Latest News and Updates

30 June 2019

June Newsletter

Conviva Precision for A/B testing

Why use A/B testing 

Let’s say you are considering introducing a higher bitrate stream for increased picture fidelity. Given the extra costs in Encoding, Origin and CDN, this is a decision that needs a clear and justified business driver. Our customers usually ask:
1. What's the industry trend on average bitrates for similar types of service, content and devices? How does this trend evolve over time and what is my competition doing?
2. How can I calculate the ROI and justify the costs? Will this decision result in higher viewer engagement and consequently audience?

For a long time we have been able to help with #1 using Experience Benchmarks, which enables smarter decision making. It gives publishers unmatched insight and intelligence into streaming experience performance against their peers.
The best way to help with #2 is with AB group testing, which allows you to compare the results in terms of QoE and Engagement on each group.

Why use Precision

For A/B viewer testing, you create two groups:
- group A are viewers for a particular series or live channel, always getting a higher bitrate stream 
- group B are other viewers always getting streams with assigned lower bitrate.




Precision will allow you to allocate the streams to groups A and B consistently. It also automatically ties with Experience Insights so you can confirm the QoE KPIs on each group of viewers and respective Viewer Engagement. Precision can assure that you have a meaningful engagement test, where the groups A and B MUST remain consistent (a viewer cannot get the lower bitrate stream sometimes and the higher another times). Then, Experience Insights will let you evaluate the results in terms of QoE and Engagement, for each group. Advanced analytics reporting capabilities can be leveraged to determine the statistical significance of the results.

How does it work

To achieve this viewer demarcation, Precision uses the same technology that already allows Publishers to Optimize by what we call a Generic Resource.

A Generic Resource is typically a CDN, or a CDN using a given Origin, or a given PoP from a CDN. In an example scenario of AB testing, the Generic Resource would be "Live Channel X with a Maximum Bitrate of 4Mbps on Roku" and "the same Channel but with 6M on Roku". In this case there is no concept of Optimization, but only load balancing two streams consistently and being able to connect this allocation to Experience Insights.

For further discussions on how to explore Precision capabilities, please contact us.


Enhancements to MetricLens Drill-Downs
and TimeLine Analysis

Conviva’s MetricLens dashboards enable you to perform advanced analysis of QoE and audience data through a combination of filters, data ranges, metric values, and dimensions. This combination of metrics and performance data allows you to quickly understand the most critical contributors - ISPs, cities, player versions, assets – of your video streaming QoE and engagement issues.

To transition from summary data to time series analysis, MetricLens users are able to ­­­­­­easily navigate between MetricLens summary and Timeline dashboards to quickly analyze timelines of the top five dimension values, determined by the number of Attempts. With the latest Pulse release, Pulse users can now also quickly analyze a timeline of each individual dimension to discover the detailed timing of anomalies in any dimension value.

Top Five Dimension Values

From a MetricLens Summary page, use the View Timeline button to view the time series of the top dimension values. MetricLens displays the metric-based time series with filters for each of the top five dimension values, in this example the top DMAs show metric spikes on Thursday 04:00.





The View Summary button returns to the MetricLens Summary page.

Single Dimension Value

To drill-down to the time series for a single dimension value for more specific dimension analysis, simply click the linked dimension value on a summary page. MetricLens displays the metric-based time series filtered for that dimension value, in this example focusing the time series analysis on the QoE metrics for the most impacted DMA, New York.





These MetricLens navigation paths are supported in both Conviva Experience Insights and Ad Insights MetricLens pages.

31 May 2019

May Newsletter

Conviva Social Insights - Webinar


Join Conviva and our customers, the Miami Dolphins and Advance Local, for a webinar this Thursday, June 6th, at 10am PST, as we share branded content best practices using real case studies. See how these industry leaders approach selling, executing and reporting on branded content. We’ll also give you the tools you need to start selling branded content on your social channels more effectively.

Topics will include:

  • Pricing content and building a rate sheet for advertisers
  • Branded content reporting made easy
  • How to best partner with advertisers
  • Case studies and experiences from the Miami Dolphins and Advance Local
Sign-up here.

Ad Slates

What are Ad Slates?

Ad Slates are short media files that play in place of an ad during a video stream ad break. Ad Slates can negatively impact a viewer’s engagement (a viewer may see a still image or a black screen) and often indicate lost ad monetization. They typically play on LIVE video streams, either during part of the break or for the complete duration. As premium OTT publishers increasingly migrate to dynamic ad insertion on LIVE streams with server stitched ads, Ad Slates become a critical issue.

Why do Ad Slates run?

