A retail analytics company, DataNova Inc. , is launching its first dynamic holiday campaign. The campaign serves personalized offers by analyzing real-time data, which grows by terabytes every hour during a 30-day period. Their data scientists use Google Cloud Dataflow to preprocess the data and store extracted feature signals in Google Cloud Bigtable for use in machine learning models. However, the team notices poor performance when reading and writing the initial 10 TB of data. They want to enhance Bigtable performance while keeping costs low. What should they do?