Analogical reasoning is essential for transfer by supporting recognition of relational similarity. However, not all analogies are created equal. The source and target can be similar (near), or quite different (far). Previous research suggests that close comparisons facilitate children's relational abstraction. On the other hand, evidence from adults indicates that the process of solving far analogies may be a more effective scaffold for transfer of a relational strategy. We explore whether engaging with far analogies similarly induces such a strategy in preschoolers. Children were provided with the opportunity to solve either a near or far spatial analogy using a pair of puzzle boxes that varied in perceptual similarity (Experiment 1), or to participate in a control task (Experiment 2). All groups were then presented with an ambiguous spatial reasoning task featuring both object and relational matches. We were interested in the relationship between "near" and "far" conditions and two effects: (a) children's tendency to spontaneously draw an analogy when solving the initial puzzle, and (b) their tendency to privilege relational matches over object matches in a subsequent, ambiguous task. Although children were more likely to spontaneously draw an analogy in the "near" condition, those who attempted the far analogy were more likely to privilege a relational match on the subsequent task. We argue that the "process" of solving a far analogy--regardless of a learner's spontaneous success in identifying the relation--contextualizes an otherwise ambiguous learning problem, making it easier for children to access and apply relational hypotheses.