Depends on whether you're at the stage where you think you can gain statistically significant insights from anything or if it's just a distraction from talking to customers.<p>Data is very much an organisation problem first, a tech problem second, you'll have to prove to yourself and others around you that you can gain actually statistically significant insights before you invest hours and dollars. So it's in everyone's interest to get "good" working fast and cheap before you worry about perfect.<p>I'd start with one of the following:<p>1) Build: A quick audit table for events<p>2) Buy: Start with segment's free plan, if you're early stage it'll give you enough time to delay more expensive decisions. If you find you're approaching the limits fast, and you're getting actual value from the data, it means you got to make your mind up quicker. Send your segment data to analytics software of your choice. If you want to join it against your application data then you can warehouse it to your DB (but best to do your querying from a replica).<p>If you find you're getting value from simple joined SQL queries, then you can think about visualisation software, plenty out there to buy (domo, looker, powerbase etc), but you can also self host something like metabase. That way you can share the value with others and prove the worth of the data by having shared queries and dashboards.<p>If all of that works and provides value, revisit the question again in a few months and see if your answers could be different.<p>Whichever you use, if you're identifying identities in your event tracking, then it's best to use UUIDs to make it easier for you to do joins across disparate data stores in future.<p>Personally I'd lean towards Option 2, which is "buy"(but not Day 1, given that there are free plans), which leaves open a pathway to Build if needed.