First explorations of television audience survey datas (from Medyatava-AGB Nielsen)

Last update : may 18th 2011

I used excel’s macros programming (cf. this article for the technique) to gather into one database several informations about the daily audience of television in Turkey since 2004 containing : target group, day of diffusion, name of the program, channel of broadcasting, score of audience, rating, and market share. The period studied in that database runs from June 11th 2004 to May 1st 2011, and represents approximately, so 2500 days and 500.000 different programs referenced.

NA : in order to share those results with turkish colleagues more easily I tried to write this article in English. I apologize in advance for the grammar faults I’ll certainly do. You can post comments if some parts seems not understandable.

I considered that part of my thesis as a complement to other datas I presented into another articles (timelinesTV production informations). On the frame of a scientific research the global intention of this new quantitive work is to prepare, by gathering various quantitive facts or infos about my subject (tv turkish series production from 1970 to now) some sufficient factual backgrounds to lead further qualitative and theorical analysis.


Before any consideration I wanted to give some infos giving an idea of the turkish population concerned by the television. In Turkey, for a population of 72 millions inhabitants and 18,3 millions households in 2009, 17,9 millions have a tv (so 96%). It means around 71 millions turks were in 2009 able to watch television at home. (source : EU, MAVISE )

The audiences datas presented below were first produced by AGB Nielsen, then published day by day on Medyatava since June 2004.

According to their website « Nielsen Audience Measurement Turkey has been conducting Television Audience Measurement (TAM) services in Turkey since 1989. Today, the panel represents the urban Turkish population with 2500 households« .

Population: 72,561,312 total / 60,264,546 urban (2009 population)
Panel Size: 2,500 households
First data production: 1989


Peoplemeters installed: 3,500 to 4,000
Universe: 51,657,783 (representing age 5+ urban population)
Data monitored: Terrestrial, cable and digital satellite
Data retrieval via: Phone lines


(Source : AGB Nielsen)

The main critic we could do about the audience survey by AGB Nielsen is the lack of precisions concerning the method they use to obtain their numbers. It’s also interesting to notice that AGB Nielsen is nowadays living the end of a crisis lasting since 2007 between them and the turkish government (cf « revue de presse » and this article), their contract coming also to an end. Anyway the interest of doing this study is the monopolistic position of AGB Nielsen in the audience measurement in Turkey. Even if the figures given by AGB are suspicious, the fact stays that every private channel has to refer to their numbers to decide of a broadcasting strategy.

Another precisions and definitions concerning the reading of the graphs :

  • Audience rating : from 0% to 100%, 0% meaning nobody watched the program, 100% the whole sample population watched it.
  • Audience score : from 1st to 100th place. A good score could be a high or low rate depending on the other audience ratings ; it’s relative.
  • Number of programs : quantity of programs broadcasted.
  • AB group / TÜM group : groups of population defined by AGB Nielsen. I don’t have yet a lot of information about the groups composition. So as a first exploration, this analysis will just focus on one : the AB group, seeming to be the « high and middle class », the one interesting the most for advertisement sellers.


I’ll present therefore a serie of various graphics intending to measure the importance of TV series audience amongst other kind of programs into the grid of the channels. At this step of the analysis the audience analysis can be approached by different ways : the program typologies, the channels grid’s structuration, and the periodicities of audience.

The audience data is a key fact for any channel studied : the audience weights the price of the advirtisement space in the broadcasting grid, and so the capacity of revenue of the channel. The higher audience gets, the more expensive you can sell advertisement. It seems quite important to analyze audiences and its impact on creation considering the behavior of other channels as in Turkey as abroad : for instance it’s well-known in America the channels are monitoring very carefully the audience’s curves and make changes to the contents, to the scenario, to get better audience rates. The situation in Turkey seems to be likewise and any telespectator in Turkey could see what importance and presence got the sponsors into the production of a tv serie.


The first graph presented above compares the audience score rates (from the 1st viewed to the 100th) of different categories of programs :

Cartoons, Concerts, Daily Magazines, Documentary, Film Foreign, Film Turkish, Health/Service Education, Humour, Magazines, Meeting/Discussion, Music/Entertainment, News report, Newscast, Politics, Reality Show, Religious, Serials Foreign, Serials Turkish, Sport News, Sporting Events, Sporting Magazine, Studio Programs, Talk Show, Theatre/Cabaret/Musicals, TV Games, Weekly Magazines, Women.

Audience scores analysis by programs typologies (click for a full view)

Reading the graph : We can see on this graph a very distinguished preferences from the audience (10 first rates ) for turkish series and newscasts, the third prefered kind of program being the TV games. The particular observation with this chart is the relative constance of those curves for every typology from 10th to 100th rating score. The interpretation we could do at this point (as a hypothesis) would be the following :

Above another graph in the same spirit but slightly different. In that graph the scores (x axis) are replaced by constant ratings, so the interpretation deals with another fact : the « left-behinds » turkish series programs.

