Sports TV Ratings FAQ

Here’s some pointers for reading my regular ratings posts.

How do I read your charts? What do all the numbers mean?
Each row corresponds to one event. The color of the row corresponds to the sport the event involves. Dark blue corresponds to the NFL, light blue to college football, yellow to the Olympics, orange to college basketball, red to the NBA, purple to Major League Baseball, dark red to NASCAR, green to golf, and sea green to the World Cup. All other sports are white.

Depending on how much information I have, there may be a ranking on the far left side of the chart. Otherwise the first column identifies the event. Generally, championships are in bold and other playoff games are in italics. Next comes the actual ratings numbers:

Vwr (mil) is the total number of viewers measured by Nielsen, in millions. The charts are usually sorted by this number, because its specificity minimizes the chance of ties, but for events on broadcast in the daytime it is not always consistently reported, or reported with as much specificity as desirable.

HH is the household rating, which is the percentage of households tuned into the event out of the Nielsen universe of 115.8 million (for the 2013-14 season), rounded to the tenths place. Household rating is another way to measure overall popularity, and is almost always available in some form. Usually viewership and household rating track fairly close to one another, except that events on Spanish-language television will tend to have a much higher ratio of viewership to household rating than English-language events.

18-49 is the percentage of viewers age 18-49 tuned into the event out of the Nielsen universe of 126.96 million (for the 2013-14 season), rounded to the tenths place. It’s well known that adults 18-49 are the “money demo” sought by advertisers, but it’s not always appreciated just to what extent demos rule television: the web site can predict the fortunes of scripted shows on broadcast based entirely on the 18-49 rating, without any reference to total viewers or household rating. This does not seem to have been widely appreciated among sports circles, however; the only sources I use that consistently report 18-49 rating are the general TV sources, not the sports-focused ones. Admittedly the question becomes more complicated for cable networks, and there are other important demos, but sports are valuable in large measure for their ability to deliver 18-49 viewers.

The time the event started is listed in Eastern time. The network the event aired on is listed in bold if cable, in normal type if a broadcast network.

What are your sources and when do you post data?
My sources can be grouped into a few categories. To understand them, let’s walk through a sample week, which we’ll call November 1-7. The Nielsen week is Monday-Sunday, so November 1 is a Monday.

Generally, ratings for Monday-Thursday will be available the day after. Ratings for Friday are available on Monday, and ratings for the weekend are available on Tuesday (with occasional delays for holidays). Two sources, both of them general TV sites, will report ratings from these daily releases: The Futon Critic and TVbytheNumbers. TVBTN will report total viewers and 18-49 ratings from broadcast primetime shows and the 100 most-watched cable shows among 18-49 (the numbers for broadcast are generally rounded to the ten-thousands place); the Futon Critic will report those same numbers for broadcast primetime shows and a somewhat arbitrary selection of cable primetime shows. (The Futon Critic never rounds total viewers and not only reports 18-49 ratings to the hundredths place but the raw number of 18-49 viewers as well, but because TVBTN doesn’t do the same and no one else generally reports any 18-49 numbers at all it’s of limited usefulness.) TVBTN doesn’t do broadcast shows on Saturday, and their Saturday cable posts tend to be a bit delayed. I generally put up the sports-related numbers from these sources once I have both of them on Twitter (though I usually don’t wait for TVBTN’s Saturday cable ratings right away); usually the Futon Critic’s post comes out later, often during East Coast primetime.

So, the numbers for November 1 get tweeted out on November 2, and this continues throughout the week until numbers for November 6 and 7 are tweeted out on the 9th. At this point, a number of summary reports come out, most of which we don’t need to worry about. By Friday, however, two more sports-oriented sources weigh in. The “Ratings Buzz” posts on Awful Announcing list viewership and household rating for the top 10 programs on each of the Nielsen-rated national sports-oriented networks; previously, their author posted those numbers for every single show to air on those networks on the Son of the Bronx blog before Nielsen’s lawyers started cracking down and he stopped posting there. In addition, SportsBusiness Daily‘s Friday issue (in this case, on November 12) always lists viewership and household rating data (to the extent it’s available to them) for almost all sports events on broadcast and the top ten events on cable for the week; however, it’s trapped behind a very expensive paywall ($120 a month!) for the first month. For most popular events, Sports Media Watch will re-report viewership and household data over the weekend, and I usually end up leaning on that when making my Sports Ratings Highlights charts, while incorporating the actual SBD data into the Top 10 Most-Viewed Sports Events charts that come out a month later. SBD also tends to not report a week’s numbers at all (especially for broadcast, which is where I most lean on it) if it’s even slightly delayed by holidays (or if Friday is itself a holiday), and I’ll lean more heavily on SMW in those cases as well.