Ad Slates can run for multiple operational and technical reasons:

  • Server side ad transcoding: Server-side ad stitching systems need to transcode ad creative media files before they can construct a stitched stream. When an ad creative is seen by these systems for the first time, they kick off a transcoding process. Until transcoding completion, these systems will serve a configured ad slate instead of an ad.
  • Ad Server delays/timeouts: Ad Servers may not be able to respond to a large number of simultaneous ad requests -- especially during LIVE events with a large number of concurrent viewers. Ad server timeouts and errors result in Ad Slater insertion.
  • Programmatic demand: The primary Ad Server may auction the inventory to multiple programmatic demand sources, which may not result in a valid ad response. Even if there is a response, multiple wrappers might not allow the Ad Stitcher to unwrap and obtain the ad creative - resulting in an Ad Slate.
  • Lack of relevant ads: Ad targeting rules (to personalize ads) or frequency caps may stop the Ad Server returning any valid ads, leading to an Ad Slate.
  • More ad inventory than forecasted: Ads can be pre-sold based on ad impressions forecast. However, popular LIVE events may get more viewers than expected leading to a larger than expected ad inventory. After delivering the pre-sold number of impressions, an Ad Server may not serve any more ads leading to ad slates.

While there might be many more reasons for Ad Slates, it's critical for Ad Ops, Tech Ops and Product teams to quickly analyze and understand - during LIVE streams - large occurrence of Ad Slates. During LIVE stream, if the issue is quickly identified, Ad Ops & Tech Ops teams can take rapid corrective action. After the LIVE stream, the teams can investigate the root cause and put preventive measures in place to minimize future occurrences.

How to Monitor Ad Slates?

Conviva’s Ad Experience module (in the Ad Insights product), tracks Ad Slates (in real time) with two metrics:

  • % Slate Duration: This metric (aggregated across video streams) measures the duration of Ad Slate play against planned duration of Ad plays.
    Example: If LIVE streams were planned to play 90 seconds of ads during an ad break, but instead played 60 seconds of ads (two 30 second ads) and one 30 second Ad Slate, then this metric would report 33% (30 second ad slate / (30 second ad slate + 60 seconds of ads)).
  • % Slate Plays: This metric (aggregated across video streams) measures the number of times an Ad Slate played against the planned number of Ad plays. Example: If LIVE streams planned to play 3 ads during an ad break, but instead played 2 ads and one Ad Slate, then this metric would report 33% (one ad slate / (one ad slate + two ads)).

These metrics are available in the Analysis, Real-Time & MetricLens dashboards.


MetricLens summary dashboard enables a user to track Ad Slates by Video Asset Name, Device, Geo, OS and other dimensions. A user can also choose to be alerted when Ad Slates duration and number crossed a percentage threshold; these alerts are available in the Real-Time dashboard and can be sent via email.

Our Ad Slates monitoring and alerting platform is a general solution not dependent on any specific SSAI system - as long as you configure Ad Slate reporting during your integration, our platform will report on it. We support Ad Slates monitoring on major SSAI systems like VDMS, Uplynk, Google DAI, and YoSpace.

Looking ahead

In the coming quarters we plan to further enhance Ad Slate monitoring by adding details such as reason for Ad Slates. In addition, we'll soon be enhancing our AI Alerts to include Ad Slates.

Metric of the Month - Ended Plays

In video streaming ‘all good things come to an end’ is an ended play. The Ended Plays metric measures number of plays that ended during the selected interval. To count as an ended play, the viewing session within the interval must have at least one video frame that played. This metric counts only viewing sessions that played and ended.



There are three ways in which a session can end: gracefully ended with/without content completion, ended due to failure or forced termination (VPF), and ended due to a timeout since the last received heartbeat. Session timeout is determined by Conviva heuristics.

This metric shows the sum of the Ended Plays aggregated across the sessions with plays that ended in the selected time period.


Ended Plays is an 'Ended' type metric that is counted only in the interval in which the play ends. You can use the Ended Plays metric to monitor the number of viewers that exit a video or streaming event during different time periods and as a comparison metric to gauge overall viewer behavior. You can also view this metric in a distribution to analyze the impact of average bitrates, rebuffering ratio, and other metrics across ended plays.

Viewer Impact

From the viewer perspective, a watched video playback ends, either in response to viewer actions or as playback termination. A high number of Ended Plays prior to play completion is usually an indication of a drop-off in viewership due to either a lapse in viewer engagement or declines in overall player performance, such as increases in rebuffering, lower frame rates, or rendering issues.

Summaries and Correlation

For analysis, this metric is presented as a summary and time series, to display metric values as data points to baseline and highlight changes. For correlation, this metric is available in Custom Dashboards as a distribution of average bitrates, rebuffering ratios, average frame rates, and connection induced rebuffering ratios. These distributions can help you to understand the impacts of network and streaming performance that may be the cause of unexpected video play exits.

In addition, the Diagnostics page can be set to show a QoE metric, such as Average Bitrate, along with the Ended Plays metric and distribution to highlight the most common thresholds related to video exits.


You can use also MetricLens dimensions to filter this metric by metadata segments for enhanced root cause analysis. For example, limiting data to a specific ISP, video asset, or player dimension can highlight the source of changes in EndedPlays to a known provider, asset, or player type.

Adjustments

None

Visit the Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics.


