Video Essentials

Using First-Party Data to Build a Personalized Media Company


If being online is about personalized experiences, delivering the most appropriate content requires deep insight to win audiences. First-party data is information companies have captured about their audience—who is viewing content, what are they viewing, for how long, and from what location. Turning this data into valuable customer intelligence can move the needle in customer engagement.

Collect Real-Time Metrics

Ad-supported over-the top (OTT) video provider Xumo uses real-time metrics to monitor user engagement times per session, understand what programs are trending, and determine which promotions are working. “Engagement, session time, and time watched on the platform is my north star when it comes to measuring the viewer experience,” says Anthony Layser, vice president of partnerships and programming for Xumo. “[Monitoring] has allowed our partners and programming team to make effective decisions that have significantly boosted engagement times and overall users.” Data capture is the company’s number one focus. “We don’t launch anything without having the right beacons, pixels, or codes in place. If we can’t measure and react, then we’re not improving.”

Beware of Siloed Data

More than half (57.1 percent) of respondents in a recent IAB study, The Outlook for Data 2017, said collecting data about cross-channel measurement and attribution is their most pressing need. Knowing what data to collect and identifying where to store it is a big task.

First-party data sits in silos throughout every company and connecting those silos is often a large custom integration. Each customer touch point generates valuable information about purchase history, viewing preferences, demographic details, and engagement levels. The problem is customer data is stored in various formats in a number of different systems within every organization.

There are two varieties of solutions for consolidating this cross-channel data. Data management platforms (DMPs) are sophisticated systems that coordinate first-party data with data from other sources. Adobe Audience Manager is a DMP used by many leading media companies. In the other option, data lakes, raw data is stored in a secure environment and is managed as needed by the media owner. More on both of these options later on. First, identify what data to capture.

Starting Out With First-Party Data

For each organization, some data is more valuable than others. Is finding out who the user is the most important thing? Or is it discovering how many seconds of advertising viewers will tolerate? Belsasar Lepe, co-founder of video platform and analytics company Ooyala, recommends asking:

  • What audience demographic information is available?
  • What does engagement time look like?
  • How many seconds of ads are acceptable?
  • What is the right price for subscription or transactional access?

“Start collecting information and collect more than you think you’re going to need,” Lepe says. He recommends collecting several months’ worth of data to validate a hypothesis. The hypothesis might be that there are 50,000 unique users viewing content and 25 percent viewers leave before seeing the second pre-roll. If shorter pre-roll ads are introduced, this should lower the viewer bounce rate.

Validating assumptions can either show things are tracking as they should, or that the initial assumption was not accurate. “A lot of those questions are very A/B testing in nature. You want to be able to run several scenarios,” Lepe says. One scenario might be shorter ads, while another scenario might be to target an additional audience segment that may be more tolerant of the ad load.

Get Organized

Nina Caruso, product marketing manager for Adobe Audience Manager, offers the following advice for developing a data strategy.

  • Get organized: Identify and segment audiences. For example, viewer group one is millennial women who play sports, and viewer group two is parents of millennial women.
  • Analyze and build:  Look for additional insights by combining first-party data with other data to build meaningful audience segments.
  • Activate: This could mean packaging the segment for audience-targeted media buys or testing what content is resonating.

Pull in Outside Vendors

When talking to vendors, ask how data is imported, stored, and manipulated, suggests Ashley J. Swartz, CEO and founder of Furious Corp., which provides an AI software platform for bespoke inventory and revenue data management. “It’s critical for brands to ensure they build and own first-party customer data,” Swartz says. “Otherwise, when you want to change agencies or vendors, you’re held hostage.” Read end-user license agreements closely for any technology partner, she adds, and be sure ownership is spelled out.

Working With DMP

DMP’s were developed to help companies build better audience groups so they can target based on information they have about their customers. DMPs consolidate disparate data from customer relationship management, point of sale, web analytics and ecommerce applications. The benefit of a DMP is the ability to store all sorts of first-party data from across an organization. This first-party data can then be combined with other data sets from partners. DMPs can identify traits for an existing customer group and define look-alike audiences that may also want a product or service because of similar interests.

Working With Data Lakes

Another approach for media company is developing their own data lake, where data is stored in a secure environment and can be managed as needed. Ooyala offers an analytics tool, Ooyala IQ, that can be coupled with professional services to develop and manage data lakes. Transform has a platform that provides storage, analytics, and social intelligence capabilities using machine learning.

Figuring out where data is—whether it’s in Excel, PDFs, or raw log files—and how to securely access it is one of Transform’s biggest challenges, says Randa Minkarah, the company’s co-founder and COO. “It’s not as simple as everybody has the same stuff. They may or may not.” Since first-party data collection can vary wildly by systems, formats, and individual company needs, this will become an even bigger area for media companies.

The ability to customize queries and use machine learning should bring great insight into what works for both content creation and marketing. When customers can specify what they choose to engage with, gaining deep insights from first-party data can mean the difference between viewers tuning-in or tuning-out.




Discussion

Comments for “Using First-Party Data to Build a Personalized Media Company”

Post a comment