There is a lot of good ideas here on how to increase voter turnout. I would like to also bring to the table the Nexus Mutual governance system(contracts on implementation are on github) which I have participated in the past and also find highly effective. A quick primer on how this functions is below:
All proposals put to a member vote are firstly put to the Advisory Board (described below), who will white-list the proposal and make a recommendation on the outcome. Members are then given a specified time-frame to vote on the proposal. If a specified quorum is met then the majority outcome prevails, otherwise the vote proceeds as per the Advisory Board recommendation.
As a reward for participating in the governance process all voting members will be awarded a share of additional NXM tokens that are generated for each proposal (chosen by the Advisory Board within limits). Additionally, any member may delegate their vote to any other member, allowing them to rely on a fellow member they believe has more expertise than themselves. In this case, the delegating member continues to earn all the voting rewards.
This model ensures that proposals which are important to the platform are actually passed even if voter turnout (quorum is not met) is abysmal. It is very rare for proposals to be decided by the advisory board since most people do end up voting as they get rewarded for participating. For those that don’t have time to vote or are not keeping up with proposals, they end up delegating (this does not mean that you are giving your tokens away) their voting power to someone else but also reap the benefits of earning voting rewards as long as the person that they delegate to also ends up voting.
The propose option to boot inactive voters would help us significantly in getting proposals passed as those folks would no longer count towards the quorum numbers. With that said, I am afraid that any option that would require us to redesign governance (delegating, expunging inactive voters, etc…) would also require modifying the governance contracts and that means time spent by developers that could be spent elsewhere (e.g. improving Bancor).
Ultimately, I think that the path of least resistance is probably to incentivize voting (and is what we should focus on). Implementation wise, this probably would require the least amount of work from developers and would allow us to gauge how effective it is in a quick manner (we can easily stop the campaign if it is not helping). If we do find that this is highly effective, then everything else might be moot.