Conviva-Facebook Partnership

In the 5 years since the Conviva Social Insights (formerly Delmondo) launch, we’ve made it our mission to provide the social analytics that matter most to premium publishers. As a result, the tools we’ve created are specifically designed to help publishers create better content, save time and drive revenue in the social ecosystem. We’ve also made it our mission to innovate by providing - nearly two years ago - the first cross-platform measurement solution to provide both analytics for Facebook Live and Instagram Stories.

We've now taken another leap forward and we are excited to announce that Conviva has been invited and accepted into Facebook’s new Media Measurement Partner program.

The Media Measurement Partner program is currently comprised of only 6 companies and designates Conviva as a world class, trusted third-party video measurement platform, endorsed by Facebook. We are extremely excited to be a part of this program since the partnership gives us the ability to further innovate our product and provide our clients with a richer set of competitive data than previously possible.

Please contact Conviva Support to request a demo that highlights the value Conviva Social Insights can bring to your organization.

Sensor Updates

We updated the following sensors:

30 April 2019

April Newsletter

Conviva Social Insights - Shoots and Scores


The Turner Sports and NCAA coverage of March Madness Tournament had record growth this year: over 100 million live streams and over 24 million live hours consumed, up 31% and 29% respectively vs last year. Conviva analytics not only supported this incredible growth but Turner Sports and the NCAA included Social Insights data in their joint press release, stating:

“Official NCAA March Madness social accounts produced a 94% increase in engagements vs. last year’s Tournament (Facebook, Twitter and Instagram). Videos across the three platforms generated over 135 million views, up 74%.“

This marks the first time in company history where a customer has attributed social and streaming data exclusively back to Conviva in a public press release. The ability to measure and compare viewer experience and data growth across all streaming channels, including social, is a capability that only Conviva can claim.

Using Social Insights to make better content

So, how would a publisher go about making such a significant increase in social engagement? When working to increase total engagement and views from one period to another, there are two things publishers could do:

  • make more content and/or
  • make better content: social posts with higher than average views/video or engagements/post.

Two key Conviva Social Insights metrics are specifically valuable in this scenario: View Conversion Rate and Engagement Rate.
View Conversion Rate (VCR) is defined as the percentage of impressions that turn into views on a video. It measures whether the video was engaging enough to capture the attention of the viewer. A higher VCR should deliver a higher number of views on a post.
Engagement rate (E/R) is the sum of all public engagements on the post divided by the total number of followers on the account, expressed as a percentage. E/R measures for how engaging a piece of content is.

Data in Action

By focusing on content themes that deliver a higher VCR and E/R from a previous period, brands can use data to inform and improve their content strategy. As an example, NCAA might increase content output based on individual players if such content showed an increased E/R or VCR compared to team content. Additionally, not all over-performing content is scalable. You can’t build a content strategy around game winning shot videos for example. However, if you were to see that celebrity appearances at games, celebrity brackets, and highlight-reels of star players performed better than most content, you can take those less obvious themes and build around them. These harder to isolate insights - which are driven by authenticated insights and can only be found by using a tool like Conviva Social Insights - should be the real focus of any content strategy.

As Conviva quickly becomes THE company that measures the entire digital streaming and video ecosystem, case studies like these serve as proof points for the growth Conviva can help fuel.

Conviva's Q1 2019 Report Now Available!

We just released our Q1 2019 insights on engagement, viewership, quality, and experience across the streaming entertainment ecosystem. Our analysis spans across 1 trillion real-time transactions/day, via 3 billion streaming applications on various devices, across 180 countries. Some of the most revealing and surprising insights include:

  • Streaming entertainment viewership is up 72% YoY, growing at a rate of consumption 49% faster in Q1 2019 than in Q1 2018.
  • The connected TV battle is heating up, Amazon Fire TV at #2 gained 7.2% connected TV device market share while Roku remains at #1.
  • Up to 47% of streaming ads are failing, signalling both tremendous risk and opportunity for publishers and brands.
  • Social media offers huge opportunity for real-time content promotion, distribution, and monetization, as viewers increasingly engage across multiple channels for peak events.

Metric of the Month - Video Playback Failures (VPF)

In the case of streaming impact, a video playback failure occurs when video play terminates due to a playback error. The Video Playback Failures (VPF) metric is an important measurement of service quality and audience engagement, especially when a large percentage of plays terminate due to video playback failures.



Two conditions must be met for the VPF calculation to count a session towards the VPF metric:

  • A fatal error is received from the Conviva device library after playback started (after the first i-frame is rendered).
  • The session ends and the total playing time from the playback error to the session end is not more than the last reported buffer length. Buffer Length is reported by device libraries. When not available, a default buffer length of 150 seconds (2.5 minutes) is used.

Conviva's VPF metric can be displayed as a sum or a percentage of the cumulative Ended Plays that terminated with a VPF error in the selected interval. The sum is a simple count of the video playback failures that occurred in the current interval. As a percentage, VPF is the sum of the Ended Plays with VPF divided by the sum of Ended Plays in the selected time period. Conviva uses EndedPlays instead of Plays to focus this diagnostics on ended plays and remove the bias inherent in considering total plays.


