Single-source data

From Wikipedia, the free encyclopedia
  (Redirected from Single source data)
Jump to navigation Jump to search

Single-source data (also singlesource, single source) is the measurement of TV and other media/marketing exposure, and purchase behavior, over time for the same individual or household.[1] This measurement is gauged through the collection of data components supplied by one or more parties overlapped through a single, integrated system of data collection. The means by which these data are stored is known as a single-source database.

In TV advertising measurement, single-source data are used to explore an individual’s loyalty and buying behavior in relation to advertising exposure within different windows of time – e.g. year, quarter, month, week. In this sense, single-source data is a compilation of (1) Home-scanned sales records and/or loyalty card purchases from retail or grocery stores and other commerce operations, (2) TV tune-in data from cable set top boxes or people meters (pushbutton or passive) or household tuning meters, and (3) Household demographic information.

The value of single-source data lies in the fact that it is highly disaggregate across individuals and within time. Single-source data reveals differences among households’ exposure to a brand’s ads and their purchases of those brands within advertising fluctuations.[2]


"Project Apollo" was designed to be a single source, national market research service based on Nielsen's Homescan technology for measuring consumer purchase behavior, combined with Arbitron's Portable People Meter system, measuring electronic media exposure. In January 2006, The Nielsen Company and Arbitron Inc. completed the deployment of a national pilot panel of more than 11,000 persons in 5,000 households. Seven advertisers (P&G, Unilever, Wal*Mart, Pfizer, Pepsi, Kraft), S.C. Johnson, signed on as members of the Project Apollo Steering Committee. The Committee worked with Arbitron and Nielsen to evaluate the utility of multi-media and purchase information from a common sample of consumers. Individuals within the sample were given incentives to voluntarily carry Arbitron's Portable People Meter, a small, pager-sized device that collects the person's exposure to electronic media sources: broadcast television networks, cable networks, and network radio as well as audio-based commercials broadcast on these platforms. Consumer exposure to other media such as newspapers, magazines and circulars were collected through additional survey instruments. The project was shuttered in February 2008 due to sample size, cost and insufficient client commitment.[3]

Prior to Apollo there were other attempts to provide single source measurement in the U.S., including Arbitron’s ScanAmerica, which used pushbutton peoplemeters; IRI’s use of household tuning meters in its BehaviorScan markets; Adtel; and ERIM. Outside the U.S. there have also been such efforts which have been discontinued except in England and in France where small single source panels survive. In Germany and the Netherlands, research agency Gfk is currently running single source panels under the name "Media Efficiency Panel" (MEP). In MEP online behavior and advertising contacts are measured in detail using a Nurago browser plug-in. Off-line media consumption is measured using validated media consumption questionnaires. FMCG purchases are captured using a household scanner and durable purchases using an online system asking respondents to check in and register what goods they bought, where they bought it and for what price. The German panel was launched in 2008 and is currently experimenting with audio measurement using a mobile phone to capture advertising contacts on TV. More than 70 studies have currently been done in this panel by a large variety of advertisers. The Dutch panel was launched in July 2010.

To circumvent the problem of unsustainable costs which brought down Project Apollo, cross-media analytics company All Media Count uses survey data from a longitudinal panel and behavioral modeling to generate individual panelists' daily media contact data for more than 10,000 media vehicles across eleven media types in China.

The term “single source” is often credited to Colin McDonald who while at BMRB in England in 1966 used purchase and viewing diaries rather than electronic means to conduct the first quasi-single source measurement. Although electronic means of data capture are preferred for accuracy and to minimize respondent fatigue, cost-effective methods for doing this do not yet exist for several media (magazines, newspapers, subway, transport, ambient). Also, many markets - such as China - do not have universal electronic measurement, even for TV. Services such as MRI, Roy Morgan, Simmons, TGI and others around the world which collect such information by non-electronic or hybrid means are broadly considered to be single source, as long as the data are obtained from a single panel of respondents.

Despite enthusiasm and the success of such data, single source has been plagued by high costs and small sample sizes, due to the cost of getting people to do all the work associated with home barcode scanning and pushing buttons on peoplemeters or even carrying and uploading from passive peoplemeters.

In February, 2008 the Wall Street Journal announced that TRA, Inc. had invented a new form of single-source which cures the problems hobbling all previous single-sourceefforts, high cost and small sample size. Coincidentally this occurred within two days of the announcement of Apollo terminating its project. The TRA invention subsequently received U.S. patent 7,729,940 on June 1, 2010, while in 2011 TRA announced that its client list includes 45 of the largest spending brands on television, major advertising agencies representing nearly half the business, ten major TV networks and a top cable operator. TRA was the first to realize that single-source could be scalably sustained at reasonable cost with the largest sample sizes in marketing/media research history, by matching already-existing “hard data” databases at the same-household level through a privacy shield, and for this invention received the ‘940 patent. The “hard data” databases combined by TRA include TV set top box data from multiple cable operators and TiVo (over 2,000,000 homes), with frequent shopper card data from a dozen national supermarket chains (57,000,000 homes). Co-founder Bill Harvey was first to publish TV set top box data research reports (1997) and co-founder Mark Lieberman was first to deploy a TV Video On Demand system . TRA also receives data for automobiles (115,000,000 homes), pharma (72,000,000 homes), and plans to receive data for Internet, Print, Mobile, Social, and other media. Single-source sample sizes represent the overlap between the media data and the purchase data, i.e. 370,000 for packaged goods, 620,000 for pharma, and 2,000,000 for automotive. Advertisers with their own lists (telecommunications, retail, financial, travel, electronic/appliance and others) can match these to TRA TV and other data for their own exclusive use through the TRA system.

TRA provides reports for media planning, buying, allocation, optimization, and post evaluation from a massively parallel client self-serve web-based system Media TRAnalytics®. The system assembles customized reports from household level data in seconds. TRA never receives names/addresses or any other personally identifiable information (“PII”). The match of the same households across disparate databases is accomplished by a double–blind match using a trusted third party that has a list of everyone in the U.S. and who never sees the set top box or purchasing records, only sees the PII and an abstract ID code for each household in each database. TRA is committed to the highest level of privacy protection and information protection and so is certified by semi-annual audits to be ISO 27001 certified. TRA also conducts ROI analyses through its Advanced Analytics Group, revealing the differential ROI rates associated with different networks, dayparts, programs, target groups, frequency levels, creative executions, branded entertainment applications, and other tactical options never analyzable on an ongoing basis in terms of ROI before TRA. TRA has resulted so far in one major discovery in marketing science, known as “Heavy Swing PurchasersTM” (HSPs). TRA coined this term to mean households spending heavily in a category who have bought a specific brand before, but not often. TRA data prove that this tends for most mature packaged goods brands except those with extraordinarily high shares, this is the “ROI Driving SegmentTM”: by targeting HSPs, a brand can increase its ROI for the same media cost outlay. This implies that these disloyal but known prior purchasers of a brand represent those most easily converted to repurchase by TV advertising. Although most TRA clients have kept their results to themselves, certain TRA clients have published findings indicating that use of TRA single-source has led to higher TV ROI for their brands[4][5]

See also[edit]


  1. ^ [Poltrack, D. F. (1996). Single vs. double source data for TV program selection. Journal of Advertising Research.]
  2. ^ Tellis, G. J. (2004). Effective advertising: understanding when, how, and why advertising works. SAGE.[page needed]
  3. ^ [McClellan, S. (2008, February 25). Arbitron, Nielsen End Project Apollo. New York, NY, United States.]
  4. ^ Shiffman, D. / Harvey, B. (2010, June) Increasing ROI: Applications for Single Source Analytics, New York, NY, United States.[page needed]
  5. ^ Larguinat, L. / Harvey, B. (2011, June) The New Research is Bringing Transparency Between Marketing and Finance, New York, NY, United States.[page needed]

External links[edit]