Developing TV projections is a tricky process filled with pitfalls for the unsuspecting strategist or planner, says Britta Reid.
There are regular updates to the Television Audience Measurement panel to ensure that it corresponds with the All Media and Product Survey (AMPS) universe, both in terms of universe size as well as the shifting demographic profile of the country. There are also adjustments to the panel weighting. Such changes impact on the results delivered by the panel.
Major schedule disruptions such as the FIFA World Cup significantly affect the viewing habits of certain audience segments. Less dramatic scheduling changes such as the flighting of a new series impact on viewership. More recently load-shedding has come into a play as a factor eroding audiences. As any planner will tell you, you need to have your wits about you when you set about projecting the audience figures for a TV plan. The past is an imperfect guide to the future, and you need to be aware of how to adjust for the multiple variables that affect audience ratings.
I was taking a rather broader view of television viewing habits, looking at the average amount of time the adult viewer spent watching TV on a daily basis. Using Telmar’s Transmit Quarter Hour programme, I had run monthly totals of time viewed from January through to December 2014. Looking at the data, something did not feel quite right. I had memories of average adult daily viewership being in excess of three hours.
To check this, I went back to 2012 and 2013 data. Reassuringly, this data corresponded with my recollections. However, the figures for 2014 were more in the region of two hours 40 minutes. It looked as if adult viewership was about 15% lighter than it had been! For a moment, I was tempted to believe that I was looking at the conclusive first signs of the much predicted death of television.
Then rationality kicked in. That rigorous panel upgrade that the National Association of Broadcasters (NAB) had overseen with Nielsen, in the wake of the TAMS Audit, was realistically the explanation for this seeming loss. In discussion, Brenda Wortley, the director of strategy and research at DStv Media Sales, reminded me “the changes that have been brought into the panel have been done by Nielsen with a great deal of transparency, particularly with the technical oversight committee as well as under the watchful eye of a respected international TAMS auditor”.
The panel was significantly expanded, and its profile was adjusted to reflect that of the evolving television-viewing universe. Corrections were made to obvious anomalies. The 50+ age group, which had been significantly over represented, was downwardly adjusted. The proportion of younger viewers was increased. This replaced heavier TV viewers with lighter ones. To be more specific, 15-24 year olds spend 21% less time watching television than the 50+ age group. If one compares these youngsters against the 65+ group, they spend 25% less time in front of the box.
The old panel also had an unusually high proportion of panel members who had participated on it for more than 10 years. (The international preference is for tenure of five and certainly not more than eight years.) Such long service would naturally translate into a higher proportion of older viewers. The rural and lower LSM groups were also significantly under-represented on the panel. Perhaps unexpectedly, rural viewers tend to spend about 18% less time watching TV than their metropolitan counterparts.
Massive technological improvements were also implemented in the TAM panel. Historically the Eurometer was used to measure TV viewing in the TAM panel. This somewhat antiquated machine did not enjoy the safety of battery backup. If a power outage occurred in a Eurometer household, the household would neither be able to view nor would it be available for ‘polling’, or data collection. The household would not be part of the released data sample (i.e. no power – no data collection).
Fortunately the Eurometer was phased out in 2011/2012 and replaced by the more sophisticated Unitam3 metering technology, which is supported by a battery. When power-outages occur in these households, viewing will be prohibited, but the battery will allow the meter to be available for polling. In other words, more non-viewing households now are counted in the sample. This, in turn, has reduced the time spent viewing. The Unitam3 meter is also able to report more sensitively on the ‘other’ viewing activities on TV, such as the viewing of DVD’s, playing games and PVR trick functions. This improvement in data collection, of course, lowers actual viewing time as well.
Not only has the number of minutes spent viewing changed but the weekly cumulative counts for Total TV has also been downwardly adjusted. This can not, simplistically, be interpreted as a ‘loss’. The fact is that the revamped panel is more accurately reflecting actual viewing levels in households. As Wortley succinctly pointed out to me, these figures do actually demonstrate “the broadcasters’ commitment to providing the industry with accurate measurement as opposed to maintaining high ratings”. This is, surely, reassuring given the deep suspicion with which the media agencies have tended to view the broadcasters’ research endeavours. I, for one, definitely feel comfortable with her reassurance that “the same rigorous oversight and refinements to the panel will continue under the Broadcast Research Council structures for the balance of the TAMS contract”.
We cannot expect upgraded research methodologies and increasingly sophisticated technologies to provide us with the same old results. With the new RACS tender awarded and the new Establishment Survey tender about to start, the industry will need to adjust to new benchmarks.