SourceCon Hackathon Fall 2024: Strategies and Lessons from the Finalists

Finalists from a recent sourcing hackathon, along with one of the people who built the challenge questions, walk through what actually happens during a timed competition against some of the toughest research puzzles in the industry. The conversation covers the tools that worked, the traps that didn't, and the mental discipline it takes to stay fast without going off track.

Key points covered include:

• Reading the full question before searching, since overthinking and chasing extra detail wastes time when the answer is often simpler than it looks.
• Using AI tools like ChatGPT as another search resource, including a case where prompting it with "translate:" unlocked an answer a direct question couldn't get, and another where rewording a prompt got around a supposed AI blind spot.
• Treating unconventional sources such as the Wayback Machine, Instagram captions, patent and academic databases like PubMed and Google Scholar, and niche certification directories as legitimate paths to an answer, since challenge questions are often built around resources most sourcers never think to check.
• Guessing strategically when a full answer isn't available: piecing together partial information such as a graduation year found on social media, since there's no penalty for wrong submissions and points are awarded only for the final answer, not the method.
• Using outside tools or custom-built search tools when available, since relying on a resource someone built themselves or a platform's hidden functionality isn't considered cheating.
• Managing pressure with practical habits: closing tabs and starting fresh after each question, working with music or a podcast to stay focused, and having someone nearby who isn't part of the competition to help stay calm after a mistake, like accidentally closing the answer-submission tab near the end.
• Recognizing that challenge design keeps shifting as sourcers get better at platforms like Instagram or TikTok, pushing question writers toward newer or more obscure territory each time.

The overall picture is less about knowing every database by heart and more about staying calm, thinking sideways when a direct search fails, and accepting that a wrong guess costs nothing while a blank answer costs everything.