jasmine mithani


Feed the Baby

December 2015 - March 2016

goals and methodology for feed the baby

Feed the Baby was a very quick project put together to show an alternate use of digital games at DoSomething.org. My manager and I had been spending most of our time on Words Matter, our idea for a standalone product, and wanted to prove that games could fit into other organizational space.

We chose to revamp the "11 Facts About Teenage Pregnancy" webpage, which remained mostly static throughout the year with a call to action to a low-conversion campaign. This page was ideal for testing because of its steady stream of monthly traffic. We tested how adding interactivity to the page could impact users leaving the current page to sign up for a DoSomething campaign.


The first prototype of the game was built by my manager in one night after a stroke of insight. He shared the project with me, and then we coordinated next steps needed in order to make it ready for public release. We strategized to increase replayability, design for virality, and how best to push users to sign up for a campaign.

I coded a game feature that made the title baby throw up if fed too much, increasing the difficulty of the game by adding another lose condition. I also randomized the win condition, to give players a different experience every time they played. I helped design A/B tests to be executed on the desktop with Optimize.ly, experimenting with placement on the page and different start screens.

Research Methodology

Our A/B test environment was Google Analytics paired with Optimize.ly. We tracked several engagement metrics, including pageviews, number of times a user chose to play the game, number of times a user replayed the game, number of conversions (responding to the call to action), and number of Twitter shares.

We had two major classifications of tests: measuring conversions on the page with the interactive game versus the previous static one, and comparing differences in conversion rates depending on where the game was placed in the page. The second test was necessary because we had to embed the game and call to action in a page where campaign conversions were not the main goal. The primary purpose of the webpage was to provide research-backed facts about teenage pregnancy, and we had to make sure that our interactive element didn't deter that steady stream (roughly 10,000 views per month) of traffic.


<My suggestion to make the game visible when the user first views the page (as opposed to having to scroll down) resulted in the highest engagement and conversion numbers.

One of the placements we tested. We received the most interactions when the game was placed above the fold.

Adding the interactive game to the page resulted in a 542% increase in users responding to the CTA. The campaign that we designed the game for occurs once a year, and is now finished. Now that we have built our game, and it has been proven as an effective conversion tool, it can be easily reactivated on the DoSomething website during the annual campaign run. Most importantly, this project proved that gaming and interactivity were valuable tools to a social activism organization, and provided key deliverables to support a new, experimental venture.