Why is it that some care order cases result in the child being removed from parental care, while in others she is not, despite the cases being similar? This paper investigates how decision-makers reason and justify different outcomes for similar cases, by an analysis of four pairs of judgments (from Norway, Estonia, and Finland) about care orders, using thematic analysis. The comparison is within the pairs and not across countries. I find that the variance in outcome and reasoning seems to be a result of discretionary evaluations: risk, cooperation of the parents, and the potential of services to alleviate the situation are interpreted differently in the cases and lead to different outcomes. This appears to be a legitimate use of the discretionary space available to the decision-makers. The decisions are justified with ‘good reasons’ mostly related to threshold, the least intrusive intervention principle, and the best interests of the child. Such justifications are suitable to provide accountability and legitimacy, but the reasoning is at times lacking transparency and thoroughness. The reasoning is longer in the non-removal cases, suggesting that more thorough reasoning is required when the decision-makers depart from the most common outcome.