Elasticiti helps broadcasters manage direct sold campaigns by accelerating data analytics during live TV events


 

Challenge

Live TV typically features short, but high-intensity events, and require accelerated data analysis to ensure direct-sold campaign commitments are met.

Solution

  • Establish processes to extract data from source providers every 15 minutes during live events

  • Data validation, in order to ensure data is accurate and analysts can trust it

 

Background

According to FreeWheel, in 2016, live streaming surged 36%, driven largely by marquee sports and political events.

For example, the volume of live sports viewed via OTT grew 161% in Q4 2016, and Super Bowl 50 set a new live stream record with 1.4 million people tuning in. The Summer Olympics and Game 7 of the World Series contributed to the momentum.

Understandably, networks are eager to monetize these events by selling premium ad campaigns and sponsorships to top brands. As is the case with all digital campaigns, sellers are only paid once they prove that ads were delivered as promised. Often, if the entire campaign or sponsorship isn’t delivered as stated in the contract, the network can’t charge for any ads delivered, which is why, for live events, networks prefer to err on the side of over delivery.

Some Live Events happen over a period of days, such as the Olympic, while others, such as the Super Bowl, last for just a few hours. Both types are very high stakes, and the time variances require time-appropriate solutions that are drastically different from traditional TV campaigns. 

Challenges

Live events, such as The Oscars or a college basketball playoff are high-intensity and short in duration (two to four hours). Traffic will ramp up quickly, and may spike as key moments are shared on social media, driving more people to tune in. The dynamics of tune-in change radically over a very short timeframe.  

“When I think of appointment-TV campaigns horizontally, campaigns that execute over the course of several months. When I think of Live Events, I think vertically, where traffic represents this huge spike at some point in the day,” explained Ben Reid, CEO of Elasticiti. “Every campaign has some risks that cause concern, but with Live Events, those feelings are escalated to fevered-pitch anxiety. Analysts need every tool at their disposal to get their campaigns delivered, because it’s not like they can make a change in a campaign, and check back the next day to see if that change was effective.”  

This poses numerous challenges to the network’s AdOps and analytics teams charged with managing direct-sold campaigns and sponsorships, and ensuring that all commitments are met, such as:

  • Typical TV campaigns occur over the course of several months to a year, and the procedures, reporting and infrastructure are designed to suit that timeframe. Live Events have much more accelerated reporting requirements.

  • Due to the short period of time to execute their campaigns, networks need to manage them aggressively, tightly and actively. This requires a lot of analysts on hand to manually crunch numbers to track campaigns. It also means those analysts must negotiate fulfillment and share impressions with their peers far more than they do for campaigns for appointment TV. That level of collaboration requires access to fast data.

  • Data in multiple systems, such as the ad server and order management, is siloed, making it very difficult for analysts getting a holistic view of campaign fulfillment.

  • Extracting data from source systems and merging it into an analysis typically takes hours, which is fine for appointment TV, but not for live events. For proper analysis in an accelerated timeframe, data should refreshed every 15 minutes. APIs make this process easier, but if the source system doesn’t support them, the network must wait for the vendor to deliver log files.

  • The campaign-tracking reports broadcasters rely on are built for the day or week level, and Live Events require reporting in hours or minutes. The reporting infrastructure isn’t set up for Intra-day and intra-hour reporting.

  • The network’s existing infrastructure isn’t designed to handle big data processes at the speed required for live events.

  • With appointment-TV campaigns, analysts tend to optimize for several KPIs that are relevant to the advertiser, with Live Events, the focus tends to be on a single KPI: delivery. There’s no time to consider the palette attributes analysts typically do with non-Live Event campaigns. This means analysts need to make a different set of choices, including which data to collect, and how much to consider.

Solution

  • Data Refresh. Elasticiti will extract source data every 15 minutes to ensure analysts are working with fresh data. Elasticiti will set up APIs, or if necessary, make arrangements with source data vendors to send log files at regularly scheduled intervals during the event. We ensure that your analysts have the fast data they need to carry on the fulfillment negotiations with their peers to ensure all campaigns are fulfilled.

  • Development of a Time-Appropriate Approach to Specific Live Events. Elasticiti provides advice and guidance that’s specific to the event type and length  (e.g. two-week vs. 4 hours). We will help you assess which KPIs are realistic for optimization, and design reports around those reports upfront.

  • Infrastructure Optimization. If required, Elasticiti will set up a temporary Massively Parallel Processing (MPP) architecture to support the accelerated analysis during a Live Event.

  • Intra-Day Data Visualization. Elasticiti can design data visualizations to help analysts quickly determine how each campaign and sponsorship is tracking.

 

Results

  • Accurately track how campaigns perform in real time, enabling onsite AdOps personnel to prioritize campaigns as required.

  • Reduce 50% of the analysts time typically consumed in gathering data, enabling them to spend more time optimizing campaigns.

 

Questions about how we can help you?