How Complex Streaming Video Ecosystem Data Impacts Business Goals
Democratizing streaming video ecosystem data, if done effectively, can both improve viewing experiences and reach business goals. The time to shift towards ROI and profitability is now. The first two months of quarantine in the U.S. saw an additional four percent of households sign up for their first streaming service, which quadrupled the previous growth rate. Viewing time for U.S. streaming services are 50 percent above 2019 levels in June. This is likely the result of streaming services launched in the last 12 months, including Apple TV+, Disney+, HBO Max (AT&T), and Peacock (Comcast).
Orchestrate data collection to power a data-driven strategy
The first goal in streaming video ecosystem data collection should be to build a high-quality and engaging user experience. To do this, it is critical to have detailed monitoring in place, have the flexibility to run queries needed in real-time, and then act on it.
This starts with understanding the current state of the video system. To start, there must be a comprehensive understanding of data access and availability. Today, companies should be capturing data and forming video metrics throughout the video preparation and delivery pipelines.
This real-time data across various workflows should be integrated to add context to insights, identify the downstream impact of failures, and pinpoint the root cause of problems. This type of data integration across systems within the same workflow helps expose underlying issues and inefficiencies at every step.
Implementing data-driven actionability
Once the necessary streaming video ecosystem data has been collected, and cross-system monitoring is in place, content providers can begin to build their strategy to turn insights into actions. Acting on data can be either “online” or “offline.”
“Online” actions can be making automated changes to encoding ladders, updating recommended content options, or changing a delivery pathway or provider given a set of conditions or thresholds that are reached. These actions may be powered by leveraging machine learning (ML) models that are refined over time and using decision trees to automate a process that may have previously been manual. “Offline” actions can be the preparation and execution of run-books in a triage scenario, or simply knowing when and where to involve experts.
Calculating ROI
Measuring outcomes against changes in business goals and KPIs will enable businesses to measure the impact that decisions have made to their top and bottom line. Ultimately, this means measuring lift and incorporating ROI calculations into technology investments. This could be looking at changes and improvements to QoE-related churns, changes to technology costs, or trends time spent to close customer support tickets or customer satisfaction surveys aligned with average customer lifetime or overall churn percentage.
State of streaming video ecosystem data
For most video publishers, the economies of scale are so great that it doesn’t make sense to build out most core services (such as players, CDNs, origin servers, ad server, and backend services) in-house, and therefore the industry has been compiling their own technology stacks using best-of-breed solutions provided by video data platforms. With this strategy, the burden lies with the video publisher to ensure the system works and is efficient end-to-end. However, the trade-off comes with third-party services building their data and analytics capabilities to measure their individual functionality, not exposing their performance within the context of a workflow.
Until recently, the strategy of third-party services was to provide a dashboard or report that exposed the usage and performance of that service. Furthermore, each provider built their data and analytics systems independently before standards were created, or simply not adhered to. Therefore there are ways to understand how a single service perform, but it’s hard to compare the performance between competitive services, and determine the impact on efficiency and performance end-to-end
As the demand from publishers to provide capabilities for end-to-end observability and optimization has grown, so has the demand for solutions to provide and expose more real-time data, so that it can be aligned with data from upstream and downstream systems. The good news is that today there is greater access to more data at faster speeds than ever before. But the challenge of cleaning, standardizing, enriching, and aligning data still remains.
The Challenge
With streaming video ecosystem data stuck in too many silos, there’s a struggle to understand the underlying factors that shape results. There is too little data standardization, which doesn’t support actionable analysis.
As an example, when ad revenue has dropped, we need to look at different variables that impact revenue, and look at various data sets to uncover the answer:
Have CPMs dropped? Why?
Data: Look for insights and correlations between player and ad server data
Have audience levels dropped, or changed? Why?
Data: Look for insights and correlations between audience and demographic data
Is ad delivery facing technical challenges? Why?
Data: Look for insights and correlations between player, ad server, and ad stitcher data (depending)
Does the content face technical challenges that cause users to drop-off before the ads? Why?
Data: Look for insights and correlations between player, encoding, CDN, and origin data
When it comes to marketing dollars maximizing ROI, the cost of acquisition compared to revenue generation matters immensely.
Are users converting from our marketing campaigns to engage with the platform?
Data: Look for insights and correlations between player and attribution data
Are those users delivering revenue?
Data: Look for insights and correlations between player and subscription (SVOD) or ad server (AVOD) data
Which campaigns deliver users with the highest MRR or LTV?
Data: Look for insights and correlations between player, 1st party data, and financial data or ad server data
Gaining actionable insights
When we’re looking to make changes, we need to have already identified the priorities and the goals. First, what does the business want to achieve: Are we prioritizing growth? Profitability? The user experience? Secondly, what does success look like: 10 percent revenue growth? 10 percent subscriber growth? 10 percent cost reduction? For each business goal, metrics need to be agreed upon that are used to measure success.
