How much is possible with email data analytics?


#1

Hi team,

I have few queries regarding email data analytics and what kind of data snowplow captures:

a. Cross browser rendering of the emails?
b. How do we report what devices were used for emails?
c. Heat maps of the email – or any other way for us to see the engagement of the end user in the window
* 1. how much did the user scroll down

  •       2.	what parts of the creative were viewed more than the others 
    

d. Geo Tagging


#2

Hi @dmah

Be aware, that most of email clients (including Gmail and other web based clients) block JavaScript in email code. Therefore it is impossible to track emails scrolls, time of being displayed and other interactive events. Usually kind of non-js tracker is used (pixel based approach), but you need to remember, that most of clients block external images until user allows to display them).


#3

Thanks @grzegorzewald - that’s really important context.

Just to follow-up on two of @dmah’s requests:

Even with our pixel tracker we can capture the user-agent, and then you can use one or both of our two user-agent parsing enrichments.

With our pixel tracker we still capture the user’s IP address, and then you can use our MaxMind-based IP lookups enrichment to identify the user’s geographic location.

Hope that helps,

Alex


#4

E-mail tracking is a bit limited by default (because no javascript). I would also recommend to use individual parameter values on each clickable element in the e-mail body. An utm_ parameter set might be usefull. With the tracked openings and the data of the clicked elements (tracked on the target domain, by analysing the utm_ parameters in the target urls) you can get some more insights about the performance of your mailing.


#5

To add to that: if the campaign attribution enrichment is enabled, the utm_ parameters will be used to populate the mkt_ fields in atomic.events (removing the need to analyse the URL).


#6

Hey dmah,
To build on what have been said and a summary would be
A) If possible use web hooks or similar so you get the full send picture (there is a guide for Mailchimp already)
B) email clients by default will block js
C) Use a pixel tracker and for users who “open” (download images) you can get location (via IP/Maxmind) and user agent.
D) For all links make sure you use campaign variables so you capture link intent

Bonus answer, although haven’t tried this yet myself but theoretically if you use the Clojure or Scala collector who that sets 3rd party cookie at server level you should be able to tie the cookie to a unique user if he/she uses a web browser for email

PS most ESP’s heat-maps are actually click heat maps of user clicks and not views only


#7

Just looking into implementing the pixel tracker to capture users’ IP addresses when they open an email. However, have been reading that Google’s relatively new image proxy/caching system would prevent the passing through of IP.

"However, Google added a twist to loading images by default; they’ll be loading them via a proxy and cache the images. This means that using images to identify the recipient’s location and browser details won’t work." from: https://customer.io/blog/gmail-loading-images.html

Shall I assume that everything above in this thread was written before Google turned on this proxy system? And that the quote from Customerio is, unfortunately, now the ground truth?