Linear TV tops streaming again after Nielsen methodology change

An unreleased report from Nielsen will find that linear television accounted for more viewing than streaming, according to a new report by The Wall Street Journal. That’s a stark reversal from a similar report from Nielsen in January and may be the result of methodological changes.

According to The Wall Street Journal, streaming represented 41.9% of U.S. TV viewing time in February compared to 47.4% for linear TV. In January, it was 47% streaming and 42.7% linear. The numbers come from Nielsen’s monthly Gauge reports.

The difference may be the result of methodological changes. The WSJ says that February is the first month Nielsen has used a study from the Advertising Research Foundation for demographic estimates. Previously, Nielsen pulled demographic data from its own volunteer panels, which may have undercounted linear TV households.

The Media Rating Council, a self-regulatory group, requested the change. According to the WSJ, the council found that the older demographic data was inaccurate.

YouTube, with a decrease of 1.5% compared to January, Netflix, with a decrease of 1.3%, and Prime Video, with a decrease of 0.8%, were hit hardest by the change. Not every streamer saw a drop, though. Peacock saw an increase of 0.8% as it benefited from streaming coverage of the Super Bowl and the Winter Olympics.

In a statement to the WSJ, Nielsen said, “different methodologies produce different results,” and committed to using the new methodology in the future.

This is the second time in the last year that Nielsen has changed its methodology for a key metric. In September, Nielsen introduced the new “Big Data” standard to measure TV viewership, which introduced expanded out-of-house data and new metrics from streaming boxes.

The change has generally resulted in increased ratings, especially for sporting events, which has led to many record viewership claims in recent months. It is Nielsen’s policy to compare Big Data viewership with the older panel standard for year-over-year comparisons.

These methodological changes may be better at getting more accurate numbers, but they stand to muddy the waters for comparisons. For those who rely on this data for analysis now, it makes it even harder to parse through the data, which complicates comparisons.

Month-to-month Gauge comparisons should stabilize starting next month because the March report will use the same methodology. Television viewership is more often evaluated on a year-over-year basis, and those Big Data comparisons will not be fully available until next September.