecture, wireframes, design kit, and high-fidelity screens. This co-creation process gave everyone the same view of the problem. It became clear that the old way of integrating and managing personalization could not scale and needs to change.

/ The Breakthrough

The solution was clear: replace scripts with a visual tool. We designed an editor that functions like a whiteboard, allowing you to drag, connect, and design data flows easily. It made collaboration between engineers, analysts, and business leaders seamless. What once took months of manual coding could now be done in minutes, making the system transparent and collaborative.

workflow mapping

concept validation

co-creation

design system setup

rapid prototyping

design reviews

roadmap development

go-live

/ Testing

Before launch, we focused on validating how people interacted with the new visual editor. Instead of testing basic whiteboard usability, we looked at whether users could confidently design and connect data flows. We set Illustrator, Shaderforge, Blender benchmarks. The in-house data team became early adopters, using the first prototype and giving feedback that directly shaped the roadmap. What started as a test quickly turned into a co-creation process with the team itself.

/ Adoption

Once the prototype was live, the challenge was no longer just technical. Success depended on convincing client-side professionals to adopt the tool and champion it inside their organizations. By highlighting clear benefits, such as faster integrations, easier control, and visible results, we turned skeptics into advocates. Even those who had preferred command-line scripts became some of the strongest supporters of the visual editor.

/ The impact

The new editor transformed how Gravity delivered personalization. What used to take months of engineering effort could now be done in minutes, reducing integration time and cost for both sides. This shift opened the door to serving smaller clients profitably while still supporting enterprise-scale projects. It also created a scalable model where data teams could own and manage flows themselves without relying on expensive, manual engineering work.

check

Integrations cut from months to minutes

check

Collaboration between engineers, analysts, and leaders

check

Visual drag-and-drop data flow editor

check

Scalable and cost-efficient personalization

check

Transparency replacing “black box” processes

/ Learnings

This project taught me the importance of winning internal champions for a new product. Not every stakeholder initially believed in a visual solution, especially those accustomed to command-line tools, but the right workshops and involvement ultimately turned them into strong advocates.

I also learned the value of co-creation: the best ideas came when developers, analysts, and designers shaped the product together. Finally, the experience pushed me to move beyond design into product strategy, leading me to coordinate the roadmap as a product manager.