How overnight TV ratings work

The move to overnight ratings is a technically challenging one, and as SAARF's CEO, Dr Paul Haupt explains, the delayed reporting during the interim phase is necessary to sort out any teething problems which may occur during the changeover. "In the interim phase, our focus will be on correct data, rather than early data," he says.

Timing of overnight releases

The first overnight reports were released to the computer bureaux on Wednesday, 21 June, and would have been made available to users by the end of the last week, once the necessary upgrades had been made to users' software, and once the data from Week 25 had been released.

Television audience data for weekdays would then be released two days after reporting - thus Monday's viewing would become available to users by 1pm on Wednesday. Thursday's and Friday's data would have been released on the next Monday, and the weekend's viewing released the following Tuesday.

The final release in the interim reporting stage will be on Monday, 17 July, when the data from 13 - 16 July (Thursday to Sunday) will be released.

From Tuesday, 18 July, the industry will enjoy true overnight ratings, with users being able to view the weekday reports by 1pm the following day. Fridays, Saturdays and Sundays will be packaged together for release on Mondays.

The impact of overnight ratings

The change to overnight ratings from weekly reporting is a major improvement, and will impact in a number of ways on the data.

In a presentation to stakeholders in Johannesburg last week, Chris Eyre, executive director of TAMS contractor AGB Nielsen Media Research, explained that users should expect a marginal increase in reach figures. Overnight ARs will be around 4-7% higher than weekly ratings. The change will be less marked with M-Net and DStv, since both channels are more tightly controlled by being weighted separately to correct for over-sampling.

He also explained that the TAMS sample will be slightly reduced due to the loss of some Good Non-Viewing households (GNVs) and late reporters. GNVs are households which usually provide regular data, but which are temporarily not downloading their data due to their being on holiday, or experiencing problems with their meter. With weekly data, they would remain part of the reporting panel for longer.

The sample will also be reduced by the omission of late reporters, households whose data are received later than the night immediately after the reporting day. In the past, late reporters were included in the weekly report as long as the data are received at the end of the reporting week.

With overnight ratings, though, any households which fail to download their data the night immediately after the reporting day will be excluded from the next morning's report. However, households that report Friday's data on Saturday or Sunday, and Saturday's data on Sunday, will be included in the Monday report.

In the event of a bad polling day, when communications problems cause significant levels of late reporting within the panel, the data for that day will be held back until polling improves. In such cases, late reporters will also be included.

Finally, Eyre added that in the short term, the transition will also make it difficult for planners to forecast from historical weekly data, and project to future overnight data.

Haupt believes however, that the benefits of overnight ratings outweigh any problems caused by the change in currency. "Overnight reporting means greater efficacy for advertising campaigns, which can now be monitored each day to ensure all spots were flighted and are delivering. Over- and under-delivery will be largely eliminated," he says. "Overnight ratings will also allow broadcasters to make better use of their inventory."

AGB Nielsen Media Research's overnight ratings presentation should be available at www.saarf.co.za.