Since launching in December, API traffic to Airtable consisted solely of GET requests, with performance averaging below 600ms and the slowest 95th percentile at around 1100ms, factoring in an expected 75ms latency due to geographical distance between AWS's us-west-2 and Airtable's us-east-1. Recently, Sequin introduced a proxy allowing customers to make write requests (PATCH/POST/DELETE) to both Airtable and Sequin databases, which revealed performance issues, particularly with bases containing many computed fields. Benchmarking tests on bases of varying sizes (1k, 10k, and 49k records) indicated that the 10k base, which contained numerous Formula and Lookup fields, experienced significantly slower write times. Further analysis showed that these computed fields consistently added overhead to requests, and their complexity, especially when interdependent, correlated with increased service times, which were exacerbated by larger base sizes. The findings suggest that complex computed fields can dramatically increase write times, making it advisable for highly transactional bases to consider alternative solutions like automations or scripts. The study also noted occasional outlier requests taking up to 30 seconds, warranting further investigation.