You can use the VPF metric to monitor the proportion of ended plays that terminate with video playback failure errors. A high VPF count or percentage, or a significant increase in this metric can be an early indication of declining overall player performance or the likelihood of increased fatal playback errors.

Viewer Impact

From the viewer perspective, playback failures terminate the streaming connection so the connection needs to be re-established before continuing with a video. A high VPF count or percentage, or a significant increase in this metric is also usually an early indication of declining overall player performance and the likelihood of increased fatal playback errors.

Summaries and Correlation

For analysis, this metric is presented as a summary and time series to display metric values as data points to baseline and highlight changes. For correlation, this metric is available in Custom Dashboards as a distribution of plays that ended with video playback failure. This distribution can help you to understand the impact of video playback failures and illustrates the play durations during which playback failures occurred the most.

In addition, the Diagnostics page can be set to show Video Playback Failures along with the Ended Plays metric to highlight the most common errors and the related streaming and performance issues.


You can use also MetricLens dimensions to filter this metric by metadata segments for enhanced root cause analysis. For example, limiting data to a specific ISP, video asset, or player dimensions can highlight the source of playback failures to a known provider, asset, or player type.

Adjustments

None

Visit the Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics.


NAB Show - Update

From Conviva’s signature cocktail party on opening night to the more than 50 sit-down meetings diving into deep conversations, demos, and hearing from our customers and prospects, another NAB has wrapped on one of our most successful U.S. shows to date!

We were excited to share more about our newest offerings Ad Insights and Social Insights, as well as a tech preview for customers to experience the latest expansion of Conviva's portfolio. Thank you to everyone who joined us! If we missed you, check out our short recap video and if you’re interested in hearing more about the latest from Conviva, we’d love to hear from you!

Sensor Updates

We updated the following sensors:

  • WinJS Windows 10 now supports DASH VOD and Live as well as APIs for manually sending seek, duration and bitrate events.
  • Android Ooyala now supports Ooyala Player SDK 4.43.
31 March 2019

March Newsletter

Alert Severity

At Conviva, we continuously drive towards greater data insights and actionability. Our AI Alerts enables automatic anomaly detection and key issue root cause analysis to prevent impact to your service. Last fall, we enhanced the actionability of CDN related AI alerts with the CDN ecosystem module. But alerts are only effective when they reach the right people at the right time.

We are now introducing alert severity awareness. This feature allows publishers to be notified of alerts based on severity level, and then direct the alerts to the right channels for action. You can now configure a single alert to send emails or chat notifications for relatively low-urgency issues, and to page the on-call engineer if the situation gets worse. For each alert state, we provide the list of impacted viewers which you can correlate with viewer reported issues as well as other external logs. You can configure 3 levels of severity information  (info, warning and critical), based on time of impact and number of impacted viewers.


With notification tiering, alerts can be tailored so that minor issues don't awaken anyone in the middle of the night, and no major issues are missed. With the right people getting the alerts over the right channels, you can reduce alert fatigue significantly!

As part of this update we have partnered with PagerDuty to offer out of the box integration that allows you to direct critical alerts to PagerDuty tools.

We are excited to bring you this capability and drive actionability further in our alerting functionality.


Metric of the Month - Connection Induced Rebuffering Ratio

Rebuffering occurs when the video stalls during playback and the viewer must wait for the video to resume playing. Various connection, bitrate, and player issues can lead to these stalls, but so can user initiated seek events. As a result, it is important to distinguish between seek-initiated rebuffering and non-seek rebuffering. Conviva’s Connection Induced Rebuffering Ratio (CIRR) metric refines rebuffering analysis to measure only non-seek or system-induced rebuffering.



The nonseekRebufferingTime in a video stream is the total time during which the reported player-state is "buffering", and excludes the buffering time before the first frame of a video is displayed, buffering following user-initiated seeks (S1 in the diagram) and any paused time. To clearly distinguish the source of rebuffering, CIRR only counts rebuffering that starts 10 seconds or more after a seek start event as connection induced rebuffering. Otherwise, the rebuffering is counted toward seek-induced rebuffering and included in the Rebuffering Ratio metric.

Conviva measures connection induced rebuffing as the percentage of total viewing time (playingTime and rebufferingTime) during which nonseekRebufferingTime occurred in the selected period. Using all rebufferingTime in the denominator of this calculation enables a more precise measurement of system-induced rebuffering over the viewer’s total viewing time.



Viewer Impact

From the viewer perspective, the viewer impact is similar to the Rebuffering Ratio metric-- the video stalls during playback while the viewer endures the spinning wheel on screen and waits for the video to resume playing. A  high Connection Induced Rebuffering Ratio or a significant increase in this metric often indicates an overall decline viewer experience and video streaming performance, often leading to an increase in the likelihood of viewer abandonment.


Summaries and Correlation

For analysis, this metric is presented as a summary and time series to display metric values as data points to baseline and highlight changes. For correlation, this metric is available in Custom and Engagement Dashboards as a distribution of play duration and Ended Plays.


In addition, the Diagnostics page can be set to show Connection Induced Rebuffering Ratio along with the Rebuffering Ratio metric to highlight the impact non-seek, system-induced rebuffering.



