(feat) Targeted Boost
Summary
One of the I've heard is that often good content in a specific niche gets a poor return with boosting in its current setup and people just opt out of doing it entirely. What if when boosting, people were presented a choice of either doing it network-wide or targeted - appearing only to people who have explicitly set up specific hashtags in their discovery?
I know there is a proposal in the issues where it was suggested to give users the option to limit their boost feed to specific hashtags. However, I believe the approach that I'm suggesting is a superior way of addressing that same point because it follows that when content creators see their boosts are more effective when boosting to relevant hashtags, much of the content will drift that way automatically without intervention.
Final caveat - keeping it transparent to mitigate manipulation and echo chambers:
If implemented, I urge that there is a strong need for whenever a targeted boost shows up in someone's feed, that there is a visual indication of the hashtag at that moment to be seen by the receiver.
For example:
- Someone boosts explicitly to #gardening, #seeds, and #urbanfarming.
- The user receiving the boost has configured their discover to find content in #blockchain, #gardening, #horses, and #urbanfarming
- There is a moment where this particular boost was matched to a user based on the common element of #urbandfarming.
- That #urbanfarming hashtag should be written somewhere in the boost header to show that this was targeted to them because of this hashtag.
Why is that important?
- It is consistent with the paradigm shift of transparent algorithms and empowers the user to be conscious of why something is appearing in their boost feed where they can always have the option to tweak their discovery settings.
- It lays bare for all followers of a hashtag to see instances where something completely unrelated to it appears to them in bad faith. Negative user response seeing this would disincentivise this from happening because receivers of the boost are empowered with the information that allows them to see that that the targeted boost is being abused.
Final thoughts:
I think this would be extremely effective and superior to existing proposals (and any possible ML solutions) because both the sender and receiver at different moments have explicitly selected their hashtags. It's a win-win, and as the network grows would likely be extensible to languages and geo-targeting (once again explicitly set up by the boost sender[chosen at boost] and receiver[if location is on profile]) which would be beneficial for local community building as well.