Website Selector

Today I'll work on the website selector.

Thinking about how to query for this. While I'm avoiding optimization at this stage, I still feel icky putting this in the query:

AND split_part(split_part(url, '://', 2), '/', 1) = $3

Think I'll go with it though. It will be interesting to see the effect it has once I start optimizing the data.

I'll add a simple endpoint that returns all available domains.

This will have to be a more elaborate setup in the future. I'll have to enable users to add authorized domains to figure out what to allow in the selector. For now though, I'll just return everything available.

I'm avoiding huge unnecessary features and scheme changes at this stage, because the requirements are not entirely clear to me.

Here's what that looks like in prod

Next up I want to think about some other stats that I want to show in addition to the 5 main metrics:

• Page Views: sources, paths

  • already done

• Visitors: countries, device (browser, os, pc/mobile)

  • most interested in these

• Sessions:

  • not really necessary for now, but some ideas
  • leaky entries: entry pages with highest bounce rate
  • dead ends: last pages with highest bounce rate

• Session Duration:

  • also not necessary, but an idea:
  • durations breakdown per bucket (30s, 1m, 3m, 5m, 5m+)

• Bounce Rate: most popular, least popular

I'll tackle the countries and devices next time.