You can use also MetricLens dimensions to filter this metric by metadata segments for enhanced root cause analysis. For example, limiting data to a specific ISP, video asset, or player dimensions can highlight the source of changes in Connection Induced Rebuffering Ratio to a known provider, asset, or player type.

Adjustments

To limit the impact of prolonged rebuffering, if a session has continuous nonseek rebuffering for 30 minutes, the play is treated as expired.

Visit the Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics.

Conviva at NAB


Meet us at Conviva’s Fairway Villa - Wynn Las Vegas
Monday, April 8 – Wednesday, April 10

Learn how Conviva, the real-time measurement and intelligence platform for streaming TV, enables you to drive engagement, increase monetization, and improve the viewer experience. Make your way to our Fairway Villa at the Wynn Las Vegas to hear more about the latest innovations in streaming. 

Request a meeting to talk to our team, we're looking forward to seeing you.
 

Sensor Updates

This month we updated the following sensors:

28 Feb 2019

February Newsletter

Integration Survey

Valued customers, we request your valuable feedback to help us understand your experience while integrating our Sensor SDK / libraries within your OTT video applications. We would really appreciate it if you could fill out the survey (1-3 mins) located at Conviva Integration Survey.

Ad Insights Introduces Ad Frequency/Unique Device Metric

Measuring Ad Frequency and Reach is critical to ensure that the ad messaging reaches the targeted audience and drives brand recall. Advertisers require strict Ad Frequency Caps and Ad Ops configures these both on the buy side and sell side ad servers. However, for multiple technical and business reasons the Ad Frequency Cap is frequently not honored. Repeating the same ad to the same viewer in a short time window leads to viewer “fatigue” and negative perceptions of the ad. It may even cause the viewer to abandon watching the content all together, leading to customer churn.

Conviva’s Ad Insights recently introduced a new metric Ad Frequency/Unique Device, which enables Publishers to monitor the ad frequency per unique device (a proxy for a unique viewer).


At an aggregate level, this metric measures the average number of ads played on devices in a certain time window (daily, weekly, monthly). In the Analysis dashboard above, the example shows that on average 8 ads played per device per day.

Using the MetricLens dashboard, this same metric can be used to identify repeated ads for specific dimensions, such as Ad Id, Ad Creative Id, and Ad Position.


In the example above, certain ads (identified by Ad Id) repeated on average over 5 times per day. You can also use the MetricLens timeline to further understand the time windows when ads repeated the most.



In this example, repeated ad impressions at 2 AM and 6 AM caused the Ad Frequency/Unique Device metric to exceed 2 ads per viewer, contributing to the daily average of almost 6 repeated ads per viewer.

Why do ads run repeatedly to the same viewer and what can you do about it? There could be several reasons: 

  • When ads are personalized for a viewer (e.g. Dynamic Ad Insertion), there may not be enough unique available ads for all viewers. The Publisher may need to identify additional demand sources to diversify the ads served to the viewers (e.g. Programmatic ads).
  • For server side ad insertion streams, there may be a problem transcoding and inserting specific ads into a single stitched stream. That might cause an alternative ad to be inserted into the stream which might be a repeated ad. In this case, the Publisher should diagnose the root cause in the Ad Stitcher and/or Ad Server to identify failing ads.
  • There may be a Frequency Cap misconfiguration on the Ad Server. In this case, Ad Ops should look at the Frequency Cap setting both at the Line Item and the Ad Creative level.
  • Multiple Ad Servers may be serving the same Ad Creative to the viewer. Since the Ad Frequency cap is per Ad Server, each one does not have visibility into the creative served by the other.
  • Viewer cookies may be getting churned (or were not available) resulting in repeated ads -- this could be because Ad Servers may be using cookies to enforce Frequency Caps.

Measuring Ad Frequency is an important first step to ensure viewer engagement and meet Advertiser commitments. Conviva’s real-time dashboards can help Publishers understand when there may be an issue enforcing frequency cap limits and to take the appropriate action: resolve the underlying technical issue, ad server configuration issue or business issue such as increasing demand sources.

Metric of the Month - Rebuffering Ratio

As consumers watch more and more video, viewers are becoming less and less tolerant of video quality issues. The percentage of time they experience buffering versus actually watching content is one of the main forward indicators of viewer engagement. 

Conviva measures this important Quality of Experience (QoE) metric asRebuffering Ratio — the percentage of total video viewing time (playTime + rebufferingTime) during which viewers experienced buffering. Both play time and rebuffering time are determined based on player state ("playing" and "buffering"), playhead movement, and frame rate progression. Rebuffering time includes both seek and non-seek rebuffering, but does not include the buffering time before the first frame of a video is displayed or elapsed time while the video is paused.



Within a video session, Conviva calculates the session Rebuffering Ratio as the total rebuffering time divided by the total video viewing time of that session.



Viewer Impact

From the viewer perspective, a high Rebuffering Ratio or a significant increase in this metric often indicates an overall decline viewer experience and an increase in the likelihood of viewer abandonment. No one expects a video play to need to stop and load more video, especially during action-packed moments of high engagement.

