The truth about TV ratings, online viewing and sci-fi shows

Contributed by
Dec 14, 2012, 4:51 PM EST

Craig is the GM and senior vice president of Syfy Digital. He'll be guest blogging about the business of sci-fi TV on Blastr and he also regularly talks to fans about Syfy and the TV industry on Twitter using @Syfy.

When sci-fi TV shows get canceled (on any network), many fans talk about how the ratings system is broken and doesn't count sci-fi viewers correctly. After all, sci-fi fans are tech savvy and don't watch live TV shows on TV ... they DVR them, they buy them on iTunes and they illegally download them from BitTorrent. If the archaic Nielsen system only took these viewers into account, many sci-fi TV shows would have massive ratings and last many more seasons.

There's some truth to what's being said, but there are also lots of misconceptions and things people overlook when the topic comes up. Since I get asked about this so often, I thought I'd try to look at ALL the issues in one place and give people a better idea of what's actually going on. My goal isn't to convince you the TV ratings system is bad or good, it's to explain how and why the system works the way it does and separate out fact from fiction.

What Ratings Actually Measure

First, we have to talk about what TV ratings are and what they're not. TV ratings specifically measure what people are watching on TV, and by that I mean content you watch on your actual television set that's sent out by TV networks in a linear stream. I don't mean what you watch on your laptop, iPhone or Roku box. We ALSO measure those, but they count separately.

This is probably the most misunderstood aspect of ratings. Usually after I say "things are counted differently" I get a knee-jerk reaction from online advocates. "See, that's the problem right there! That's just stupid. Why, why, why aren't they all counted together?!?" Good question. Here's why.

TV ratings don't just measure how many people watch a TV show, they measure how many people watch the ads in TV shows. This is hugely important to both TV networks and advertisers because billions of dollars each year are spent buying TV ads, and everybody wants to know what they're getting for their money. The Nielsens provide a way for ad sellers and ad buyers to get data on a level playing field (i.e. every network agrees to use the same system) that comes from an impartial third party, so we're all comparing apples to apples.

So the first big reason TV and online viewing count separately is because the same ads don't run on air as online (where, in some cases, NO ads run). Naturally enough, advertisers paying for TV ads don't want online viewers lumped into their ratings, since those viewers aren't seeing the ads.

So why don't the same ads run online as on air? Because advertisers don't want to buy them that way (believe me, it would make it a lot easier for us if they did). Since it's their money, they get to choose how to spend it. Later this year Nielsen is going to roll out a new rating that combines TV and online views for shows that run online with the same ads as on air, and that may entice more advertisers to buy their online and on-air ads in sync. Until they do, there is a real business need to track them separately.

Factoring DVRs and Ad Skipping Into the Equation

From here people often segue into another big anomaly about ads on television: DVRs. Ah ha, they say, I have a DVR and I skip ads. How do you factor that in? As it happens, we're well aware of that fact, as is Nielsen. (We also know people watching live TV get up during commercials to use the bathroom, raid the fridge, etc.) To take that into account, Nielsen changed their system to include DVR usage, and they're continuing to hone their reporting as DVR usage keeps growing. Right now, we get reports that break down viewing like this:

Live - Just as you'd expect, this measures how many people watched a show live, not recorded. When you pause a show via DVR for 25 seconds or more, you're kicked out of the "live" category.
Live + Same - This is how many people watched a show live or via DVR on the same day the show aired.
Live + 7 - This is how many people watched a show live or via DVR in the seven days following the show's broadcast.
C3 - This is a measure of the commercials watched both live and within three days via DVR playback. It's the metric under which much of prime-time advertising is bought and sold.

Another common misconception is that online viewing is destroying the TV model, when in reality DVRs are a far bigger issue. We've already seen shows where more than 50 percent of viewing was done via DVR recording instead of live, and DVRs aren't even in every household yet. Since we make most of our money from ads you can see why, if half of viewers are potentially skipping them, that poses significant issues.