There is one other thing to take into account: Nielsen considers the airing of cable NFL games on local stations in the participating markets as effectively a syndicated show, and counts the airing of the games on cable when calculating syndication numbers. As a result, during NFL season Monday and Thursday Night Football will show up on the top 25 most-watched syndicated shows list with higher numbers than originally reported, incorporating the numbers from the local market stations. These numbers come out a week later than the other recap posts, so the syndication numbers for November 1-7 will come out on the 16th. This is generally the soonest the ratings post, with the Sports Ratings Highlights, comes out. If I’ve incorporated syndication numbers the network will be listed as “ESPN+Locals” or “NFLN+Locals”.

Finally, the SportsBusiness Daily roundup comes out of the paywall on December 11, and I incorporate its numbers into the Top 10 Most-Viewed Sports Events that goes up, along with the November 29-December 5 highlights, on December 14 at the soonest.

Sometimes I’ll see a “share” reported for a given event, or I’ll see something like “4.4/6” listed. What does this mean?
The share is the percentage of televisions tuned in to the event in question at a given time. It is always reported as a whole number; in the example above, the 4.4 is a rating (usually the household rating), while the 6 is the share. 4.4% of all households in the Nielsen universe were watching the event at that time, which also represented 6% of all the televisions that were turned on at that time.

Sometimes I see “overnight” or “fast national” numbers reported for certain events. What do these mean?
Hey, I saw a number reported for an event that’s completely different from what you have! What gives?
Hoo boy, this is going to be a long one. The short answer is that you can completely ignore these numbers, as they are pretty much completely useless, although the fast nationals are less useless, and generally you can get some idea of the general area the final ratings will end up in from the overnights or fast nationals. The long answer requires a more detailed explanation of how Nielsen works.

Nielsen collects data from three sources: set meters, which are what are installed in the 56 so-called “metered markets” to track what people are watching on their local stations there; People Meters, which track who specifically is watching what program, and are where demographic data comes from; and paper diaries, which are used to track what people are watching on local stations in the other 154 markets and are only used during the “sweeps months” of November, February, May, and July, and never for any publicly-available national ratings data. In addition, Nielsen has begun collecting data from cable set-top boxes on a limited basis, but mostly to improve the accuracy of data in the 154 non-metered markets.

Strictly speaking, overnight ratings consist only of data collected from the metered markets, and consist only of a household rating and share, while fast national ratings incorporate People Meter data from all over the country to calculate more specific viewership and demographic data. Most people in television don’t pay attention to overnight ratings anymore, except for local ratings for individual markets, but confusingly, some sources, like TVbytheNumbers, will refer to fast national ratings as “overnight” ratings. (A “metered market” rating is always an overnight rating.) SportsBusiness Daily reports overnight ratings from weekend sports events on broadcast every Monday, but that just shows how ignorant even the most knowledgable people reporting about sports ratings are about how Nielsen ratings actually work. The only thing I use those overnight ratings for is as a starting point for what to look for when SBD then doesn’t post the final ratings at the end of the week.

(Overnight ratings are theoretically useful in comparison to the final numbers to compare an event’s popularity in urban vs. rural areas… but because Nielsen market boundaries are county-based, there can be a considerable amount of variety within a market, and “urban” markets can be a lot more “rural” than you think. Plus, I don’t think the metered markets are strictly the 56 largest markets either. I believe the smallest metered market may be Fort Myers-Naples, Florida, which is 62nd overall.)

Fast national data is usually available to networks and other sources by 8 AM ET the following day, regardless of weekends and holidays, and is usually pretty close to the final ratings; some sources will even report a “final” rating “according to Nielsen fast nationals”. But fast national ratings still don’t incorporate a number of factors that can result in adjustments in the final numbers. For primetime sports events and other live events in particular, the fast nationals are completely useless because they count whatever would be in the normal time slot on the West Coast, so a sports event that airs from 8-10 PM ET will have fast nationals including whatever aired on the West Coast at 8-10 PM PT, even though it aired at 5-7 PM PT.

Overnight numbers are irrelevant and fast national numbers are almost never reported for cable networks.

My question isn’t on here, or one of the answers just confused me even more!
Leave a comment here, on one of my ratings posts, or at mwmailsea at yahoo dot com, and if I think you have a good point I’ll edit this FAQ. Keep in mind, though, I’m not sure how well I understand all of this myself!