TabSplit is a web app I built as a direct alternative to Splitwise. I had always found the standard bill-splitting apps to be plagued by confusing UIs and highly unintuitive setup processes. I wanted to create a tool that made group ordering and settling up genuinely frictionless.
Building an app in a controlled environment is easy; deploying it to a table of ten people trying to split a restaurant bill is hard. Once TabSplit hit real-world usage, multiple edge cases immediately surfaced.
What happens when the network drops? What do you do when a friend at the table doesn't have their phone on them? The interface had to accommodate manual overrides, allowing the "initiator" to manually add people to the tab who weren't actively using the mobile app.
Testing the app in the wild brought in a continuous stream of feedback, particularly around how information was presented and manipulated. I had to rapidly iterate on the UI to solve friction points:
I learned a massive amount from this project. It proved that UI problems cannot be fully solved in a vacuum—they have to be tested against the impatience of hungry people trying to figure out who ordered the extra fries. By embracing multiple rounds of field feedback, TabSplit successfully streamlined group ordering and optimized an inherently messy social interaction.