Read on Twitter
To any companies who spent a bunch of $ on Super Bowl ads only to wake up & read (below) that Nielsen couldn’t deliver a low-tech ratings report for over 24 hours: if you’re crazy enough to still be relying on Nielsen for insights, you’re shit out of luck http://tinyurl.com/gsfazumj
Epic fail. But if you're an http://EDO.com client, you don’t really care bc you had a much higher-value ad effectiveness report on this critical ad event by 3am, ready to discuss Monday morning (public version 
, clients get much richer data) http://tinyurl.com/3v5oa4v4
, clients get much richer data) http://tinyurl.com/3v5oa4v4
It took around 36h for Nielsen to process data it collects from only 100K households. A pretty rudimentary system to then make a high-school math level extrapolation against US pop of 320million toward the stunning insight: How many people probably watched (big yawn)
Meanwhile, it took EDO a few hours after the game to complete its full SB report. During it we collected, processed & analyzed anonymized minute-by-minute searches of 100's of millions of people to assess effectiveness of every ad airing in & around the SB, virtually in real time
By 3am, EDO wraps its everyday routine, processing 1+ Petabyte of data covering anonymous, aggregated behavior of over 90% of the US pop. Producing investment grade insight on which ads most effectively drove actions tightly correlated w/ purchase intent & actual economic outcome
EDO achieves this w/ just over 50 brilliant people on our team. We process 1TB of video & 3.6B fingerprint hashes per hour, storing it all in the cloud. We know how to automate. We have some of the legit top ML / AI DS & Eng pros in the world. That drives powerful efficiency
Nielsen has about 46,000 employees but it seems they can’t collect & sort data from 100k set-top boxes w reliable speed. You pay for that headcount, not speed or quality of insight We’re glad someone still does exposure but who on earth would price ad inventory off that in 2021?
Great data engineering & world-class data science allow EDO to handle enormous, unruly data sets with statistical rigor in near real-time & produce finance-grade actionable analytics in cloud-based software, reports, and APIs. That's a modern tech platform.
Convergent TV advertising supports a $100B creative industry & underwrites the work of some of our most brilliant creators & storytellers Counting how many people (probably) sat in front of a screen when an ad ran is entirely insufficient as a pricing metric & has been for years
And we’ll throw down & say that efficacy analysis rooted in real time Search Engagement Rate is the new gold standard metric for cost/performance analysis. Bullshit social sentiment or weak sauce multi-touch attribution signal, sorry…no. Debate us on this, we welcome it.
Data based on mid-funnel consumer behaviors *predictive of eventual sales* brings what digital ad platforms have been able to deliver to the urgently needed evolution of Convergent TV. If your company’s research budget doesn’t reflect this revolution, you’re behind the curve.
If anyone is wondering why I’m writing this: I co-founded EDO with my friend Dr. Daniel Nadler (founder & ex-CEO of Kensho). Biased? Yup. But we assembled this all-star team & we’re incredibly proud of the brilliant work they’re doing. @karaswisher @profgalloway @jonahgoodhart
@pkafka @zachrodgers @bmorrisey @DeadlineDominic @tiffkhsu @edmundlee @kenli729 @sherman4949 @chrispalmeri
