opening pandor’s set top box - rsmb ltd...opening pandor’s set top box: from concept to practice...
TRANSCRIPT
OpeningPandor’s settop box:from conceptto practice
Todd Juenger (TiVo)Jim Kite (StarcomMediaVest Group)
November 6th 2009
Today’s Themes
• An overview of the current digital set top box mediameasurement landscape
• How the STB data debate is evolving in the US• MediaVest’s use of TRA data in 2010 Upfront• TiVo cross-media measurement with Quantcast• Conclusions and predictions
Digital Set Top Box data and media measurement:Who is doing what
Around the World
• TNS• UK Skyview (33k)• South Africa (4k)• South Korea (4k)• Australia (10k)• New Zealand (6.5k)
United States
• TNS• DirecTV (100k)
• RenTrak• AT&T U-verse (4mm)• Dish (11mm)
• TiVo• TiVo (375k)
• TRA• TiVo (1.7mm), Charter (300k)
• Google• Dish (5mm)
• TNS, RenTrak, Nielsen, TRA• Charter 300k
From Nielsen alternative to marketing effectiveness tool
Nielsen
• Age/Sex Demos(with NAD breaks)
• MxM & TimeShifted Data
• Panel Based
1st Gen.Set-Top-Box
• Sec x Sec Data
• LA Market Only-(300k homes)
• Household dataonly
Single Source
• Sec x Sec data
• National +….2MM Homes
• ‘matched’with same hh
level databases
The True ROI Accountability for Media (TRA) Architecture
MEDIA PURCHASE DEMOGRAPHIC
ADVERTISERS AGENCIES
Privacy Shield (no names/addresses)
1.5mm TV National HHs*Cable, Satellite, OTA
54mm HHs8 Supermarket Chains
100mm HHs250 fields of data
NETWORKS
Media TRAnalytics™ SaaS Reports (ISO 27001 certification)
data data data
AudienceMeasurement
MediaPlanning
Return onInvestment
370,000++ Single Source National HHs370,000++ Single Source National HHs
TRA lines up with currency
TRA/Nielsen Rank Correlation Overall Eight Broadcast Networks Applicable Programming National Broadcast Week of November 17, 2008(Broadcast Networks Include: ABC, CBS, FOX, NBC, CW, MNT, Univision, Telemundo)
TRA/IRI Rank Correlation - TOTAL U.S. – Food, Calendar Year 2008 ending Dec 28, 2008, Based on Volume Share
Ratings Correlation
Brand Correlation0.871 0.837
0.959 0.916 0.844 0.850 0.879
0.0
0.2
0.4
0.6
0.8
1.0
Coffee Hand & BodyLotion
LaundryDetergent
RTE Cereal Yogurt Shampoo Average
0.879 0.898 0.857 0.882 0.864 0.912
0.7670.870
0.000.100.200.300.400.500.600.700.800.901.00
17-Nov 18-Nov 19-Nov 20-Nov 21-Nov 22-Nov 23-Nov Weekly
MediaVest/TRA Partnership
• Set-top box viewing data from nearly 2 million cable/TiVo Households• Robust demographic & psychographic data from Experian• A sub-set of 370,000 homes with real product purchasing data from
frequent shopper card data
• As charter subscribers, MediaVest has helped guide the selection of productcategories:
• Cereal - Hand & Body Lotion• Soft Drinks - Laundry Detergent• Beer - Candy/Chocolate• Coffee - Cookies/Crackers• Yogurt - Pet Food• Shampoo - Paper Towels/Tissue
Media Viewing Behavior
Purchase Behavior
Page 8
Audience Measurement(national + live + time shifted)
Media Selection:Using Purchaser Targeting
ROI Validationof Media Plan
Use ROI for improved media planLinear AdsAddressability
Delivering continuous insights
TRA changes three areas of media measurement
1. Audience &Advertising Analytics
2. Targeting
3. True ROI
Audience / Ad Analysis
• More granular audience analytics• Measurement beyond what Nielsen sample can measure• Also accounts for variables beyond min x min data
% Switch Away: Percentage of Households tuned at first frame of an ad who tunedaway before last frame of that same ad
% Perfect Play: Percentage of Households tuned at first frame of an ad whoremained tuned in to the last frame of this ad.
% Jump In: Percentage of Households not tuned at first frame of an ad who tunedin before last frame of an Ad
Audience Retention Report(Daypart Example)
Improved targeting potential
• Targeting that leverages two massive databases
• Experian• Robust household-level demographic, lifestyle, ownership data from
Experian’s National consumer database• E.g. car buyer behavior, pet ownership, home furnishing enthusiasts,
home value, type of residence, etc
• Shopper Data• Real purchase-based targeting via shopper-card data• Light/medium/heavy category users• Brand users v competitive users• Light brand users v heavy brand users• Geo-targeting down to the zip code level
Target Report (Example)
Top Broadcast Network Programs – Soft Drink Category
Network Program Index Avg. Rtg.
FOX Chaotic 169 0.87
CBS CBS Morning News 163 1.74
NBC Early Today 154 1.40
FOX Entertainment Studios.com 153 0.97
FOX This Week in Baseball 152 0.82
FOX Judge Mathis 150 2.29
CBS CBS Up to the Minute 149 1.46
FOX Frasier 146 1.40
FOX Sherlock Holmes in the 22nd Century 145 2.04
FOX Entertainment Studios.com 144 1.26
True ROI
• TRA provides real-world, accurate link between advertising exposure &changes in consumer behavior
• Measures same HH’s viewing/ad exposure against that HH’s purchasing
Advertising
They did notbuy Brand X, but they
did buy Brand Y
Theysaw advertising for
Brand X
We know that theybought Brand X after
seeing ads for Brand X!!
ROI Report Example
Brand: Cereal
Target Audience: Married Home Owners
In Tab Households: 1595
Of thosereached, 7%
bought more ofthis product inthe past month
This equatesto an ROI of
$6.10Those 7%,accountedfor $10,371
in sales
ABC’s ROI was8% higher than
the averagenetwork
What this all means
• Provide more granular TV audience data• Larger samples• Second by second data• Improved ad-viewing measurement
• More targeted TV inventory evaluation & selection• Find the inventory that reaches certain types of purchasers, multi-
layered demographic segments, etc.
• ROI/Effectiveness measurement & evaluation• Prove ROI delivery by schedule, station, target, etc.• Use past ROI data to guide inventory selection
2010 UpfrontTRA Swing-Purchaser Guarantee Parameters
• Network/MediaVest to use historical TRA data to establishTRP benchmarks among “Swing-Purchasers”• Based on network delivery – by network/daypart
• For each brand campaign, Network to deliver +20 pointindex improvement vs. historical norm
• 1+ reach equivalency – min. 100 index on 1+ reach vs.demo target reach
• Network/MediaVest agree to track & report on ROIperformance for each campaign• To set ROI norms/lay groundwork for future discussions
TRA Swing-Purchaser Guarantee Parameters
High category purchase (spend $8+ onRTE Cereal in selected timeframe); lowloyalty (<50% brand SOR)TV Buying Target = A45+
Top 10 indexing programs among Grape NutsSwing-Purchasers
Top 10 Based on Swing Purchaser Index* Ntwk HH Rtg Index*
Shark Attack : Megalodon SCFI 0.17 533.1
Dragonheart SCFI 0.15 379.4
Shadow Man USA 0.16 323.6
It Waits SCFI 0.12 317.7
Taxicab Confessions : All's Fare in Love & Vegas BRAV 0.10 271.2
Children of the Corn IV: The Gathering SCFI 0.15 266.9
Dungeons & Dragons: Wrath of the Dragon God SCFI 0.26 253.9
Debt to Wealth USA 0.13 235.1
Three Wise Guys USA 0.19 190.4
Supergator SCFI 0.13 190.1
*Index vs. HH Rtg**Notes: Minimum HH Rtg requirement = 0.10Excludes NBC Network, Telemundo, Sleuth
0.0
10.0
20.0
30.0
40.0
50.0
60.0
Swing Purchaser 1+ Reach
Buying demo 1+ reach
Swing Purchaser 1+ Reach 0.0 11.6 21.1 28.3 33.5 37.8 41.4 44.4 46.9 49.1 51.1 52.8 54.3 55.7 56.9 58.0 59.0 59.8
Buying demo 1+ reach 0.0 14.2 23.3 29.9 34.9 38.9 42.3 45.2 47.8 49.9 51.9 53.6 55.1 56.5 57.7 58.7 59.7 60.6
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350 375 400 425
Source: TARDIIS / Nielsen Data 4 weeks starting week of 1/31/00 and 2/7/05
Increased TRP’s against Purchaser HH, whilemaintaining equal 1+ Reach
Primary & Secondary 1+ reach guarantees
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
Swing Purchaser 1+ReachBuying demo 1+ reach
Swing Purchaser 1+ Reach 0.0 11.7 21.1 28.5 34.8 40.2 44.5 48.2 51.6 55.6 59.0 62.4 65.1 67.2 69.1 70.7 72.3 73.5
Buying demo 1+ reach 0.0 14.3 23.5 30.3 35.3 39.6 43.3 46.5 49.2 51.5 54.0 56.1 57.9 59.6 61.1 62.6 64.6 66.0
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350 375 400 425
Secondary Guarantee – Increased TRP’s againstPurchaser HH, and increased 1+ Reach
• Quantcast & TiVo, two leaders in online and TV media analysis arecombining forces
• New joint program offers unprecedented levels of measurability andaudience insights between television advertising and onlinecustomer activity, using anonymously joined sample
• Analysis measures cross-platform dynamics between TV andInternet, including: overlap (cannibalization?), and downstreamconsumption (i.e. exposed to promotion, does viewer take action?)
Cross-Platform Media Nirvana
21© Quantcast 2009. All Rights Reserved.Confidential
Panel intersections are tiny
TVUniverse
OnlineUniverse
The Challenge
22© Quantcast 2009. All Rights Reserved.Confidential
TVPanel
OnlinePanel
Vanishingly small intersect
TVUniverse
OnlineUniverse
What if we used the entire online universe?
A Better Way
23© Quantcast 2009. All Rights Reserved.Confidential
Intersect covers entire TV Panel!
TVPanel
TVPanel
TVPanel
Quantcast participants:all digital media consumption
and purchases for participatingmarketers and media companies
How it works
24© Quantcast 2009. All Rights Reserved.Confidential
TiVo Power||Watch panel(35,000 hh’s, opt-in) with
complete TV/DVR programand advertising consumption
Anonymously joined,single-source cross-platform media set
World’s largest single-source TV and Internet panel
• Largest sample size to date – 35,000 homes measuring both TV and Internet• Eliminates “noise” – population is large enough that a few aberrant households
can’t swing the results one way or another• Precise measurement
• Large sample size with digital measurement techniques• Measurement of specific commercial exposures (on both tv and web)• Measurement of precise timeline of web activity
• Diverse panel reflecting digital tv household population• 37% > $100k income; 57% broadband; 37% w kids; 65% married; median age 39;
median hh size 2• Completely passive observation with no self-reported behavior
• No surveys – not a single question is asked• No software to install, buttons to push, or reports to fill out. Panelists just go about
their lives.
Key Attributes
25 © Quantcast, TiVo 2009. AllRights Reserved. Confidential
Power||Watch Details
Panel Composition• Median HH size is 2• 37% have children in the home• 37% HH income 100K+• 65% married/partnership• Median respondent age is 39• 78% employed full-time• 81% ethnicity caucasian• 57% have a broadband connected DVR
• 35,120 Households• Most recruited via on-box invitation/
recruitment• On-box recruitment via TiVo Central (main
menu)• Emails to random and targeted samples
of TiVo subscribers to recruit specific -characteristic households
• Incentive for each round of recruitmenttypically is a drawing for a TiVo HD unitwith service.
• Daily viewership log collection• Low touch panel: Panelists typically receive
one survey once or twice a quarter
• Understand interplay and correlation between TV and Internet advertisingand content consumption
• For AdvertisersOverlap: Combined versus independent Reach, frequency, GRP
of Internet and TV advertisingDownstream Behavior: How does TV and/or Web ad exposure
influence Internet behavior (i.e. “are people who saw my tv admore likely to go to my web site, or search for my keywords?”)
• For Networks/Media CompaniesOverlap: TV-only versus Web-only versus both TV & Web
household viewershipDownstream Behavior: How does TV consumption
predict/influence Internet behavior, and vice-versa
What you get: New Insights on Overlap and DownstreamBehavior
27© Quantcast 2009. All Rights Reserved.Confidential
Networks: Example Reporting
28© Quantcast 2009. All Rights Reserved.Confidential
Web-only TV-onlyWeb & TV Notexposed
Exposure to ProgramPromotion
• R/F/GRP
•TV-only, Web Only, TV + Webhousehold viewership
• Among demographic segments
• Live & time-shifted TV
Downstream Behavior• TV promo driving online behavior
• Online promo driving TV behavior
• By frequency of exposureWebsite activity
TV viewing
Advertiser: Example Reporting
29© Quantcast 2009. All Rights Reserved.Confidential
Web-only TV-onlyWeb & TV
Exposure to Advertising
• R/F/GRP
• Among demographicsegments
• Live & timeshifted TV
Downstream Internet Behavior
• By frequency of exposure
• Across campaigns and specificcreative
•See previous slideWebsite activity
Notexposed
• What is my combined reach andfrequency?
• Daily, Weekly, Monthly timeperiods
• All campaigns, per campaign &per creative
• Live, Delayed, Fast-Forwarded
Example report: Overlap
30 © Quantcast, TiVo 2009. AllRights Reserved. Confidential
• Attribution of TV advertising to downstream web activity (paid search, site visits,conversions)
Example Report: Downstream Behavior
31 © Quantcast, TiVo 2009. AllRights Reserved. Confidential
Paid
Purchase
Brand AssetInteraction
Time to Convert Site Visits
1 hour 2,276
2 to 12 hour 1,515
13 to 24 hour 621
2 to 7 days 482
8 to 14 days 307
14+ days 522
Average 19.5 Hours
# of CommercialViews Reach Site Visit Site Visit Rate Site Conversion Site Conversion Rate
TV to ConversionRate
1 500,000 1,000 0.200% 150 15.00% 0.030%
2 1,200,000 8,000 0.667% 900 11.25% 0.075%
3 1,400,000 12,000 0.857% 1100 9.17% 0.079%
4 1,820,000 21,000 1.154% 2400 11.43% 0.132%
5+ 4,000,000 8,000 0.200% 1000 12.50% 0.025%
Final thoughts on Digital STB data
• Digital STB data is more than competition to panel audiencemeasurement
• Methodological deficiencies and industry inertia will hinderprogress into a trading currency
• Providing “nice to have” media effectiveness data is not abusiness model
• Separate efficiency and effectiveness metrics are now a reality• The future depends on media owners and advertisers• We are all waiting for that “tipping point”