What’s next is to ensure the underlying data is accurate behind the calculations/goals and can be measured to show progress. Establishing a single methodology for what and how things are monitored and measured will create a common understanding of system performance. Therefore, when insights are tied together, it can easily deduce what variables impact end-to-end workflows and thus control the outcomes and results, matching the business metrics.
Democratizing the complex streaming video ecosystem data: Why it matters and what the future could hold
Without the ability to understand what’s possible and to make improvements, so much of the efficiency and value of using best-of-breed solutions will be lost. Data is the only way to uncover and measure the impact of actions taken, and to track progress against business goals. Looking into the future, service providers should be evaluated not only for their capabilities but also for their ability to be a strong provider of data.
The result will be greater efficiency for the video streaming industry. Traditional “video analytics” tools or other single-system analytics diminish, and end-to-end analytics and workflow analytics rise in their place. With data as “building block” solutions that support data engineers in reducing the time instrumenting, piping, standardizing, and cleaning data will grow and power new innovations for real-time adaptability.
As our technology improves, how we approach measurement should evolve as well. Analytics only matter if customers know how to leverage these insights and transform them into impactful changes for their business. It’s critical that everyone in the organization understands the same goals, uses the same metrics, and speaks the same “data language.” In 2021, the industry will focus on making this data align with business goals for the organization – ensuring it is made accessible and contextualized for all relevant stakeholders.
Creating sustainable growth in the video business is about setting goals and priorities, defining metrics, and measuring progress in the context of revenue, lifetime customer value, and profitability. To get there, the industry needs investments in data, collection, standardization, and enrichment and measurement at every stage of the business.
Have questions? Get in touch, and we’ll get you the information you need — it’s what we do.
How AVOD Services will Stabilize Heading in 2021
The state of today’s AVOD services has intensified dramatically. Users are streaming more than ever before, dropping out of cable subscriptions and adopting a mix of SVOD and AVOD services to replace their needs. This is particularly the case with families feeling the long-lasting economic stress that 2020 has inevitably left in its wake.
With the stress comes the opportunity to streamline. Some users are already replacing SVOD with AVOD services. Since the pandemic, 35 percent of users now watch AVOD content, while Nielsen reports that overall, AVOD has grown 3x since the beginning of the pandemic. With limits to the number of SVOD services someone is willing to pay for longer-term, AVOD services are here to stay. They fill gaps for fractured audiences looking for niche content, where SVOD is without.
However, it’s clear market turbulence has been a driving force to make AVOD companies find their profitability. When the economy took hits in April, advertising revenue was down 35 percent. The profitability mark for AVOD services is coming sooner than we think. We first saw AVOD services primarily offered by larger organizations who could use this as a loss leader to attract new audiences or mine data before the service itself hit profitability. More recently, the independent AVOD services which have come to market and have been swallowed up by larger enterprises and conglomerates. But now that we’ve seen how quickly the economy can impact advertising budgets, fewer newcomers will likely enter the market without the backing of deep pockets and investment.
Challenge
There’s an opportunity for AVOD services to be a key part of OTT and CTV streaming services, but first, we must address the value-recognition problem in ad-supported streaming.
The previous goals for launching an AVOD service was:
Get to market
Grow the audience quickly
Push for profitability
The problem there is when we go into an economic freeze, advertising budgets tend to be the first thing to get cut. Today, there’s more pressure than ever to reach profitability; do more with less.
Solution
Since AVOD revenue is subject to the swings of marketing and advertising budgets, we need to do a better job of matching supply and demand. Real-time bidding (RTB), which is a subset within programmatic advertising, enables advertisers to bid for an ad spot in real-time. Ultimately, the more bidders, the higher the price.
There’s an opportunity for “header bidding” within RTB to grow, but this will require more data collection, more data standardization, and more data sharing. The trade-off we make with moving to header bidding is reduced targeting capabilities and reduced control over the user experience than what you’d find using a primary ad server.
Focus on ad viewability
In comparison to an ad seen on linear television, a study by Unruly TV found that a user watching ad-supported Connected TV was 71 percent more likely to tell a friend about a brand, 53 percent more likely to search for a brand, and 48 percent more likely to have an improved opinion of the brand. Moreover, 52 percent were more likely to buy a product, and 45 percent more likely to visit a store or website.
Despite greater effectiveness, advertisers still prefer to advertise on broadcast television than digital advertising. Part of this could be due to the claims of CTV fraud and the inability to measure things the same way as linear television
Again, data is the only thing that can come to AVOD’s rescue. The ad viewability and verification are equally as important as understanding the audience. As a publisher, if you can differentiate your inventory by offering new metrics or providing direct access to real-time data, you’ll help your advertisers better measure their ad dollars’ reach. Placing a focus on providing more data for advertisers, which can, in turn, help them optimize their campaigns, can create a multiplying effect on CPM’s.
