This study investigates the effectiveness of concordance feedback on EFL learners’ writing. The research questions are as follows: (a) Is the concordance feedback effective for improving accuracy in the learners’ English writings?, (b) How do the participants use concordance feedback to correct errors in their writings?, and (c) What are the participants’ perceptions on concordance feedback? Using mixed methods research, this study measured the effectiveness of concordance feedback quantitatively and examined how the participants used and perceived it qualitatively, in an effort to support the quantitative method results. The researcher made online concordance information accessible to 42 intermediate EFL graduate students in China, so that they could correct their writing errors by using this concordance information attached to their writing errors as feedback. To measure the effectiveness of concordance feedback on the accuracy improvement in the EFL learners’ writing, the error rates from the pre- and the post-writing tests were compared by using a paired-samples t-test. To examine students’ self-correction process, correction strategies, and perceptions on the concordance feedback, this study used NVivo for qualitative data analysis, including error analysis forms, open-ended questionnaires, interviews, and classroom observations. The findings from the quantitative data analysis were (a) the concordance information as feedback on sentence-level written errors had a significant effect on writing accuracy improvement, and (b) most errors were successfully corrected by the participants themselves. The qualitative data analysis revealed students’ self-correction process and correction strategies. The students’ self-correction processes involved phases of “perceiving error”, “guessing”, “noticing input”, “comparing”, and “understanding error”. Since the process of error correction triggered inductive learning, the result implies that the concordance feedback approach provided students with an optimal environment for intake facilitation. In terms of error correction strategies, this study revealed that accuracy improved students tend to use cognitively demanding strategies (e.g., “analogy”, “autonomous corrections”, “accepting but changing”, and “creative correction”) than accuracy unimproved students. More importantly, it showed that “complex feedback” seemed to provide an environment for correction failure because it apparently increased the level of complexity to students’ correction process. The qualitative data analysis also revealed the participants’ perceptions of the concordance feedback. The participants found that the concordance feedback: (a) provided them with a useful resource for authentic language input, (b) motivated them to pursue more authentic English, (c) led them to become confident in correcting their own errors, and (d) helped them take more active role in their learning. The findings reported here may be seen as an indication that using concordance feedback may provide a convincing way of introducing concordance-based methods to new EFL corpus users and a motivating way of preparing students for independently using a corpus technology.