In a previous client letter, I had promised a look at “data federation” as one of the topics that our industry would start hearing more and more about. Like cookieless in 2018, federated data is not getting much hype currently. My prediction is that data federation within analytics teams will be as fluent as the acronyms ETL, SQL, GDPR etc. by EOY 2023.
When we look at the ever changing landscape of ad targeting, we follow a timeline of great advancement in tech > a catalyst event(s) pointing to areas for improvement > innovation and investment resulting in new better technologies to achieve that improvement. For the ad targeting story, we start with a robust tech ecosystem (think the lumascapes of 2013 for those in the industry). Targeting was quick and easy and people were drunk on data. Boom — the 2016 election and subsequent data privacy related conversations (ie: Cambridge Analytica, GDPR), resulted in more commercial investment in privacy tools and applications, and the eventual full release of cookieless solutions from everyone under the sun (although statistically only a few took off, shameless plug).
The same thing is happening in customer data analytics. Right now, I would argue we are in the robust tech ecosystem phase. Headlines like “X company looks to Y cloud to streamline business” are frequent. Clients are investing in clouds, CDPs, ETL, ELT, visualization tools, etc. to connect legacy systems into clouds and vice versa. All that is healthy and good and something LiveRamp spends a good amount of time doing. But what is next? The frustrations in breaking down data silos with cloud migration and data transfer are a. It takes too long, b. It requires multiple public clouds, c. it requires different tools, resulting in MORE data movement, not less, and thus d. It has created more silos which means more of the problems above!
So what is data federation? To do customer analytics today, there is a massive undertaking of “data consolidation”. This involves physically piping data into a centralized area so you can analyze it. This is why a lot of diagrams have arrows, pipes, etc. on them. Data federation leaves an organization’s data where it is, but provides a unified view using virtualization. In other words, data federation offers a means of querying and analyzing information from multiple systems as if it all resides within a single, harmonized environment (I plagiarized this). Imagine, cross channel analytics and MTA done without having to actually touch or move the data out of its current marketing application, or cloud, or even organization — the applications are endless. David Gilmore, LiveRamp head of Privacy Tech, who leads our federated data charge, gives two phenomenal examples in this podcast.
For example, imagine being able to go to your IT team, say to them that you are removing the need to move data from one platform to another, reducing the overall security risk. Now imagine also going to your business team, and saying that not only would there be zero reduction in capability with this change, but latency would actually go down due to the removal of time traditionally used for movement, and the amount of accessible data would actually go up due to more data sets being available now that they no longer have to be moved to be accessed.
LiveRamp’s mission has been and always will be to enable companies to use data effectively without having to sacrifice privacy and security. Data federation is a no brainer investment for us and for our clients. I look forward to watching this space grow together with you!
Have a wonderful week!
Eli