For electoral accountability to be effective, what is better: a fragmented opposition that gives plenty of alternatives or an opposition with fewer stronger parties? While oppositions play a key role in democracies, the effect of their fragmentation on accountability remains unclear. On the one hand, the more the opposition parties the easier for voters to find ideologically close alternatives. On the other hand, high levels of fragmentation can also make accountability more difficult because in these contexts voters may struggle to find and coordinate over viable alternatives. Thus, fragmentation of oppositions generates two opposite effects that have arguably led to inconclusive empirical results so far. To solve this puzzle I first analyze whether the effect of oppositions’ fragmentation on accountability may follow a non-linear pattern, that is, whether fragmentation may initially foster accountability while making it more demanding when many alternatives exist. Moreover, I distinguish between ideologues and pragmatic voters and differentiate ideological and valence issues to assess the specific conditions under which fragmentation may enhance accountability. I test these hypotheses using a survey experiment. In the experiment, voters were presented with a case of a government accused of corruption, and were asked to express their willingness to support or punish it. I randomly manipulated the framing of the scandal and the degree of opposition fragmentation. The results show that fragmentation clearly enhances accountability and that this effect is significantly larger for ideological voters.