Abhishek Choudhary

Analyzed large datasets to find the right features of the page to prioritize (since bytes are at a premium).

The Amazon portfolio of widgets is in the thousands, and I could only select a handful for this project to be successful. Jointly designed (with engineering) a rendering pipeline that can significantly prepare, render, and flush the prioritized part of the page. I created a roll-out strategy across product lines and user types to mitigate risk and maximize the returns from collected data, eventually crystallizing into a worldwide release strategy.

The Amazon detail page for the vast majority of products is now faster to respond to user interactions by multiple seconds in most instances, which has resulted in a much better experience for the customer in regions where lots of data and computing power are still at a premium; additionally, Amazon is now able to tap into tens of millions of dollars in latent revenue.

Testing non-public websites is a USP for any web-testing platform, and BrowserStack enabled this with a CLI application called BrowserStack Local. However, that application had a very high failure rate due to user error, given the technical nature of the work and the lack of good UX guardrails—up to 75% of users failed to use the application on the first try.

With data analysis of the largest pain points, we developed an MVP for a GUI dashboard within the application to help users cross the hump. We were able to reduce the onboarding friction almost entirely.

As a mature product, Automate by BrowserStack suffered from the issue that a single new feature was not going to help significantly improve NPS, which had been stagnant for some time. With a mixture of rigorous data analysis, customer interviews, and primary usage, I identified and prioritized a set of technical and UX optimizations to solve customer issues. For example, creating fallback-to-default workflows in incorrect or incomplete inputs for testing parameters and indicators of potential customer errors during test setup. Cumulatively, our NPS number moved by 10% across the quarter after being stagnant for two years.

As a market leader in the web and app testing space, BrowserStack was hitting a glass ceiling—a lot of the customer dissatisfaction with BrowserStack was dissatisfaction with the state of the testing market itself. In order to efficiently change the direction of the entire market, we needed to be able to set the direction of the market and best practices outside of our paid products.

As the open source product management lead for BrowserStack, I was managing BrowserStack Automate as well as Selenium and Nightwatch.js. Using this vertical integration, I created the open source strategy for BrowserStack with a combination of acquisitions and new feature development to help uplift the state of the testing market and improve the workflow for all users interested in testing better. This includes everything from onboarding to testing tools, reporting, analytics, and advanced features like self-healing tests.