Ribant: Datorama avoids privacy pushback by training on user actions, not data
Katrin Ribant · Marketing With Data // Katrin Ribant, Datorama [FirstMark's Data Driven] · Nov 9, 2016 · at 15:30
At Data Driven NYC, Datorama CSO Katrin Ribant answers Matt Turck (FirstMark) about whether enterprise clients object to machine learning models learning across customer data.
0:00 / 0:28exact quote · 29.0s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“No, never because we don't really, I mean, we don't at such use your data, right? We use your actions and you self benefit from your own actions as well. So you are, you know, integrating a first data source and maybe it's the first time we see it. So you're doing, you know, a whole bunch of operations and you're mapping the data. If you have a similar data source and you're eating rated a second time, probably 80% of the work will be done for you. So I, we've never really had any complaints from that perspective.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Katrin Ribant
Ribant: Harmonizing marketing data is harder than simply integrating sources
“All of those execution platforms need to be connected together, so the data needs to be extracted and integrated. So there is one challenge around connecting to the data source and actually integrating the data source. But the second challenge, which is a litt…”
Ribant: Traditional BI specs become outdated before they are even finished
“In most of the projects that I have been involved in, if you do something like this, by the time that you have finished the specifications, everything or half of what's in the specifications has already changed.”
Ribant: Datorama mostly competes against fragmented, multi-tool traditional BI stacks
“Mostly what we see in terms of competition is, if you remember the slide where I had the whole process with the, all the different pieces, an ETL tool, a database, et cetera. That is mostly the competitive set that we, ah, that we see.”
Ribant: Marketing analysts rarely input the data they are forced to analyze
“The person who analyzes the data is rarely the person who's responsible for putting the data into the platform that will produce it.”
Datorama planned to apply crowdsourced machine learning to marketing insights
“We plan to take this approach further into insights over the course of next year, and kind of bring an ability to crowdsource the same type of machine learning to you know, to help marketers understand across the complexity of KPIs that they have in their plat…”