DVRs aren't the only way TV viewing on your TV has changed. A lot of linear content is now also available on demand (both free and paid) from cable service providers, so TV companies need to figure out how that fits into the equation. Here the Nielsens come into play if the same ads that run on live TV also run in on-demand shows, but that doesn't happen very often.

We also track all on demand viewing another way, whether it runs with ads or not. This is done via another an impartial third-party system, called Rentrak, that gets data right from cable system. Because you have to request a show, it's easer to track directly than if you just dip into and out of the various gushing streams of content that TV channels provide.

How Online Views Are Counted

Then there's online viewing to contend with. For the sake of brevity I'm lumping everything like Xbox, PlayStation, BitTorrent, iTunes, Netflix, Amazon, Hulu, etc. into the "online" category. There are dozens of ways to watch content online, but broadly speaking they fall into three categories, each tracked separately:

1. TV shows that stream with ads in them: Like TV advertisers, online advertisers want to know how many people see their ads, so we track that data wherever streaming occurs.
2. TV shows that are sold online: These are shows you pay for, either by episode or by season. We get an accounting from all the digital retailers like iTunes of exactly how many shows are sold.
3. Illegal downloads: Although this contributes nothing to the business model, most TV networks have anti-piracy groups that track illegal downloading.

I won't go into mobile downloads and distribution other than to say they're handled pretty much like online and we count it. I'm also not going to get into DVD sales because most TV networks don't profit directly from DVDs. (To make a long story short, TV networks usually license shows from a TV production company, and that company sells and profits from DVDs.)

How Nielsen Ratings Work

Okay, that's how most "TV viewing" occurs. There's one other factor we have to take into account before we can discuss why all of these viewing methods are -- and aren't -- counted together. That's methodology. Nielsen TV ratings are determined via sampling, while most of the Internet methods are tracked via "actual" downloads or streams. Sampling is another of those terms that sounds like exactly what it is. Nielsen looks at a sample of households around the United States, and from that sample they make a statistical estimate about what everyone else is watching.

Sampling doesn't measure direct views because when Nielsen started estimating TV audiences back in the early days of TV, that simply wasn't possible. Today it's theoretically possible to measure direct views for cable subscribers, but practically speaking it would be so hard that it's not possible. Sort of like how living on the moon is theoretically possible, but practically speaking it's just not.

This issue of sampling vs. direct views is another hot-button topic for online advocates. Some people don't believe sampling works. And they think it works especially poorly for sci-fi shows because they feel sci-fi fans aren't adequately represented in the sampled homes. If only Nielsen adequately counted sci-fi fans, sci-fi shows would do better in the ratings.

Yes, Sampling Seems Weird

I get it. Sampling is a weird concept, it doesn't "count everybody," and sci-fi shows seem to get canceled a lot. Sci-fi fans are getting screwed! But we're not getting screwed, because Nielsen ratings don't exist in a vacuum. All of the other data we look at ... shows people watch on demand, DVD sales (we often get this data even though we don't usually share the profits), digital downloads from iTunes and Amazon, streams on the Internet, visits to show Web sites, even piracy ... give us different metrics to look at alongside TV ratings to make sure nothing really weird is going on.

Comparing all this data reveals that highly rated shows are streamed more frequently online, sell more DVDs, have higher sales on Amazon and, yes, are pirated more often. When you account for variables that impact all these metrics (e.g., some movies bomb in theaters but later sell well on DVD, younger viewers are more apt to watch things online than older viewers, etc.), we don't see the crazy variances that you'd expect if ratings weren't very accurate.

Also, you have to consider that while some sci-fi shows fail in this system, others thrive. Some start out thriving, then decline when they don't keep viewers interested. If the ratings unfairly punished sci-fi shows, you'd never see sci-fi shows rate well. What's more, you can often figure out how a show is doing by paying attention to what people are saying about it.