Session & Aggregation Levels

For interval measurements, we aggregate the session-level rebuffering that occurred during the interval sessions into an aggregate total for that interval.



 

Summaries and Correlation

For analysis, this metric is presented as a summary and time series to display metric values as data points to baseline and highlight changes. For correlation, this metric is available in Custom and Engagement Dashboards as a distribution of play duration and Ended Plays.



In addition, the Diagnostics page can be set to show Rebuffering Ratio along with Video Start Failures and Average Bitrate metrics to highlight the impact of performance and quality standards between video starts, video quality, and buffering time.




Adjustments

To limit the impact of prolonged rebuffering, if a video session experiences continuous rebuffering for 30 minutes, Conviva expires the video session.

Visit the Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics.

1. The current industry mean Rebuffering Ratio is 1.12 percent (Conviva’s 2018 Annual State of the Streaming TV Industry).

Sensor Updates

We have created a new page on the Community to provide you with a quick reference of all releases. Please review at Platforms Releases link, under the Platforms menu.

This month we updated the following sensors:

31 Jan 2019

January Newsletter

Announcing the Next Generation of Ad Insights 

We are very excited to announce the next generation of our Ad Insights product with the new Ad Experience capabilities. Premium publishers can now leverage our real-time measurement and diagnostics to identify root causes for lost ad impressions and poor ad creative quality.



Ad Experience Analysis dashboard: Summary metrics from ads requested to completed ads. The “Ad Delivery” tile tracks lost impressions due to failures, the “Ad Experience” tracks poor viewer experience due to creative quality and the “Ad Completion” tile tracks ads viewed to completion.

OTT Video Advertising Growth

Our latest 2018 census report showed that streaming TV viewing hours grew 89% and the number of plays grew 74% last year. Naturally, advertising revenue is growing with these viewers, with U.S. digital video ad revenues on track to exceed $13.9 billion in 2018 (S&P Global Market Intelligence).

Our Ad Insights (Ad Breaks) product has tracked this significant advertising growth. In Q4’18 alone Ad Insights monitored over 50 million advertising hours - a 16% increase over the previous quarter. The average viewer watched 8 ad breaks per day (with one or more ads in a break), with an average duration of 64 seconds per break.

OTT Advertising Operational and Technical Challenges

Despite this growth, premium publishers are still learning how to best monetize and operationalize ads as they continue to invest in streaming TV. Common challenges include the transition from Linear TV to digital advertising technologies, adoption of programmatic ad workflows and managing client-side and server-side ad insertion.

There are multiple failure points in the complex ad delivery ecosystem: starting with video applications (e.g., failing ad requests) to ad servers (e.g., decisioning and waterfalling), creative delivery (e.g., site-served vs 3rd party creatives) and dynamic ad insertion issues. At the same time, ad delivery standards for Streaming TV are still evolving, including VAST 4.1 finalization, ad verification approach (i.e. Open Measurement), impression tracking for server-side ads and viewability on streaming apps.



Real-Time Measurement and Issue Diagnosis

Ad Insights, with its Ad Experience module, empowers a publisher's Technical and Ad Operations teams to identify and diagnose failing ad attempts or playback issues, at the individual ad video level. They can narrow down the root cause to the offending entity in the ad ecosystem: the application, ad server or creative media. With the real-time dashboard and alerting capabilities, they can react quickly to minimize impression loss or viewer abandonment caused by poor ad creative quality.


Diagnosing Lost Impressions


Ad Failure Diagnostics in Analysis dashboard: Select a time window to identify ad start attempts with the most failures and the associated error messages. Identify ads with dimensions such as Ad Id, Ad Name, Creative Id etc.

Diagnosing Creative Quality (Impacting Viewer Engagement)


Ad Rebuffering Diagnostics in Analysis dashboard: Select a time window to identify ad creatives with the most buffering - these have creative quality issues. Identify ads with dimensions such as Ad Id, Ad Name, Creative Id etc.

Track Ad Viewing Experiences Across Major Platforms

Our Ad Insights SDKs are available for:

Reference implementations of the Ad Insights SDKs are available for:

  • Server-side: Google DAI, Verizon Uplynk
  • Client-side: Google IMA SDK, Freewheel SDK

Closing

As publishers increase their investment into ad monetization for streaming TV apps, we strive to help you navigate the complexities of the ad ecosystem by providing real-time diagnostics that maximize monetization, minimize lost impressions and monitor ad video creative quality.

Metric of the Month - Concurrent Plays

We ended last year with a look at the Plays metric, so let’s now continue the monthly metrics with a look at the maximum number of plays in an interval, the Concurrent Plays metric. Concurrent Plays measures the maximum number of simultaneously active sessions during a given interval. Active sessions have at least one played video frame, and do not include unsuccessful play attempts or sessions that have ended. 

Concurrent Plays is a time-trended metric with multiple data points during each interval. The data points measure the number of concurrent videos playing or simultaneously active sessions in 5-second data buckets throughout the interval. The maximum number of simultaneously active sessions from any of the 5-second data buckets determines the number of Concurrent Plays during that interval.



