Decoding the CFL’s TV numbers (and the Twitter beef over them)

On Wednesday, the CFL put out a series of Tweets trumpeting the TV viewership for its games over the Labour Day weekend.

Those seem like pretty impressive numbers, as was noted by Toronto Sun columnist Steve Simmons.

But on Thursday, Sportsnet’s Arash Madani – who, fairly or not, has a reputation as someone who enjoys poking the CFL with a stick from time-to-time – questioned some of the league’s numbers.

Those are still really strong numbers. But the discrepancy comes from the fact that the league is using one set of numbers – the total number of people who tuned in over the course of the broadcast, known as “reach” – while Madani is reporting the average number of people who watched the game, known as “average.”

Now, “average” has been the traditional metric used to measure viewership. When media reporter Chris Zelkovich says that CFL ratings are down 15 per cent in 2015 or up 13 per cent to start 2016, it’s average he’s referring to. Madani is accusing the league of misrepresenting “reach” as “average.”

The CFL, however, is hardly the only entity to use both “reach” and “average” when trumpeting their success these days. CBC did it during the Olympics and Madani’s employer, SportsNet, did it last fall during the Jays’ playoff run. Start throwing in digital numbers – people watching on tablets, phones and computers – and the numbers get even more muddled. I think it’s vitally important – and this might be the biggest takeaway for the CFL – that leagues and broadcasters be crystal clear as to which numbers they are using.

So what were the actual TV numbers for the CFL games this weekend? These according to the CFL…

• the Hamilton Tiger-Cats game against the Toronto Argonauts was watched by an average of 902,000 viewers with a reach of 2.9 million.

• the Winnipeg Blue Bombers game against the Saskatchewan Roughriders was watched by an average of 891,000 viewers with a reach of 2.3 million.

• the Calgary Stampeders game against the Edmonton Eskimos was watched by an average of 891,000 viewers with a reach of 2.3 million.

Because “reach” numbers weren’t commonly referenced until recently, it’s hard to compare those numbers year over year. But we can look at the “average” numbers, again using the archives of our good friend Mr. Zelkovich.

Game 2014 2015 % 2016 %
Argos-Ticats 796,000 559,000 -29.8 902,000 +61.4
Esks-Stamps 885,000 757,000 -14.5 781,000 +3.2
Bombers-Riders 1,144,00 1,118,000 -2.3 891,000 -20.3
Totals 2,825,000 2,434,000 -13.8 2,580,000 +6.0

With these numbers in hand, let’s go back and look at some of the league’s claims.

When the CFL tweeted that “Labour Day match ups were up 30 per cent” they are taking the average audience of the two games played on Labour Day (Argos-Ticats, Esks-Stamps), a combined total of 1,316,000 in 2015 and 1,683,000 in 2016. That’s an increase of 27.9 per cent, which the CFL rounded up to 30. But what they didn’t say is that the vast majority of that increase came via the big jump (61 per cent) in viewership of Argos-Ticats (the 2015 game was a Hamilton blowout early, killing the average.) The Esks-Stamps game was up 3.2 per cent.

Meanwhile, viewership of the Bombers-Riders game was down 20.3 per cent from last year and 22.1 per cent from 2014 while Esks-Stamps is down 11.8 from 2014. Viewership for the three games is up six per cent this year but still down 8.7 per cent from 2014.

Overall, Zelkovich, says CFL ratings are up seven per cent with a 14 per cent bump in the 18-to-49 demo. After drops of 15 and six per cent the last two seasons, that increase is exceptionally good news – especially in light of competition from the Rio Olympics and a strong Blue Jays club.

A 2.9 million viewership “average” would have been insanely high for the Ontario version of the Labour Day game and it’s important to note that there’s nothing factually incorrect in the CFL’s tweets. Given the limitations of 140 characters, providing a link to a more fulsome press release that broke down the numbers, something the league and TSN have done before, likely would have helped.

Ultimately, whether or not you believe the CFL was trying to mislead their Twitter followers depends largely how nefarious you think the league is.


Simmons, meanwhile, felt the need to apologize for the conclusions he drew based off the CFL’s initial tweets…

Bottom line: the CFL ratings are largely good news. How you interpret the rest…

Drew Edwards is the founder of 3DownNation but has since wandered off. Beard in the photo not exactly as shown.