Evolving video content recognition and the analytics for data driven marketing all in real-time. Bridging linear & non-linear data to optimize cross-channel marketing campaigns connecting high intent users with brands/marketers.
2. No real-time data.
No real-time content recognition.
Industrial-age“analog”analysis tools.
Very difficult to target ads in a time-sensitive
fashion.
Very difficult tomeasure ROI.
Very difficult to obtain intelligence on
competitor’s campaigns.
The problem
TV broadcast advertising flies “blind”
3. THE PROBLEM
TV broadcastadvertising is getting killed by
innovationindigital analytics:huge data gap
REAL-TIME & AUTOMATED
23 years of rapid advances
NOREAL-TIME DATA!
60 years of stagnation
4. The solution
The IDENTV Media Intelligence Platform
A full suite of proven automated content recognitiontechnologies.
Brings TV broadcast advertising intothe digitalage. FINALLY.
5. FEATURES (All Over-the-Top without any intrusive watermarking)
Real-Time VideoContent Detection
Real-Time Historical Media Analytics
Built-In Customized Reporting and Push Alerts/Reports with webhooks for
integrationwith Programmatic Ad Insertionand attributionanalytics capabilities
Real-Time and Retroactive Content Recognition processing of recorded streams for
any number of days/months of Online/Linear Video
CompetitorCampaignTracking
The SOLUTION
The IDENTV Media Intelligence Platform
7. THE PROBLEM
BST had little luck moving their Mobile App
client budgets from costly and saturated
digital/online marketing ($5Mn/mo. Facebook)
to high-return geo-targeted spend via TV
campaigns.
The pipeline of cash rich Mobile App clients cited
the disparity between Digital & TV data as the
leading reasonbehind the hesitance.
USE CASE
bringing TV to the digital age for BSTand their clients
8. THE SOLUTION: IDENTV BRIDGES THE GAP
Using the platform’s features to provide
standardized TV data delivered in real-time,
allowed clients to begin using it for cross-channel
decision making and thus re-allocating spend to
TV!
RESULTS: Real-time data enabled instant analysis
& targeting that led to lift of 20% in downloads
and 16% in usage over thebaseline.
Use case
bringing TV to the digital age for BSTand their clients
10. Gets rid of manualand unverified processes
Automated Logging with direct integration
Real-Time Spot Tracking & Live Post-Logs
High accuracy of 99.5% leads to:
• No delayed logs
• Reduced A/R
• Reduced operating inefficiencies
VALUE PROPOSITION
for Sell-Side, Stations/Networks:
Vastly improved workflows
12. HOW DOES THE ADS/CLIPS TECHNOLOGY WORK?
DETECTS
Detects ad breaks, segments and
identifies unique ads, (which currently
are manually tagged) with metadata,
tracking the majority of Ads airing at a
given time in the U.S.
SEARCHES
Industry leading search component
compares millions of grouped data
points simultaneously using a vector
based multi-dimensional search engine.
MATCHES
Simultaneously monitors hundreds of
live channels for thousands of indexed
ads and returns matches on ad-airings
at close to 100% accuracy in real-time.
17. Optical Character Recognition & Closed Caption Extraction –
Integration in progress, CC if available – in use
Automated Speech Recognition & transcription with translation,
subtitling & even captioning services– in use
Audio Fingerprinting – module ready in testing
Image Fingerprinting & Scene Detection – In use and training
new/better models
Facial Recognition – Prototyped preparing integrations
Logo Learning & Detection - Prototyped preparing integrations
Object Recognition – Expected at end of Q2 2016
Speaker Identification & NLP/Semantic Technologies –
Prototyped and some in use
End Objective: To automate the collection of video assets and
automated uploading processes through the life cycle extracting
and enriching metadata inreal-time.
Product roadmap
building the intelligent video container
for video based search engine
20. Value Proposition for entire ecosystem: fulfilling the promise of ProgrammaticTV
IDenTV willplaya key role in
enabling the establishment of
programmatic marketplaces in
television.
OBJECTIVE
IncreaseROI of Brands/marketers,while booking sellers open inventory toclose
the loop.
22. THE MARKET
MANAGED IP VIDEO, THE GIANT THAT CAN’T BE IGNORED
Managed IP video is IP traffic
generated by traditional
commercial TV services
This traffic remains within the
footprint of asingle service
provider, so it is NOT considered
Internet traffic
Source: Cisco VNI, 2015
23. THE MARKET
THE COMINGEXPLOSION OF GLOBAL VIDEO
The sum of all forms of video
(TV, video on demand [VoD],
Internet, and P2P) will be in
the rangeof 80 to 90
percent of globalconsumer
traffic by 2019.
Consumer VoD traffic
willdouble by 2019.
Source: Cisco VNI, 2015
25. Undefined content - Live TV channels
• Indexed in real time
• It is transformed into “recordedcontent”
Variant content -Games
• Splash screens are indexed
• It is taken as “recorded content”
Recorded Library content - VOD, SVOD,
YouTube
• This is the main objective of the solution
• The worst case is taken into consideration…
The scope
INDEXING APPROACH
26. Quick facts
ON INDEXING
Indexing Live Content
• Indexing real-time channels consumes less
than acore (x86 laptop).
Indexing Regular Content
• One hour is processedby ax86 core under
10 minutes.
Every indexed content is availablefor
searching within 30 seconds
27. Approximately 2,000 years of video
Growth: 1 additionalyearevery 15 minutes
(increasing rate)
Assuming an averagedurationof 3 minutes,
there are 350,400,000 videos (350M).
The scope
INDEXING THE WORST CASE