Capacitive Deionization (CDI) is an energy-efficient desalination technology that utilizes an electric field to extract ions from water. Flow-through CDI systems show potential for superior desalination performance compared to traditional flow-by CDI; however, they face the challenge of increased occurrence of Faradaic reactions, leading to undesired by-products and reduced energy efficiency. In this study, we constructed a flow-through CDI cell and investigated the desalination performance of the two possible cell configurations: upstream anode mode and downstream anode mode. A series of experiments were conducted, measuring conductivity and pH of the effluent solution during charging and discharging phases. The results were analyzed in terms of salt adsorption capacity and charge efficiency. We used pH fluctuations in the effluent solution as indicators of Faradaic reactions. It was found that upstream anode mode yielded superior desalination, with a salt adsorption capacity of 6.79 mg g^-1 and charge efficiency of 64.3%, compared to downstream anode mode, which displayed a salt adsorption capacity of 5.19 mg g^-1 and charge efficiency of 50.8%. However, upstream anode mode also produced more pronounced pH oscillations, suggesting a higher occurrence of Faradaic reactions. Reconciling these conflicting results and shedding light on the complex processes within the CDI cell calls for further investigation.