For real-time data, the Concurrent Plays metric is the maximum number of active sessions in the current second. The Concurrent Plays metric only counts active sessions where the video has started playing before or during the data bucket, including sessions with rebuffering, pause, and other trick modes that do not exceed a defined inactively limit.

Viewer Impact

From the viewer perspective, the Concurrent Plays metric shows the maximum number of attempts to play video that result in at least one played frame of video in the current second (real-time data) or in any 5-second data bucket ­­­­­­­­­­­­within the selected int­­erval (historical data).

On the Real-Time and Custom dashboards, the Concurrent Plays Map also shows the maximum number of active sessions in the current second, so you can easily view real-time data at the city, state, or country level. 

Session & Aggregation Levels

For aggregated intervals, Concurrent Plays is a snapshot of the maximum number of simultaneously active sessions during the selected time range.

 

Summaries and Correlation

For analysis, this metric is presented as a summary and time series to display metric values as data points to baseline and highlight changes. For correlation with the different viewing patterns, the Diagnostics page can be set to show Concurrent Plays in a customizable time series to compare viewership trends across different time periods.




In addition, you can use MetricLens dimensions to filter this metric by metadata segments for enhanced root cause analysis. For example, limiting data to specific ISP and player dimensions can highlight the changes in Concurrent Plays and peak viewership to a known provider, player type, browser version, or other dimension.

Adjustments

This metric does not count unsuccessful attempts. Also, Conviva sanitizes the active sessions to eliminate prolonged periods of rebuffering and other modes that may indicate the viewer is no longer engaged with the video.

Visit the Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics.

Conviva @ CES



After another very successful CES for Conviva, all we can say is thank you!

Our suite at the Nobu Hotel was a constant bustle as our executives sat down with top customers, prospects, and ecosystem partners to dive into how we can continue to improve streaming for our customers and their customers. It was our best CES yet, and a great kickoff to the year ahead! Special thank you to our customers who took time to share their thoughts with us, for the very warm welcome of our newly launched Ad Insights, and for the amazing feedback provided during our days at CES.

As we expand our portfolio to encompass social media video with Social Insights and continuous streaming video ad measurement with Ad Insights, we were excited to share with customers how we can truly be an end-to-end solution.

Sensor Updates

As of  February 1st, 2019, the following Sensors will be entering the "End of Support" phase:

  1. Roku Native
  2. Xbox 360 Lakeview
  3. Xbox 360 Primetime
  4. PS 3 Touchfactor
  5. Samsung Orsay
  6. Adobe Primetime HTML5 TVSDK
  7. Flash Native
  8. Flash OSMF
  9. Flash Primetime
  10. Microsoft Silverlight

These Sensors are being discontinued as part of our product lifecycle review process to manage technical debt, align with market trends and accelerate future development. For customers using the impacted media players, we offer alternate Sensor options on the Conviva Community website, at the relevant Experience Insights integration pages. Please reach out to your Conviva Solutions Consultants for further assistance.

31 Dec 2018

December Newsletter

 What an Epic 2018! 

2018 was a fantastic year at Conviva as we keep innovating and providing the Streaming TV Ecosystem with solutions that deliver the best possible viewing experiences to connected audiences across the globe. Our 2018 highlights include:

Team

We welcomed our new CEO, Bill Demas, and we substantially expanded our team across all regions, including Silicon Valley, New York, Bengaluru, Beijing, London, Singapore, and Madrid.

Products and Innovation

We had several exciting product announcements in 2018.

We enhanced our award-winning Video AI Alerts with more granular root-cause analysis and automated workflow capabilities via Webhooks integration. The improved root-cause analysis pinpoints issues closer to the source and Webhooks support now enables our customers to directly integrate viewer-impacting alerts into their incident workflows. 

We also launched the Ecosystem Module. This enhancement to Video AI Alerts enables publishers to resolve streaming quality degradations faster by routing alerts in real-time and with granular diagnostics information directly to Content Delivery Networks (CDNs).

We enhanced Precision, our multi-CDN optimization solution, with Saved-session Reporting and Policy Manager controls to preemptively mitigate video delivery degradations. Our customers can now update policy configurations in real-time and get visibility into Precision actions to protect viewers.

Finally, we are extremely excited with our first-ever acquisition. Delmondo joined the Conviva family in the fourth quarter of 2018. As a result of the acquisition, we announced Conviva Social Insights that provides streaming TV publishers with audience and content intelligence from all major social media platforms. This acquisition fits directly into our strategy to provide a platform for real-time measurement and intelligence for streaming TV. With Delmondo, we are the first company to provide full streaming TV measurement across direct-to-consumer and social video based distribution. We are very excited with the prospects of bringing these two datasets together to provide new and unique insights into streaming experience and behavior.

On top of this, we have been working tirelessly to incubate multiple new products that you will start seeing right from the beginning of 2019. Stay tuned!  

