Legacy data warehouse projects historically took up to two years
Satyen Sangani · Fake News, Alternative Facts and the Enterprise // Satyen Sangani, Alation (FirstMark's Data Driven) · May 24, 2017 · at 6:17
Satyen Sangani, Co-Founder and CEO of Alation, speaks at Data Driven NYC about the evolution of enterprise BI stacks from centralized legacy architectures to modern self-service tools.
0:00 / 0:09exact quote · 9.9s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“To stand up a business object's universe and cube in a Teradata warehouse, or a Netezza warehouse, or whatever warehouse could take anything on the order of two years, right?”
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 Satyen Sangani
Enterprises suffer from internal fake news and alternative facts
“You have an alternative facts problem that exists inside of the enterprise with as much fake news, with as much lack of understanding, as you do in the public domain.”
Enterprise data analytics investments create as much noise as signal
“My contention to you in this audience is that the investment in data analytics that every single person in this room is making for the companies that you are making is building as much noise as signal.”
Measuring churn against contracts up for renewal is most accurate
“Churned revenues are the revenues over the things that were up for renewal. Now, in theory, that's the most accurate claim. That's the most accurate metric, because if something wasn't for renewal, then you can't really say that it churned, right?”
Controlling data allows people to construct whatever narrative they want
“You control the data, and you build a narrative in a way that supports the claim that you want to make.”
Information overload drives confirmation bias because parsing data is too painful
“You look for the information that supports the claim that you want to make, because parsing through all the information is just too much of a pain.”
Organizations must require analytical labeling to clarify data context
“We've got to kind of figure out analytical labeling, right? That it's, that's not accept enough, good enough for us to kind of accept information that's proffered to us without understanding how that information was produced, where that information came from, …”