Cloudfront Collector to Scala (or Clojure)

[Robin_Darrah]

We have been using the cloudfront collector for over a year now as our source of truth for web analytics. Recently we’ve started to do more advertising, and would like to start understanding post impression conversions as well as post click - cloudfront is a bad fit for this as it can’t assign user ids to new users.

Is there a fairly simple way to run the impression tracking through one of the Scala or Clojure collectors, and then link those network_userids up with our already existing domain_userids (set by our client side js tracker)?

If not are there instructions for migrating all our collecting from Cloudfront to Scala, and preserving as much userids and similar as possible?

Thanks,

Robin Darrah

[ihor]

@Robin_Darrah,

cloudfront is a bad fit for this as it can’t assign user ids to new users

What do you mean here? It’s true that network_userid is out of the picture. However, domain_userid and user_id are all to be considered.

Is there a fairly simple way to run the impression tracking through one of the Scala or Clojure collectors, and then link those network_userids up with our already existing domain_userids (set by our client side js tracker)?

Typically, we recommend the approach described in Identifying users (identity stitching)

are there instructions for migrating all our collecting from Cloudfront to Scala, and preserving as much userids and similar as possible?

You might consider Javascript tracker: How to keep track of users when changing the cookie domain? [tutorial].

[mike]

Are you measuring impressions / clicks on your own first party or owned domains or are you embedding Snowplow within a creative that’s being served through 3rd party networks?

I ask because network_userid might be useful but is less useful at the moment (depending on your use case) given ITP and Chrome changes around the way that first and third party cookies are handled in the browser.