Predicting the click-through rate of an advertisement is a critical component of online advertising platforms. In sponsored search, the click-through rate estimates the probability that a displayed adver- tisement is clicked by a user a er she submits a query to the search engine. Commercial search engines typically rely on machine learn- ing models trained with a large number of features to make such predictions. is inevitably requires a lot of engineering e orts to de ne, compute, and select the appropriate features. In this pa- per, we propose two novel approaches (one working at character level and the other working at word level) that use deep convolu- tional neural networks to predict the click-through rate of a query- advertisement pair. Speci cally, the proposed architectures only consider the textual content appearing in a query-advertisement pair as input, and produce as output a click-through rate prediction. By comparing the character-level model with the word-level model, we show that language representation can be learnt from scratch at character level when trained on enough data. rough extensive ex- periments using billions of query-advertisement pairs of a popular commercial search engine, we demonstrate that both approaches signi cantly outperform a baseline model built on well-selected text features and a state-of-the-art word2vec-based approach. Fi- nally, by combining the predictions of the deep models introduced in this study with the prediction of the model in production of the same commercial search engine, we signi cantly improve the accuracy and the calibration of the click-through rate prediction of the production system.