FlashForward, V and The Walking Dead

Take V, FlashForward and The Walking Dead, for example. All of them were highly anticipated and started out with HUGE ratings. Then the ratings for each show pretty much mirrored what fans thought about them. FlashForward was canceled, V barely hung on for a second season, and The Walking Dead grew its audience and was renewed for another season with more episodes than the first. Based on the consensus of viewer opinions I read online, that's just what I would have expected.

Overall I don't think there's any evidence to support that Nielsens are wildly inaccurate or especially harsh on sci-fi shows. And sci-fi shows are actually canceled no more frequently than other genres. The reality of TV is that most shows fail, in any genre. That's endemic to all entertainment businesses. Most movies aren't successful, most books don't become best-sellers, etc.

The Nielsens aren't perfect, but no system trying to measure what 100 million+ households are watching ever will be. Even if you could precisely measure what was on every TV at all times of the day or night, you wouldn't be able to figure out if the person watching them is paying attention, or how many people are watching that TV, or even if anyone is watching. What you need is a system that's as reasonably accurate as you can expect, and that everyone on all sides of the TV business accepts as a standard. For now, it's Nielsen ratings.

How Views Count Toward Renewal

Now that you know what we count and how we count it, we can tackle the most crucial aspect of all ... how much each kind of viewing counts. TV is incredibly expensive to make, which means you need to bring in a lot of revenue just to cover the cost of producing shows, let alone make a profit. In the business model of TV, views translate into money, which is why everyone talks about them.

Since different kinds of views bring in different amounts of revenue, you can't treat all views equally. And the difference in revenue derived from TV views compared to online views is massive. If you add up all the money you get selling ads in live and DVR viewing and stack that against all the money you bring in through every other kind of viewing method, you'd probably be lucky to get $1 in online revenue for the same number of views that would bring in $10 on TV. And that's likely an optimistic number.

If you do the math, you'll see why combining all kinds of views into a single "super rating" won't work if you want to figure out whether, financially speaking, you can keep a show or not. At that 10-to-1 ratio you'd need 10 million online views to equal 1 million views on TV. And those have to be views you make money on, not pirated content. To look at it another way, if you add the income from 1 million TV viewers and 1 million online viewers, it gives you the same income as 1.1 million TV viewers would.

What makes it more lopsided is that the pool of online viewers is smaller than the pool of TV viewers. Although it feels like everyone in the U.S. is connected to the Internet and watches TV online, only about 2/3 of houses in the U.S. have broadband Internet access. And only 50-70 percent of households that have the Internet watch video online. And of the people who watch video online, most of them are watching short videos, not full TV episodes. And we can't include international viewing because the production companies we license shows from reserve those rights to include in their international deals.

A "Super-Rating" Isn't That Super

So even if you did add online viewing to TV viewing, you're not going to double or triple your Nielsen rating, you'll increase it by some percentage, and probably not a huge one. Remember, the rating you usually read about is for the first broadcast of a TV episode, but that episode repeats, and all those repeat viewers count for the TV model. If you stack all TV viewers of all the new and repeat TV airings against all online viewers, TV will be the bigger number by far.

That's why combining every possible view into one super rating in order to look at the "real" audience for a TV show doesn't gain you much. Sure, you'll get a bigger number, but it's not much bigger, and even if it was, that number isn't going to provide much revenue, which at the end of the day is what enables you to actually make TV shows.

Is the system going to stay this way forever? Not at all. The TV and online industries are both in massive flux right now, and that will continue for a long while. Five years ago, online revenue for TV shows was counted in pennies, and now it's counted in nickels. Hopefully it will get to quarters in the next few years, and then online viewing might really start making an impact on the ability of TV networks to renew shows.

But as of early 2011 this is how the TV industry stands, and hopefully this gives you some insight into the way things work.

P.S. I'll try to stick around and answer questions in the comments below if I can.