This year’s edition of TV Pulse newsletter presents eight industry papers, including works from agencies like MediaCom, Lintas Media Group, ATG and Lodestar among others.
The newsletter puts into scrutiny various aspects of TV viewing and TRPs. Whether it is ATG’s effort of quantifying relevant TG for high-end products or Lodestar’s attempt to predict TRPs of blockbuster films, TV Pulse 2005 presents thoughts that are beyond ratings.
TV Pulse Newsletter is a combined initiative of the Joint Industry Body (JIB) and TAM Media Research. The issue this year comprises MediaCom’s paper of viewing patterns of second TV households, ATG’s ‘Deriving further insights from TAM data after a psychographic definition’, a joint study by IMRB-PQR and TAM Media Research on understanding the low take-off of set-top boxes in Chennai, Lintas Media’s new tool - Intelligrip, Lodestar’s tool Reel Pointer and papers on reason of shifting loyalty and a study on programme promos.
Viewer Profile – ‘Deriving further insights from TAM data after a psychographic definition’
In a bid to quantify relevant target for the high-end products, ATG’s paper on ‘Deriving further insights from TAM data after a psychographic definition’ (based on an exercise done by MCI and TAM Media Research) looks at questions like ‘How does one size up a psychographic target?’, ‘What are the ways in which one can derive some broad media consumption learning?’ and ‘How can one then use some kind of ‘bridge’ definition of a target audience that would help one analyse dynamic media information through viewership databases like TAM?’
The paper takes the example of Category X and attempts to choose channels / programmers for this high-end category that would deliver least duplication and reach the relevant target.
A preamble to the study is that viewership of channels with international news and entertainment could comprise broadly two kinds of audiences – one that is interested in such content and the second that flirts with such channels. A hypothesis is that Category X consumption is skewed to a particular cluster, which is characterised by heavy viewing of English programming.
The first objective of this analysis, therefore, was to determine the various audience clusters of Category X and size up those who are skewed to English programming. This would substantiate the hypothesis that Category X consumption is skewed to a cluster that is characterised by heavy English programming viewers and determine the overall size of this cluster. This allows the marketing manager a realistic idea of the cluster.
The second objective was the RLD analysis of heavy viewers on English entertainment and the skewness of various channels to this definition of the target. This would give marketing managers a fix on the channels that offer ‘hard-core viewer’.
GroupM’s 3D study comes in play here. The study has the advantage of doing product linkage based psychographics. A battery of over 290 statements allows the analyst to do various combinations before finalising on internally homogenous and yet differentiating clusters. A combination of correspondence and cluster analysis offered five media consumption clusters for Category X.
The ‘Media Rejecters’ (low consumers of all media) formed 24 per cent, the ‘Junkies’ (heavy consumption of all media) were 9 per cent, 29 per cent were ‘Info Seekers’ (heavy on print and news, skewed to Hindi / regional news), and ‘Others’ (Medium TV viewers – regional, entertainment, devotional songs, radio) were 27 per cent.
The required audience, termed as ‘International’, who are Internet users, heavy viewers of English movies, English news, business and science shows were 21 per cent. Given the number of choices available today and keeping in consideration that mass channels would form a part of plan, the challenge is to choose international programmes / channels that have the least duplication with Hindi / mass channels.
To achieve this, MCI conducted a special analysis using Respondent level data in partnership with TAM Media Research. An interaction index with mass entertainment channels was computed, observing the time spent by viewers on mass entertainment channels.
The study indicated the channel with the lowest reach to also have the lowest interaction index, which proves to be a better vehicle to deliver exclusive viewers.
The bottom-line of this paper was that combining a psychographic definition along with a special analysis using TV data helps in making a smarter channel selection for the high-end brand.
Reel Pointer – A tool to predict ratings of blockbuster movies on TV
This paper by Lodestar Media generated from industry examples like ‘Devdas’ delivering only 3.7 TRPs, despite the acquisition of Rs 12 crore and ‘Humraaz’ beating ‘Kabhi Khushi Kabhie Gham’ by getting a TRP of 12.
The challenge was to bridge the information gap in the current TV programming scenario – blockbusters are amongst the biggest properties, but lack data to evaluate and hence, price them.
The methodology used was of regressing modelling, where TRPs of past blockbuster movies were modelled against factors that would help predict future performance. Some of the factors involved here are Box Office collections, how recent the movies were, repeats, promos, channel of telecast, daypart, day and opposite viewing.
The paper began with the ‘Modelling Process’ that included retention of strong predictive relationship – only those factors that gave a high coefficient of correlation and a low standard error were retained in the model.
The second aspect of the paper was ‘Multicollinearity’. A low multi-collinearity (variables having low correlation with each other) was seen as a necessary condition to for being considered for the model. This was tested by calculating the coefficient of correlation between the independent factors.
The next consideration is ‘Curvature effects’. The relationship between TRPs and a factor like Box Office collections may not be linear beyond a point. The plot of this is curved and the equation was suitably adjusted.
‘Stepwise Regression’ is the final point here, which implies addition of independent factors step wise. This process was done one at a time and was carried out till the highest ‘goodness’ of fit was attained.
Some of the findings of the paper were that a good prediction model had a high ‘goodness of fit’. Giving a few examples where the Reel Pointer was of use, the paper quoted movies like ‘Humraaz’, ‘Ishq Vishq’ and ‘Mujhse Dosti Karogi’ among others, where the agency predicted numbers that were very close to what ratings showed – in effect helping in the right movie choice.