LinkedIn emplea cookies para mejorar la funcionalidad y el rendimiento de nuestro sitio web, así como para ofrecer publicidad relevante. Si continúas navegando por ese sitio web, aceptas el uso de cookies. Consulta nuestras Condiciones de uso y nuestra Política de privacidad para más información.
LinkedIn emplea cookies para mejorar la funcionalidad y el rendimiento de nuestro sitio web, así como para ofrecer publicidad relevante. Si continúas navegando por ese sitio web, aceptas el uso de cookies. Consulta nuestra Política de privacidad y nuestras Condiciones de uso para más información.
Se ha denunciado esta presentación.
¡Ha desbloqueado descargas ilimitadas en SlideShare!
IntroductionTAM data fulfils the need to understand Indiantelevision audience behavior.Over the past decade, the data and its dimensionsitself have vastly increased;more sample size, more geographies to analyze (28reported markets in 2008 to42 reported markets in 2013), platforms like Digitalemerging, explosion innumber of active channels 427 in 2009 to 509 in2013 etc
1. Realize that this data comes from asurvey based on a statistical sampleLet’s say you were analyzing a programme’s performance forMumbai for theC&S 4+ target group on June 11, 2012 that ran between 20:00– 20:30 hours onChannel ABC. You ran the analysis in Media XPress and foundthat the reach ofthe programme was 5% and the associated sample size forthe analysis targetgroup was 2000 individuals. These 2000 individuals comefrom a set of homeschosen to be representative of Mumbai’s C&S populationaged four years and
3. Business significance is as importantas statistical significance.Every standard statistics textbook will talk about the difference inestimatesbeing ‘statistically significant’ and “practically significant”.Consider the graph below. At first glance it seems that Channel B is wayaheadof Channel C and definitely ahead of Channel A. But a little attention totheGuidelines on Standard Usage of TAM Dataestimates themselves shows that the difference in estimates is verysmall: 0.01,0.03, and 0.02 between Channels B-A, B-C and A-C.Assuming these differences were even statistically valid, from acommon-senseperspective these differences are extremely unlikely tobe large enough to impact business.
4. Trend the dataThe biggest advantage of the TAM panel is that one is measuringmore or lessthe same people (20%-25% average yearly churn) over time. Due tothis, changes inviewing behavior are reported more precisely than if we were tosimply conducttwo separate surveys at different points in time. But the power of thepanel is notabout just calculating a change over some two points in time. We cansee howgradually or suddenly this change is happening by also changewas a ‘one-off’ or part of a systematic change.
5. Use the right basket foranalysisTAM does not classify channels into specificgenres. But it is expected that theuser includes all channels of a genre whenanalyzing data for that genre. This isbased on her domain knowledge. Similarly thecase of programmes though thisis admittedly more complicated. But the fullattempt should be to give an honestrepresentation of the environment in which thechannel or programme existswithout prejudice to any other player.
Use and/or Interpret Time Spentper Viewer (TSV)with discretionCount all the minutes that a channel is viewed in ageography. Divide it by the population estimate(viewers and non-viewers of that channel) of thatgeography. The result is Time Spent Per Universe(TSU). This is what TAM reports.Now divide the total man-minutes by only thenumber of viewers of the channel. This is TimeSpent Per Viewer (TSV). It stands to reason that TSV>= TSU.As a hypothetical example, take the case of a marketwhose population is 100 individuals. Let’s assumethat only two channels (A and B) are available to beviewed in this small market. Their viewing behaviorfor a certain time period is given on the left
End-NoteA rise or drop in viewership for a programme need notmean that TV viewers havesuddenly started liking or disliking content. Analyzing TVviewing behavior is morecomplex than it seems. There is a range of factors at play.Environmental factors likepower cuts, competitive programmes, effect ofpromotions, special events likecricket etc. all contribute to behavioral and therefore datachanges. These must bestudied carefully before jumping to conclusions.