According to an aspect of the present inventive concept there is provided a method for identifying a cause for a failure of a computer-implemented test performed on a latest software code revision in a sequence of a plurality of software code revisions, the method comprising: identifying a first revision of said sequence of revisions for which said test passes and a second revision of said sequence of revisions which is subsequent to the first revision and for which said test fails, determining a first change between the software code of the first revision and the software code of the second revision, generating a first modified version of the software code of the latest revision using the determined first change, determining whether said test passes or fails when performed on the first modified version, in response to determining that said test fails when performed on the first modified version: identifying a third revision and a fourth revision of said sequence which are intermediate the latest revision and the second revision, wherein the third revision is a revision for which said test passes when said revision is modified using the determined first change, and wherein the fourth revision is subsequent to the third revision and is a revision for which said test fails when said revision is modified using the determined first change, determining a second change between the software code of the third revision and the fourth revision, generating a second modified version of the software code of the latest revision using the first change and the second change, determining whether said test passes or fails when performed on the second modified version, in response to determining that said test passes when performed on the second modified version, providing a message indicating at least the fourth revision as including a cause for said test failing when performed on said latest software code revision.