Using data to understand cost
The other half of the profitability is cost. Having the right data and using the right metrics is key to knowing why users stay and why they churn. It’s important to understand that you can add efficiency to costs. Asking questions like, “What percentage of leads from marketing watched the content?” Or, “What’s the cost of end-to-end delivery vs. CPM?” The need to connect attribution data to engagement data or content preparation and delivery data to users and sessions should take significant consideration.
In fact, granularly understanding of ad revenue needs to improve dramatically overall. You’ll want to understand how much money users earned you on X devices, watching Y content, in Z geography. What cohort of users earns you the most money and what about them is unique? If you can answer those questions, then you’ll want to learn how to optimize marketing, support, and operational budgets to protect those viewers.
If you’re detecting certain users, campaigns, or pieces of content are not delivering on the profitability targets, you could decide you may want to:
Prioritize bug fixes or improvements on those platforms
Make changes to your content library or content recommendations
Change the available bit rates
Optimize ad pod count, structure, and positioning
Shift marketing spends towards those places with the most ROI and decrease the spend on other platforms.
Glimpse in the future: Data will provide AVOD’s Moneyball Moment
Unless you can take effective actions to improve profitability with the metrics, then they’ll have to change, and you’ll need data to do it. It’s the key to targeting and is a rising tide that floats all boats.
Moneyball was a movie about how the same data, analyzed differently, led the Oakland A’s to a historic performance and changed the game of baseball forever. Data isn’t just for reporting anymore. More companies are defaulting to wanting access to data versus metrics because granular insights are what will enable greater optimization for the user experience and create more efficiencies across this end to end network.
Lastly, advertising will always be an ecosystem endeavor. Therefore, optimizations likely need to happen within and between multiple parties. This means that the future of advertising will include new types of data and data exchanges between different systems. Investing in systems today that enable the capture, standardization, enrichment and delivery of data will help lead to the more efficient and profitable advertising landscape of tomorrow.
Have questions? Get in touch, and we’ll get you the information you need — it’s what we do.
What Segment’s Acquisition Means and the Opportunity for Data Standardization
It’s likely you saw that Segment, the leader of the Customer Data Platform (CDP) category, is being acquired by Twilio in a 3.2B all-stock deal. This effectively doubles their valuation from their last funding round from a little over a year ago in April 2019. The thesis behind this acquisition signals the value behind companies focused on data standardization and reduction of data silos.
As someone immersed in this space, I can’t help but see the similarities between Segment and Datazoom: where Segment captures, standardizes, and routes customer data from end-users’ digital touchpoints, Datazoom does the same across video’s touchpoints, which includes the end-user’s application. While Segment’s data is primarily used by marketing teams to improve customer identification and segmentation, Datazoom is used by product, engineering, and infrastructure operations teams to improve observability and optimization efforts. Both are focused on leveraging data to make it actionable for their audience.
I’m so glad Segment has been recognized for the value they bring to customer data, but the opportunity for data standardization is still much larger.
The vision behind Twilio’s acquisition of Segment is to be able to gain a more complete understanding of the customer (Segment), and leverage that data to act intelligently based on those unique insights (Twilio). The challenge to apply what these two companies have built so far to the world of streaming video is that the relevant data needed to take action comes from an ecosystem that hasn’t been explored by Segment — it’s data created by video players, CDNs, ad servers, encoders and other pieces of an end-to-end system. Furthermore the actionability of video data extends far beyond only the content publisher – to really drive value and improvements with data in the video space will require not only data standardization but real-time industry alignment and coordinated actions across an ecosystem of third party vendors, cloud solutions, and internet service providers.
Additionally the volume of data created by these systems, and the speed at which this data needs to be leveraged, is best served using a different infrastructure and business model than the current CDP model. To be generous, a CDP might log 15 events per customer interaction. By comparison, you might track 120 events across the end-to-end video delivery for a 2-minute video. Twilio claims that they will be close to powering 1 Trillion customer engagements this year. By comparison, Youtube alone will stream 1.8 Trillion videos this year. That’s not including other industries incorporating video beyond media & entertainment, such as distance learning, remote work and telehealth.
With the OTT video market projected to reach $169.4 billion by 2023, better harnessing video data is critical to understanding viewers and content engagement, improving and measuring video advertising, and optimizing the end-to-end workflow for improved efficiency and streaming experience. And this is just the Media & Entertainment vertical. Video is becoming a service within other verticals such as Education, Remote Work and Telehealth.
Datazoom is well-positioned to tackle the ecosystem that needs to be created around video data in the same way that Segment has tackled the ecosystem around customer data. Segment’s announcement cements even further the unique market opportunity we have in front of us in the Video Data Platform market.
Have questions? Get in touch, and we’ll get you the information you need — it’s what we do.