Paul Boocock of Snowplow Analytics joins me for this episode.
Paul is a brilliant engineer, helping build Snowplow into the analytics pipeline powerhouse that it is. For the uninitiated, Snowplow is a modular, open-sourced, end-to-end analytics pipeline that anyone (with the skills and resources) can setup to run their own analytics collection, processing, and reporting system.
This episode isn’t (just) about Snowplow, however. Instead, we discuss the world of online analytics with Paul.
As seems to be the trend in this podcast, our talk touches upon the difficult compromise between privacy and feature engineering.
How do you build an analytics platform so that it is as flexible and multi-purpose as possible without introducing features that make it easier to mess with user privacy?
How do you build features while observing the ever-changing world of browser standards, the increasingly more intimidating spread of privacy regulations, and the entire complexity of user intent, especially when set against the backdrop of the organization’s data collection needs.
Listen to the episode to hear Paul’s (and mine) thoughts on these very difficult topics!
Listen to the episode using the player or find it in your favorite podcast service.
Topics
- 00:00:00 – Introduction
- 00:06:31 – Paul explains how tracking beacons work and how an analytics server could make use of them.
- 00:09:31 – Why are image elements specifically being used as the predominant tracking beacon technology?
- 00:12:18 – How much do analytics engineers need to be aware of the changing browser landscape?
- 00:14:28 – How POST requests can be utilized instead of image beacons, and what other collection methods are there?
- 00:17:55 – In 2021, why would (or should) ANY user ever consent to their data being collected by the site?
- 00:21:46 – How do you encode different levels of consent and/or anonymity into an analytics platform?
- 00:25:52 – How much do privacy regulations and ethics play a role in the mindset when engineering features into Snowplow?
- 00:31:08 – How do you reconcile with the pressure from so many different directions – your customers, browser changes, laws and regulations?
- 00:36:29 – How important is it to keep Snowplow as flexible as possible, especially for the larger enterprise customers who expect the best quality data?
- 00:41:43 – Is the increasing use of ad blockers a threat to online analytics?
- 00:43:56 – How much crossover is there between online trackers for web browser vs. trackers for mobile apps?
- 00:48:02 – Does the collector have to be reconfigured to accommodate hits from mobile apps, or can it process both browser and app requests without any changes required?
- 00:50:47 – If you could change any aspect of the web with regard to your work in online analytics, what would you do?
- 00:52:40 – Granularity of privacy controls in the browser.
- 00:42:54 – Outro
Notes and references
- Browser tracking protections (cookiestatus.com)
- Snowplow Analytics
- doNotTrack
- Global Privacy Control
- How To Leverage Cookieless And Anonymous Tracking With Snowplow
- Snowplow JavaScript trackers
- Client hints
- Snowplow release notes: Removing user fingerprinting
- Apple’s ITP CNAME restrictions
- Apple’s AppTrackingTransparency framework
- Snowplow for mobile apps
- I Don’t Care About Cookies (browser extension)
- Paul Boocock on Twitter (@Paul_Boocock)