42% of media CIOs struggle to manage the immense volume [1] of content being produced. On the other hand, the audience is ever-hungry for more and more content and stories that educate, inform, and engage them. To compound the issue further, the pandemic has led to pausing or slowing down the creation of new content, causing a content gap.
This is the paradox of media content today. Production is not able to match the demand, and thus the production that is happening is leading to overburdened content management teams.
IDC and Seagate published a white-paper, that claimed the following — in five years, entertainment content alone will stand for 25% of all the world’s data. Assuming the pandemic will not have any impact on this number — even though content is being consumed faster than ever before due to the pandemic — that is a lot of data to deal with.
In this blog, I will present two solutions to the challenge of managing your content effectively and making more sense of and from it, which will eventually address the content gap. For the second challenge of impaired production teams and mitigated creation cycles, refer to this free eBook where there are some modern work recommendations for your Media and Entertainment company.
You already have the solution to the paradox
It resides in your archives and repositories of pre-existing content! Furthermore, it is highly likely your existing content is under-utilized and its maximum ROI not achieved yet. It doesn’t make much business sense to abandon them to continue creating fresh and new content. A middle route is needed here — some of your teams produce new content, while you delegate a few of your resources and time to efficiently manage your libraries and back catalogs to discover and repurpose relevant, timely content for your audience.
Media and Entertainment companies have traditionally adopted manual, time-consuming, and painstaking methods to tag, discover, and reuse content. At the pace and scale at which content is needed now, there has to be a shift from manual to digital. That is — a review and revival of your technology, a digital transformation.
Hence, I spoke with the digital experts at MediaAgility now part of Persistent Systems to understand how you can improve your content management and workflow, and unearth deeper consumer insights with technology. Here are the two digital recommendations that stood out:
1.Generating intelligent metadata
Metadata, simply put, is data describing your audio, video, and text — what it is about, when was it created, what does it hold, and more. Tags can be leveraged to identify specific data such as actors, genres, plots, locations, languages, themes, actions and so on.
Metadata opens up a huge opportunity to discover and reuse content, optimize workflows around program scheduling, content moderation, and localization, and recommend contextual media content feeds to your audience.
For example, you could search for all “Christmas movies” from your media library to play during the holiday season or search for all “Dwayne Johnson movies” to recommend to the wrestling fans.
You can also leverage Natural Language Processing (NLP), computer vision, speech transcription and other deep learning techniques to generate metadata.
The company receives over 600 audition videos every month. The whole process of reviewing the videos, classifying them into relevant categories, and rating them was carried out with a time- and effort-intensive manual exercise.
Furthermore, the client evaluated how ML could help them understand their media content better. Together with our team, they developed a solution to ingest batches of videos, run them through multiple sophisticated ML models, and extract granular impressions of the context-aware data.
There were three phases or pipelines:
- In the first pipeline, the model was trained to identify generic-labeled real-world objects in the videos, like — drums, trampoline, chair — and also people by their professions, like — singer, acrobat, and trumpeter. The occurrence times were projected on a horizontal, interactive timeline with a corresponding confidence score.
- The second pipeline was to analyze the audio content. The audio in the videos was transcribed by passing them through a speech to text system. The transcript was later fed into a Natural Language Processing(NLP) system to identify important entities like the performer’s name, the school they attended, their experience, and more. Here is a brief demo depicting how to pull insights from audio.
- The third pipeline helped to understand the demographic distribution.
As a result, the client could find the right mix per their requirements with “Smart Search” — a search for specific artists based on talent, gender, age or ethnicity. For example, “search for mid-aged male trumpet artists who can also speak french”. This has led to increased productivity and less time to production.
2. Understanding your consumer engagement data
The second recommendation also revolves around a pre-existing content — content on social media and your other tools and platforms. Analyzing this data will help you better understand your audience preferences and adjust your content strategy for increased engagement.
You can start by (a) analyzing the first-party audience data with the latent features about users and content and (b) tap into social media signals for insights. This helps in creating more accurate user personas that can be targeted with personalized content from your archives to optimize engagement.
But viewing all the siloed data within the same context and at the same time is a challenge that impedes delivery of relevant content to the audience. Here is where you need to delve into more organized content operations on a data platform that integrates, analyzes, and reports insights. AI/ML capabilities further accentuate the usefulness, speed, and reliability of such a data platform.
One of our clients adopted this method. A video intelligence provider modernized their data ingestion, processing, and classification process to set up a standard data warehouse. Integrating all the data not just gave them a finer view of their audience, but also provided their ML teams a structured dataset to train and deploy ML models, that further enhanced the time to achieve results.
Stories are at the heart of any media company. And to continue to tell stories to your audience in a scalable way, you need to alter the story of your content creation and production. In this new content story of your media company, produce new content and also look back at what you already possess with data analytics, cloud, AI, and ML.