Retaining Google for multiple 6-figure renewals
May 2024 - August 2025
The problem
Our team and I needed to prevent Google from churning by delivering a series of improvements to Workflow (Ansarada’s task management product) to retain them for FY24-25.
Backstory: In July 2023, Google's corporate development team used a customised version ofWorkflow our team and I built for their team to run targeted acquisitions. However, due to internal resource shifts in Ansarada, the MVP was left unimproved until early 2024, jeopardising Google’s renewal for FY24-25.
The second priority of this project was to scale our targeted acquisition solution built for Google to reduce the increasing churn rate of our other customers running acquisitions on Ansarada.
Outcome and impact
My design helped secure Google’s renewal for FY23-24 and, also for FY24-25.
The release has also improved the overall usability of Workflow, with the biggest design change launched in July (Auto-move of documents) eliminating ~17% of setup clicks from our 2023 MVP.
The latest experience can be found in the section below titled Latest design: 2024-25 solution.
Lastly, since the latest updates in July, the selection-to-conversion rate increased from 8% to 13% for the targeted acquisition use case, suggesting improved product value fit for customers.
My design process
Diagnosing Google’s pains
To understand Google’s pain points, I led two discovery interviews with Google. These sessions revealed 13 pain points ranging from minor frustrations to blockers that slowed initial setup.
To create clarity, I mapped Google's pain points across their set up journey. I identified it took approximately 317 clicks and steps before they could even begin their RFI process.
This system map helped me align product, engineering, and leadership on the pain, scale and urgency of the problem. Most importantly, it created shared understanding that UX improvements needed most at setup.
Design validation with Google
Using the pain points uncovered in our discovery sessions and system map, I created low-fidelity concept designs to validate possible solutions with Google and clarify uncertainties.
These designs became critical for ongoing alignment—not just with Google, but also internally with stakeholders like our CFO, who had experience in targeted acquisitions.
We ran continuous feedback loops with Google, refining areas like security, commenting, and document handling within Workflow. Today, we maintain a quarterly update cadence to share progress and ensure our solution continues to meet Google’s evolving needs.
Scaling beyond Google
In between our design updates with Google, our team and I validated the design with 5 other customers using Ansarada for targeted acquisitions.
These sessions proved that our solution was desirable and could help other customers beyond Google.
Importantly, their feedback were important data points to guide our product roadmap and priorities.
What we delivered for Google
Below is the latest Workflow design for FY24-25 that has ~80% of the previous setup friction and improves usability for Google and other customers, the details of the key changes are:
1. Automatic move of documents (BTS): Documents now automatically move on upload from the target once the tab is shared, eliminating 17% of the previous friction.
2. Enhanced usability: We reduced the font size and introduced text wrapping in Workflow to improve information density and surfaced comments in Workflow to improve collaboration.
3. Integrated access (BTS): We integrated the access settings from Workflow to the Document index, eliminating the previous setup pain of requiring Administrator’s to do it manually, reducing, another ~60% of setup friction.
Where we started: 2022-23 MVP.
At a glance, it resembles our new version of Workflow for targeted acquisitions but on closer inspection, it lacks the improved information density and other improvements.
In the below GIF, I show how our customers or Google would have activate ‘auto-move’ manually per tab in our MVP. In our newest version of Workflow, this process is completely automated.
