As @weaselhouse noted in his response, self-policing is counter productive as it imposes a set of rules onto Steemit which the platform itself does not uphold or expect users to abide by.

For example, many users complain about self voting abuse but it's permitted by the platform. So until Steemit stops allowing it, then complaining about it is inconsequential. The founders themselves acknowledged its propensity as a feature to be abused in the Steem White Paper, where it's said: "...each individual voter has incentive to vote for themselves at the expense of the larger community goal." And here we are today, two years later and self voting is still an issue.

In reference to @grumpycat's initiative, it appears to be predicated upon a desire to root out users that engage in reward pool abuse, with an emphasis on targeting those users and voting services that seek/offer upvotes 3.5 days after a post is published. The problem with this is that Steemit itself allows upvotes on posts up to day 7. And there are some voting services that are manually operated and thus have backlogs - a user may pay for a vote on day 2 but not receive it until day 4 or 5, for example.

As the saying goes, "Don't hate the player, hate the game." There are no doubt users that play the game here, but they are doing so within the rules of the platform. So for wide reaching and constructive changes to be implemented, compliance initiatives such as what Grumpy Cat is engaging in are better enforced by being coded into future iterations of Steemit itself. In this manner, the updated rules of the game apply to all users equitably. And that requires buy in from Steemit Inc and the Witnesses. That is who Grumpy Cat and others should be engaging for reform.