Every company I meet today has a data platform. And if they don’t have one, they want one. Problem is that building and maintaining a data platform is not trivial. First, multiple tools need to be integrated together: Airflow, Spark, Presto, Kafka, Flink, Snowflake, and potentially many more, but, more importantly, a dedicated engineering team must be setup to maintain and make sure everything runs smoothly. And, what usually happens is that, after data has been accumulated for months and months, the cost of running such infrastructure is higher than the benefit.
So the question is: do you really need a data platform ?