Media Companies are just beginning to scratch the surface of AI/ML

Why ARPU and Customer Loyalty Optimization is not the last step of User Data Monetization

Published in

Provoke Solutions

5 min readAug 3, 2021

In general, a content consumers value to a media company can either be defined by ARPU (Average Revenue Per User) or customer loyalty, depending on the monetization model. Let’s look at Disney’s various services offerings and some key metrics:

Disney Plus Hits 103.6 Million Subscribers

Disney issued a dramatic upward revision on its forecast for Disney Plus, projecting 230 million to 260 million total paid subscribers by September 2024. Two years ago, the company told Wall Street it expected 60 million-90 million Disney Plus customers by that time.
ESPN Plus Customers increased 75% year over year

Disney Subscriber Growth Slowing like Netflix

Disney+’s average revenue per user, excluding India’s Hotstar, was $5.61 per month. Netflix’s ARPU last quarter in the U.S. and Canada was $14.25 per month — up 9% from a year ago.
Disney’s Hulu subscription video-on-demand service has higher ARPU — $12.08 per month — but its growth was negligible, up just 2 cents per month from a year ago.

When looking at AI/ML capabilities in the media streaming space, we either think of increasing ARPU (either through Ads or Service Offerings) or increasing Customer Loyalty through bundling and/or reducing competitors’ space to perform. It all starts with data collection.

When companies go “vertical” by controlling the entire consumption processes (i.e., through set-top boxes, mobile apps, and bundling various streaming services), they can acquire a good deal of information about their users. This information can then be integrated with their content metrics (i.e., advertiser metrics, streaming media consumption, and media purchases), creating a very in-depth picture of their content consumers.

For many media companies, an ARPU is whatever the monthly streaming fee is. However, for ad-supported platforms, an ARPU can be modified by better advertising rates (either by micro-targeting or increasing exposure to ads). This later model requires better and better personalization so that the consumer consumes more content on the platform and/or more information is known about the consumer so that ads can be better targeted. This personalization is an ideal application of a recommendation engine, which provides more and more targeted choices to the consumer to drive the desired behavior.

Customer Loyalty is another excellent example of a need in search of a recommendation engine. Determining a customer’s viewing behaviors and the context around those behaviors provides the training data to better surface content that the consumer would find enjoyable. The more data that you have, the better the suggestion is going to be.

Increasing the information that trains the recommendation engine is also valuable beyond recommending the next piece of content to the consumer.

Instead, that information can be used to determine what kind of content consumers are yearning for and can help dictate content for media companies to acquire or produce.

While the average consumer consumes 6 hours of video a day, we know that serialized media is not the only medium. Short-form content is a significant part of our daily media habits. For example, of the total number of YouTube users

28% say the site is very important to just pass the time
19% say the site is very important to understanding things happening in the world
51% say the site is very important to figure out how to do things they haven’t done before

This study implies that the knowledge that YouTube has gathered is helpful to show what interests their userbase has in non-serialized information, which media companies can use to influence documentary creation, reality shows, etc.

It’s also worth noting that U.S. citizens spend two hours every day reading. This information is also valuable to showcase an entire persona of a consumer. Whether it is information to supplement the YouTube information above, fiction choices by those users, or current news, all of this information can go back into the media company’s recommendation engines and increase micro-targeted advertisements or increasing Consumer Loyalty by meeting needs through new content creation or targeting.

Understanding consumers’ complete media consumption habits can increase ARPU by providing another vertical play, i.e., leveraging data to create sales/affiliate links for non-streaming media, whether you sell books, magazine subscriptions, or products.

As you can see in the chart above, the boxes in green (about eight hours) represent the time that content consumers don’t already have committed during their day ready for content consumption. Audio, Video, or copywriting are all available to a consumer, but with media companies building profiles, they are in a unique position to market and fill those times with their content.

Don’t just use your data to maximize your current profits.

Media companies should use their predictive models to predict what consumers will want and meet that demand. By diversifying your data sources, you can build a more in-depth profile of your users and start a cycle of constant refinement, increasing your ability to personalize your recommendations and find more ways to monetize their consumption.

Reach out to us today to schedule a conversation or visit www.provokesolutions.com for more information.

Andrew Crider is the Head of AI and Machine Learning Applications at Provoke Solutions and has a long and successful track record, often working with Forbes 100 companies integrating all facets of technology to help with their journey to better operations through AI and ML.

Sources:

Media Companies are just beginning to scratch the surface of AI/ML

Why ARPU and Customer Loyalty Optimization is not the last step of User Data Monetization

Written by Andrew Crider