Audience rating analysis by programs typologies (click for a full view)

Reading the graph : In my interpretation of this graph, the « left-behinds » turkish series are the programs in orange at the left side of the graph, composing the big orange curve. Those programs get a very low audience (less than 10% on the x axis) but are massively produced ! That fact seems quite paradoxical but could be seen as an effect of the inflation of the tv series production – the so-called « Yesilcam » era for tv series. It could be explained by the will of producing more and more series, hoping one of them gets good rates. Series being the most successful way to have audience, more producers invent series concepts. Logically this « liberal » way to produce generates more offer than demand, and we can imagine half of series production has just succeeded in starting but not lasting. This could be the dark side of series production : the inflation of ideas and careers left behind because of their lack of capacity to get audiences.


The previous graphs were more dealing about the repartition of the whole programming between various typologies, without precising on which channels they were broadcasted.

The next graph shows how every channel compose its own grid of programs.

Channel programmation composition by typologies (%) (click for a full view)

The next graph shows the numbers of TV programs rated from 1st to 10th, classified by channels.

Scoring audiences by channel (click for a full view)

There is another way, perhaps more explicit, to show those figures using the percentages.

Audience scoring audiences by channels (%) (click for a full view)

Reading the graph : On the 18 channels presents in the AGB Nielsen datas, only 4 are getting more than 80% of the 10th first rates. So, if we trust AGB sources, the spectators are mostly watching ATV (in blue), KanalD (orange), Show TV (clear blue) and Star TV (pink). It’s also good to notice in the margin that TRT is treated differently by AGB Nielsen since 200…

The interesting fact showed by these numbers is the constance of this repartition in the 10th first rates.


Another important factors of the composition of broadcasting are the cycles of programs typologies choices within a day, within a week, or within a year. The time factor is a complex data to compute on a day-by-day database.

The two next graphics shows clear cycles on year-scale illustrations :

Audience ratings for 1st score by major typologies (click to a full view)
Average audience rating by day of broadcasting for the 1st rating score (click to a full view)

Reading the graphs : every summer (between june and august) since 2005 registers a clear decrease of audience. We could also note that turkish series (in blue) gets most of the highest rates the whole year compared to other programs, as we can see by the « clouds » of high rates.

The serie of the next three graphics focuses on turkish series programmation and their periodicity by month  :

Turkish series analysis by month - average / quantities (click to get a full view)

Turkish series analysis by month – average / percentages (click to get a full view)

Turkish series analysis by month, year of programmation and audience scores - quantities (click to get a full view)

Reading the graphs : As we’ve seen before, the summer period could be considered as an empty one for series. But here the curves for all audiences scores and 10th first audiences scores are very different : as  the total of programs broadcasted is merely stable the whole year (just february, march and june registering a light decrease), the curve of best scores is much more flexible and increase from january to march. We could input that to a big difference of strategy of diffusion between series in general and series gaining the first 10th scores. Perhaps after a couple of month of tests, the less competing series are ended at the benefit of the other, then more broadcasted

Furthermore the last graph shows the increase of serie’s diffusion from 2005 to 2010 in the absolute, with two years having real particularities : 2008 (the month of August for first audience score shows a peak) and 2010 (the absolute quantities of series broadcasted for all scores and for the first score).

The next graph focuses on a week-scale analysis, showing the percentages of programs broadcasted by their typology every day.

Typologies by weekdays (%) (click for a full view)

Reading the graph : Once more we can see the importance of turkish series in the grid compared to other kinds. Series are mostly working-days programs, the saturday and sunday gaining good audiences with other kinds of programs like TV games. For instance the mondays, the first 10th best audiences are usually generated by turkish series (50%), then by newscasts (25%), and the rest stays under 10% or less.


To be understood clearly, I wanted to precise I tried to keep in mind during this study that the figures given by AGB Nielsen should be taken with many precautions : first because of the different complaints opposing AGB and the government on the accusation of manipulating the datas, and secondly because of the absence of another major channels like TRT. So it’s clear AGB figures doesn’t have to be taken as a reflect of the reality of the audience, but as a tool inside the TV production chain.

Taking in consideration those limits the interest of this database lies anyway in the possibility to show how the different program typologies, the channels, the schedule of broadcasting and the ratings are interrelated, constituting a field in a Bourdieusian way to speak. To complete those diagrams by a field analysis, a factorial correspondance analysis would be necessary and should be the main topic of a next article.

As a first conclusion it’s obvious to say turkish series have a strategical importance in the channel’s capacity to get audience, and mechanically to get revenues from advertisement. More than strategical it’s nowadays the first kind of program getting best audiences.  This capacity reveals interesting characteristics of the TV production for private medias in Turkey : even if in Turkey the topics of series are very dependent with local cultures and so differentiate production from US, the logic of « market » and the inflation effects related are similar.

A second conclusion to this article would be the help those analysis gave to give more precision to my phd thesis. It’s one thing to know that generally tv series are important in the television panorama, it’s another to be able to give precise datas about it. Furthermore, crossing those reflexions with previous ones, those analysis allow me to choose with more arguments a reduced panel of channels (ATV, SHOW, KANAL D and STAR TV) being the channels whiches are in the same time major producers of TV shows, and the most dependent on audiences rates in their capacity to attract advertisement funding.

Laisser un commentaire

Votre adresse de messagerie ne sera pas publiée. Les champs obligatoires sont indiqués avec *