Making a website findable is critical nowadays, so web analytics tools that monitor website traffic have become key online marketing tools. Our aim is to assess the changes over time in the performance of a website using Google Analytics data and within a statistical framework that enables us to test the effectiveness of online campaigns. A time series econometric methodology is applied based on Structural Time Series Models that take into account the non-stationary behaviour usually displayed by socioeconomic time series. The results show that online campaigns affect traffic volume positively but their effect on its quality is ambiguous. Analyzing paid and unpaid traffic separately, we find that the increase in traffic volume is not always due to the paid keywords and that the lowest quality visits come from paid traffic. This analysis may help webmasters to design successful online advertising strategies to enhance the competitiveness of their firms.