Are bounties on Stack Apps productive?

Question

Stack Apps doesn't have a Featured tab like other SE sites. Is starting a bounty to promote your Stack App a stupid idea?

Answer 1

My opinion is that they probably do not help, but that could change if you post a large enough bounty and are somehow able to "advertise" it (maybe in chat) to enough people that can answer (or upvote your app) and who are not regulars on Stack Apps, but willing to join the site.

There are very few people who regularly visit Stack Apps. Of those, I estimate that maybe 5, or so, will answer a general question with any regularity. And frankly, would probably answer it anyway, if they thought it would help -- no bounty needed.

Criteria to measure Bounty performance:

What positive effects would a bounty ideally have?
Here are some possible benefits:

1. Increase the number of views to a question.
2. Increase the number of answers on a question.
3. Increase the number of "new" users viewing/answering a question.
4. Increase the odds of an accepted answer.
5. Awarding full bounty to the answer the bounty provider most likes.
6. Awarding partial bounty to a well-received answer if the provider does not make the award.
7. Increase the number of votes on the question and its answers (often a function of higher viewcount in a short amount of time).

What does the data say about Bounties on Stack Apps?

See this SEDE query. It shows:

1. For the entire existence of Stack Apps, there have been only 23 bounties offered.
2. One of those is on a deleted question, so further information on it is limited.
3. Only 7 of the remaining 22 have an accepted answer.
4. 5 of the questions have no answer.
5. 11 of the questions have only 1 answer.
6. Only 10 had the bounty awarded (so 12 had no "worthy" answer added by the bounty).
7. One of those bounty awards was not for the full amount -- suggesting the award was automatic and the bounty provider was not happy.
8. 3 of the bounties were awarded in 3 days or less. This is usually to "reward" existing answers rather than to seek new ones.
9. The average question score was 12.1, but only about 3.9 for questions opened since 2013.
10. The average view count was 770, but only about 358 for questions opened since 2013.

---

Note: Those view and vote figures should be calculated relative to non-bounty questions and to how many days old a question is and what date the bounty was placed. I'll leave all that as a future exercise.