This varies SO MUCH, but I can try to generalise.
Cost of sending people vs reach/impact we would get is a big factor. Tiny events with a homogenous audience that are very far away (I was in Stockholm, Sweden, so for me this meant a lot of single-university hackathons in the US), were usually off the table. I highly recommend looking for local sponsors where possible!
We also looked at how hands-on people were likely to get with our APIs. If I can run a workshop to show people the ropes and then run a challenge to use the APIs, it’s more likely that hackers will use our stuff and we’ll get more value out of it.
Diversity was a large factor for us as well. We hosted a couple women/nb in tech events in our offices and sponsored Nonbinary in Tech.
Things we didn’t care about were getting resumes or having recruiters around. This would be different for companies where the recruiting/TA org does sponsorship, but in our case we were in engineering and more focused on getting people using our stuff.
Most unique, though questionably useful, was in a music hackathon where we had access to a VIP room where recording artists were competing in some sort of song competiton. Also Ninja (the streamer) dropped an album…it was a weird event.
Some of the best ones were proposing our own challenge, and running workshops (I hecking love workshops, exhibit a, exhibit b). When in doubt, ask the sponsor what would be cool for them!
Spotify works on a lot of cool problems and builds a product I use daily, which was awesome. It’s also a very product-driven and agile org, and I ended up learning a ton about product management for consumer-facing products while there.
I moved on to a newly formed game studio which was just too cool to turn down! I would have also liked to see Spotify invest more in developer relations and 3rd party developer strategy - there is a lot more I would have liked to do if I had the chance.