FEED Issue 10

71 START-UP ALLEY EDO

EDO

COUNTRY: USA STARTED: 2015

Applying financial data science to measure the effectiveness of linear TV advertising is the idea behind EDO, founded by AI entrepreneur Daniel Nadler and Oscar- nominated actor Edward Norton. Nadler is the founder of AI start-up Kensho (of which Norton was an earlier investor), which uses machine learning to make complex financial analysis as easy as a search on Google. The technology was a hit on Wall Street and the duo have now turned their attention to TV advertising measurement, which still tends to rely on old-fashioned panels and survey methodologies. Noting that the film industry was heavily reliant on TV advertising, Norton felt that EDO could bring ‘the rigorous data-driven and actionable analytics – that digital advertising has had for years – to the huge budgets on TV’. Venture capitalist Jim Breyer was among the firm’s early backers while Kevin Krim, former head of digital at CNBC, was hired as its chief exec, along with a small army of Harvard maths graduates. Ed Norton points out that the firm was able to pull together “a level of talent that would not normally be inclined to focus on these specific issues.” But why invest this level of talent (and money – the firm recently raised $12m in Series A funding) in a medium that’s seen viewer numbers dwindle? According to Krim, TV networks and ad execs know that on-air advertising works

on a macro level. In addition, the price of air space has actually risen, with TV sales departments adjusting their ad formats to compensate for viewer decline. Krim adds that in the era of Netflix, Amazon Prime and Hulu’s premium ad-free services, there are fewer places to advertise against premium content, making the rates broadcasters charge hold strong, and any data on ad effectiveness more desirable. EDO’s approach to measuring this effectiveness involves gathering three data sets. The first involved creating a benchmark in the form of a database that contains every ad that has aired on live television since 2014 – which to date totals 47m linear ads. Krim explains: “We run live feeds into DVR Raids on the East and West Coasts and apply automatic content recognition on these feeds to discover the ad breaks. We then record every ad and match it against our existing database of ads.” This data set requires petabytes of storage and it is retained in the cloud. EDO also overlays TV schedule data and audience metadata from sources, such as Nielsen and Kantar Media, while the third data set involves gathering anonymous consumer behaviour data from search engines and content sites where people do research before making a purchase. So far, EDO has measured 2,100 brands and products, with each data set receiving a rigorous validation and normalisation process to weed out anomalies and factors

DO THE MATHS Edward Norton, actor and co-founder, felt EDO could bring the data and analytics of digital advertising to TV

unrelated to the underlying effectiveness of an ad. Once all these data points have been gathered, EDO can give brands a strong indication of future sales, in real time, which Krim likens to an electrocardiogram of consumer intent. “We watch these EKG lines, look at departure from the base lines and align them with the TV ads that are occurring,” Krim says. The three-and-a-half-year-old firm now works with a number of “large paying” clients – including Paramount, ESPN, Warner Bros and Turner. Interestingly, the sellers of advertising are also embracing the quant method – with a sports broadcaster recently using EDO to quantify ROI for advertisers paying a premium for high-profile events. The project worked out that someone who viewed an ad unit during the Championship Game was 2.30 times more likely to engage online with the brand being advertised than someone who viewed a TV ad unit during a regular prime time broadcast break. EDO worked out that the advertiser would require 10.48 TV ad units during a regular primetime broadcast to generate the same amount of impact. “Used in this way it’s as much of a tool for selling advertising as it is for measuring its effectiveness,” adds Krim.

feedzine feed.zine feedmagazine.tv

Powered by