1/ @jonathanmendez& #39;s wonderful post omitted the potential role of @GCPcloud and BigQuery in particular. https://twitter.com/jonathanmendez/status/1195383697781989376">https://twitter.com/jonathanm...
2/ Google Tag Manager works well for owned site-side interactions and site analytics, but doesn& #39;t offer much help for off-site paid media exposure (both targeting and measurement) other than loading retargeting pixels (which will all soon be blocked anyway).
3/ So one of the post& #39;s most ambitious statements is only half-fulfilled by Google Tag Manager. What will be the "lifeblood" on the paid media side? Especially in the era of Privacy?
4/ Started wondering today whether one of Google Marketing Platform& #39;s key approaches to managing #privacy will be to deprecate ALL direct I/O with respect to user-level data within the Google paid media environment (e.g., Input = data onboarding; e.g., Output = ad server logs).
5/ Note "within the Google environment." This doesn& #39;t really apply to Google Tag Manager (or GA360 to some extent), which is usually used in Owned environments (e.g., your own website). Also GTM doesn& #39;t need to pass data to Google in order to be useful.
7/ But...if Google Marketing Platform were indeed to deprecate all direct access to user-level data, what would fill the vacuum? Digital paid media targeting & measurement don& #39;t work very well without some kind of two-way street for data.
10/ OK, so back to that "lifeblood" question. How do you satisfy Google Marketing Platform& #39;s (presumed) distaste for user data I/O while still allowing some kind of "lifeblood" to flow within paid digital media?
11/ BigQuery and @GCPcloud may be the solve. Together, they can be the & #39;ligua franca& #39; on the paid media side where Google Tag Manager is the & #39;lingua franca& #39; on the site/owned side. ADH is, after all, a custom application built on top of BigQuery. https://developers.google.com/ads-data-hub/guides/intro">https://developers.google.com/ads-data-...
12/ This is where it gets interesting (to me). ADH architecture assumes you have your own Private GCP Cloud environment where your user/1P data lives and Google can NEVER SEE IT. This solves so many privacy problems for Google Marketing Platform! (it also builds adoption for GCP)
13/ And the data interchange (targeting, bidding, measurement, personalization) between your Private GCP Cloud and Google Marketing Platform is all driven by Differential Privacy. https://arxiv.org/abs/1909.01917 ">https://arxiv.org/abs/1909....
14/ And of course there are all sorts of fun things you can do with your data in GCP. It& #39;s a cloud platform. The sky& #39;s the limit on I/O and transforms / queries / applications. As long as you can get into or out of BigQuery, it& #39;s the ultimate extensible playground for your data.
15/ So maybe you& #39;ll be able to use GMP tools like DV360 in "dumb" mode with all data cut off (except basic reporting), or you& #39;ll be able to spin up a prepackaged Private GCP cluster and start doing analytics and activation right out of the box.
16/ I& #39;ve no idea whether this is actually Google& #39;s roadmap, but it feels right because it& #39;s kind of a best-of-all-worlds stack. I just have this feeling that Private Clouds are going to figure more and more centrally into #adtech / #martech architecture as things like CCPA land.
You can follow @Myles_Younger.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: