There are numerous observational studies on intraspecific variation in avian nest building and a single experimental manipulation. The general consensus is that birds build nests in response to environmental conditions, but it is not clear whether such flexibility in nest building is reproductively advantageous. To test the relationship between building flexibility and reproductive success, we allowed captive zebra finches to build their first nest, using string, and to breed in temperature-controlled rooms held at 14 or 30 °C. Once the offspring had fledged, we returned half the pairs to breed at the same temperature while half the pairs were switched to the alternative temperature. We provided all pairs with string and left them to build and breed a second time. For their first nest, pairs that built at 14 °C used more string than did pairs that built at 30 °C, and pairs that bred successfully built a nest with more string than did unsuccessful pairs. When pairs built their second nest, however, temperature no longer explained the number of pieces of string they used; rather, irrespective of the ambient temperature, pairs that had successfully produced young from their first nest used the same amount of string for their second nest, whereas those that had failed to reproduce with their first nest used more string. These latter pairs were then more likely to reproduce successfully. Ambient temperature, therefore, did affect the nest the pairs built but only in the absence of reproductive experience.