One of the key elements in the Snowplow technology stack is the Hive SerDe. This is what makes it possible for Elastic MapReduce to read the Cloudfront log files generated by the Snowplow javascript trackings tags, extarct the relevant fields and make these available in Hive as a nice, clean query table. (The structure of the Hive table is documented here).

A number of improvements have been made in the new versions. However, the most significant is that the 5 utm_marketing fields have been added, so that campaign attributes are now available for analytics.