At me&u, we amassed a wealth of customer data but had room to further leverage it for actionable outcomes. During our inaugural hackathon, our team (myself, our Head of Product, Henry Hill, and Senior Full-stack Developer, Nadav Oran) sought to utilize existing purchase data to increase average transaction value.

Strategic Considerations

Key insights guided our approach:

• Upsell recommendations must integrate seamlessly into the ordering flow to minimize friction.
• Suggested items should focus on low-cost, impulse purchases.
• Merchant management needed to be minimal for scalability. This feature should 'just work' with intelligent, data-driven automation.
• Recommendations must be timely and contextually relevant, such as aligning with our pre-defined time-of-day menus.

‍

Development

The two-day hackathon imposed strict time constraints, requiring parallel design and development efforts. Initial discussions aligned our tasks and acknowledged potential bottlenecks to ensure we stuck to the deadline. We established these parameters:

• Front-end changes should prioritize inline modifications to avoid introducing new modals or pages.
• Daily batch data processing with overnight updates was an acceptable solution.
• Merchant configuration would be limited to a feature toggle, reducing onboarding complexity.

‍

‍

Beyond the Hackathon

Despite not securing a hackathon win, our concept demonstrated compelling potential. 'Quick Sells' found a path to production during a quieter development period. Merchant response was overwhelmingly positive, resulting in measurable uplift in average spend throughout our network. This initial success also generated numerous feature enhancements which we iteratively shipped in the following months.

‍

‍

Impact

Within a month of deployment, 'Quick Sells' delivered a notable 28% increase in average transaction value. Over the course of a year, it proved instrumental in adding over $1M in revenue across our merchant partners.