Platform and Events

We blew through our previous platform concurrency records by reaching 9.12M viewers during the Croatia vs France World Cup game. This far exceeds our previous record of 5.5M during Super Bowl 2018 and 3.29M during the previous World Cup in 2014.

We also saw tremendous traffic increase with over 63% viewing hours growth from Q3 2017 to Q3 2018. We are excited to see this level of rise in streaming video and look forward to more in 2019.

Awards, Awards and Awards

We are now officially certified as a Great Place To Work, selected as one of the Top 50 Companies that Matter the Most in Online Video by Streaming Media. We also won several awards for our products: Precision was recognized by TV Connect as The Best Network Optimization Implementation and the Experience Insights Ecosystem Module won Streaming Media Europe’s Best Streaming Innovation and TVB Europe’s best of IBC awards.  

Thought Leadership and Research

It’s a Conviva tradition. We are proud to illuminate the Streaming TV Ecosystem with periodic reports reflecting the trends and insights gathered from publishers across the globe.

In 2018, we published four all-screen streaming TV census reports: Census Report: full year 2017, Census report: Q1 2018, Census report: Q2 2018 and Census report: Q3 2018. Please be on the lookout for our full-year 2018 census report in January 2019!

We also partnered with nScreen media in 2018 to publish Part 2 of our highly anticipated Secret Life of Streamers study that revealed eye-opening findings as it relates to evolution of streaming video consumption.

30 Nov 2018

November Newsletter

 Introducing Conviva Social Insights 


Last week we publicly announced the acquisition of Delmondo. Founded in 2014, Delmondo aggregates video consumption and audience data across Facebook, Instagram, YouTube, Twitter, and Snapchat. Delmondo is renamed Conviva Social Insights and is now a new product within our expanding product lineup.

Delmondo earned a reputation as the leading cross-platform social video analytics solution for top-tier publishers including Turner, Viacom, WWE, ABC News, Fox Sports, and NASCAR. We inherit Delmondo’s designation as official Facebook Media Solutions Partner and expand our customer base and capabilities across social media video streaming measurements.

Using Conviva Social Insights, streaming TV publishers can maximize engagement and viewership across a variety of social video platforms to drive subscriptions and viewership or build engagement with content. Additionally, by understanding viewer demographics and behaviors across social media video, our customers can identify unique revenue opportunities in native advertising, sponsorships and social media partnerships.

Conviva Social Insights receives all consumption and audience data through APIs provided by Facebook, Twitter (GNIP), and YouTube. The data is aggregated and made available to customers via an intuitive front-end analytics solution. While some of the APIs are publicly available, our customer relationships and the social media platforms themselves give us access to a richer set of data points, including real-time Facebook Live and Instagram Stories intelligence.

The following content types are measured once a customer’s account is authenticated by Conviva and the social media platform:

Facebook - Posts, Video, Live Video, 360 Video, Facebook Watch Show Pages
Instagram - Posts, Videos, Carousels, Stories
YouTube - Video
Twitter - Tweets (including replies and retweets), Videos, Photos
Snapchat - Organic Profiles, Discover Channels

As we continue to expand our full suite of measurement tools for streaming TV, tying viewer-centric social video consumption data with session-based analytics offers our customers unprecedented streaming video intelligence. For more information, visit Conviva Social Insights or check out our Press Release.

Metric of the Month - Plays

Now that we have explored the definitions and details of Conviva’s startup metrics (Attempts, Video Startup Time (VST), Exits Before Video Start (EBVS), and Video Start Failures (VSF)), let’s look at the metric most visible to viewers, Plays.

A play occurs when a viewer sees the first frame of a video.


The Plays metric counts the number of plays that started in the current interval, not including unsuccessful play attempts. Unsuccessful play attempts count as Video Start Failures (VSF) or Exits Before Video Start (EBVS).  

As a percentage, this metric shows the percentage of Attempts that resulted in plays in the selected interval.

Plays is a 'Started' type metric and is counted in the interval in which the play started.

Viewer Impact

From the viewer perspective, the Plays metric measures the number of attempts to play video that result in at least one played frame of video ­­­­­­­­­­­­within a specified time int­­erval. A play results in a video being viewed until an Ended Play occurs or an early termination, for example Video Playback Failures. 

Session and Aggregation Levels

For session calculation and aggregation, Conviva provides a simple sum of all the Plays in the selected period.

Summaries and Correlation

For analysis, this metric is presented as a summary and time series to display metric values as data points to baseline and highlight changes. For correlation with the different startup metrics, the Analysis Window in the Diagnostics page can be set to show the Plays (number and percentage) along with the other startup metrics to compare startup behaviors and successful Plays across different dimensions, such as operating system, browser, and CDN.



In addition, you can use MetricLens dimensions to filter this metric by metadata segments for enhanced root cause analysis. For example, limiting data to specific ISP and player dimensions can highlight the changes in successful Plays and startup behaviors to a known provider, player type, or browser version.

Adjustments

None.

Visit the Conviva Metric Dictionary in the Pulse Help Center to learn more about Conviva's metrics. 

Sensor Updates

The Conviva Sensor Product team is excited to share our November sensor updates for the following platforms: