One of the curious phenomena in political activism (as famously lampooned in the scene linked below from Monty Python's "Life of Brian") is that political activists often reserve their greatest criticism for their natural allies, rather than their natural enemies. As the linked sketch suggests, this behaviour appears to be quite irrational. However, it is possible to justify this behaviour as being rational over the short-term, even if it fails to be rational in the long-term.
Here is a simple model to illustrate this. Suppose there is a political issue X, on which one can take just one of two positions: pro-X and anti-X. (In the example below, X would be "Judean independence".) For sake of this discussion, we assume that there are no intermediate or neutral options available. Given a rational actor A, how would A choose which of these two positions to take? One can postulate a payoff function P( A, p ) which specifies the payoff (or "utility") that A would obtain from taking position p, thus P( A, pro-X ) is the payoff A would get from declaring to be pro-X, and P( A, anti-X ) would be the payoff A would get from declaring to be anti-X. A rational actor A would then take the position that offered the larger payoff. For instance, for the Roman authorities in the example below, P( A, anti-X ) would presumably be considerably larger than P( A, pro-X ) (as long as the level of unrest was manageable), whereas for rebels P( A, pro-X ) would presumably be larger than P( A, anti-X ) (as long as the level of suppression by the authorities was not too great).
Now, suppose there were several actors A for which the payoffs P( A, pro-X ), P( A, anti-X ) were close to equal. A pro-X activist could then attempt to tip the balance for these actors to join the pro-X camp, by actively criticising (or boycotting, etc.) those actors who decided to choose anti-X, thus effectively lowering the payoff P( A, anti-X ) for such actors to the point where it would be rational for A to choose pro-X instead. On the other hand, for actors A for which P( A, anti-X ) was significantly larger than P( A, pro-X ), such criticism would be substantially less effective, as it would usually not be enough to tip the balance in one's favour. Hence, the rational use of resources (in the short term) for a pro-X activist is to focus one's attacks on "moderates" that already have significant pro-X sympathies, rather than "natural enemies" that are strongly in the anti-X camp.
However, despite this strategy being rational in the short-term, it is detrimental in the long-term, because it creates a disincentive for strongly anti-X actors to moderate their anti-X position. One can come up with many examples in real life in which an anti-X actor tentatively introduces a somewhat pro-X policy, but then comes under attack, not just from anti-X activists, but also from pro-X activists for not going far enough. Perversely, the long-term effect of pro-X activism is then to encourage anti-X actors to simply remain anti-X and not attempt any pro-X initiatives whatsoever, so that pro-X activists do not even bother to attempt influencing them via boycotts or other negative actions. Another unintended consequence is that long-term-oriented pro-X activist groups may begin to criticise more short-term oriented pro-X groups for sabotaging their long-term cause.
The moral here is that in order for activism to be successful in the long term, it has to pressure "unpersuadable" actors at least as strongly as "persuadable ones", lest one create perverse incentives.