A trade group says that newspapers like the New York Times have seen large increases in circulation, but that’s partly because they are allowed to count their readers multiple times. The industry needs to do better.

There has been much hue and cry about the New York Times passing USA Today in circulation to become the second-largest newspaper in the United States, thanks in part to a boost from the NYT’s digital susbcription plan, which reportedly boosted circulation to almost 2 million daily readers. These numbers are notoriously dodgy, however — and if anything, they have gotten worse instead of better with the arrival of online measurement and new digital devices.

The real bottom line is that until newspapers start coming clean about their readership — both to themselves and to their advertisers — they are going to continue to miss the forest for the trees.

The latest circulation gains for the NYT and others came courtesy of the Alliance for Audited Media (formerly known as the Audit Bureau of Circulations), an industry group composed of advertising agencies and publishers. The group noted that the numbers are not really comparable to the previous year’s results for a number of reasons, including the fact that some newspapers have launched new subscription formats, stopped printing every day and so on.

Counting readers multiple times

As Edmund Lee at Bloomberg points out, the AAM survey — which is somewhat ironically locked behind a paywall — also allows publishers to count their readers multiple times, according to rules adopted recently by the group. In other words, newspapers can count someone who reads the newspaper in print, on the web and on their Kindle as three separate readers. But doesn’t this inflate their readership numbers unreasonably? It sure does. The bottom line is that no one really knows what the “real” readership numbers are for newspapers.

Some argue this has always been the case with newspapers, which is true: publishers have routinely engaged in all kinds of shady tricks to boost their circulation — including special discounts for bulk purchases by hotels and airlines and other giveaways, and even dumping large quantities into ravines or pulping them after printing. On top of that, many papers have inflated their readership numbers for years by claiming that each copy gets read by as many as five people, an estimate that borders on the ridiculous.

Newspapers need to come clean

This defence boils down to: “Newspapers have always done this, and no one believes these numbers anyway, so what difference does it make?” A pretty weak defense, you might argue — and you would be right.

The other line of defence is that online measurement is also chaotic and confusing at best, and that since websites can’t even agree on whose numbers are correct, why should newspapers be any different? It’s true that measurement of online traffic is murky, with providers like comScore often giving wildly inaccurate estimates when compared with a site’s internal numbers. But this is a little like saying newspapers don’t have to tell the truth because no one else does either.

If newspapers are competing with online publishers and digital-native content companies for both readers and advertising, which they clearly are, then they have to be better than their competition — being just as inaccurate is hardly helping their cause. And they should be spending a lot more time on trying to measure real engagement (repeat visits, time spent, etc.) than on simplistic and flawed vanity metrics like raw circulation numbers. That is a mug’s game.

Post and thumbnail photos courtesy of Shutterstock / Donskarpo

You’re subscribed! If you like, you can update your settings

  1. Martin Langeveld Wednesday, May 1, 2013

    Matt, one thing the analysts are ignoring is that in the AAM report, there is a new metric called “total customer accounts” which eliminates the duplication (but leaves out single copy sales, and those are not separately broken out). That metric only started being included in the Sept. 30 report, so we don’t have YOY comparisons yet. But looking at just the Boston Globe, which elsewhere is reported to have gained “total” circulation, there is a substantial drop in the total number of accounts from just Sept. to March. (About 7 percent on the daily, 1% for Sunday.

    1. That’s interesting — thanks, Martin.

  2. David Thomas Wednesday, May 1, 2013

    “But doesn’t this inflate their readership numbers unreasonably? It sure does.”
    > If they’re selling different ad space on the print, web, and tablet then no, it doesn’t seem unreasonable.

    “On top of that, many papers have inflated their readership numbers for years by claiming that each copy gets read by as many as five people, an estimate that borders on the ridiculous.”
    > This is quantifiable, and while the number five just “seems” out of bounds, what is the baseline? One person? Two? When an item is read in print, then circulated among friends & family via the web, doesn’t that contribute to the overall total?

    “But this is a little like saying newspapers don’t have to tell the truth because no one else does either. ”

    > Yeah, but they’re fighting for the same ad dollar (don’t tell me about separate budgets). What’s more, it is quite a common practice among event organizers, sports stadiums, and theaters.

    You want newspaper organizations to be the white knight? Isn’t that like telling them they need to be falling on their own swords? Nielsen and others content with this issue all the time — if the TV is on at the bar and there is 40 people at the bar, the TV station says its 40 people watching.

    I agree that there should be a commitment to the reality of the measurable and unbiased estimates, but the older institutions are up against some amazing upshot blowhards accustomed to keeping their own numbers and in the habit of throwing out BS and being rewarded with big VC bucks. Shouldn’t the outrage be aimed at the likes of FB and others to lead the way?

    1. If they were selling completely different ads for each platform, then maybe — but most are not. And the pass-along figures are a joke — no one really believes them, including newspaper editors themselves. You might as well argue that because my wife leaves her computer on with the newspaper website up, they should be allowed to count all five of us in the household as online readers.

    2. David – actually most TV stations don’t do that. In major markets, their are LPMs that users have to actually for measurements to be attained – the samples aren’t big enough and their are other factors , Smaller markets are still diary based and only get ratings during sweeps months. Again these measurements are not perfect, but believe me TV buyers are not taking TV stations word on who is watching.
      Also readership and circulation are 2 completely separate measurements

  3. Gunther Furbush Wednesday, May 1, 2013

    So it’s only newspapers that give advertisers the most positive audience numbers possible? I don’t think so.
    If I watch my local TV news at noon, 6 and 10, they count me as three viewers. Radio ratings are a joke. And don’t get me started on the unreliability of web traffic stats.
    This issue goes far beyond print. Paid Content should know that.

    1. Your analogy is false. TV counts you as three viewers because (1) those are three different programs, not one; and, (2) TV audience is measured by daypart, and each of your three examples falls within a different daypart.

      1. Dayparts allow TV stations to count one viewer multiple times, right? The analogy is valid.

        1. No, it’s not valid. THEY’RE SEPARATE PROGRAMS. And, for any program that spans multiple dayparts, those numbers are broken down by daypart.

          You want to do that with a newspaper? Publish separate editions, or separate titles.

          1. It is valid. If a newspaper reader is reading in print, online and on a Kindle, do you think they’re reading the same article over and over again? If TV can call the same viewer in three time slots three different people, then the newspaper reader using a variety of devices should be counted separately.

        2. Guther – What you are not considering is differences in the way media is measured.
          and the way these questions are asked in the survey.
          When measuring newspaper readership – again not circulation – but readership, respondents are asked “Did you read x newspaper yesterday?” “Did you read it past Sunday?” from there they extrapolate daily, Sunday and 7 day readership
          (they ask the same question of the web property)
          But with TV & radio, the actual time period is measured.
          So if you watched the news at 8am, 6pm and 11pm, you would be included in the measurement for each. but if you pulled a report for viewing from 8a-11:30p, you would be counted one time, not three.
          With newspaper, a person who looks at the front page carries the same weight as the person who reads it cover to cover.
          It’s not a fair measurement.
          With web its even more problematic for an advertiser. Most web sites are sold on an impression basis. So not unique visitor has even the opportunity to see your message.

  4. Sites with paid memberships are be in a better position to aggregate reads from multiple devices because they can roll up all unique views by ID. I wonder if this acts as a disincentive for free sites to add a login or paywall because doing so takes away their hall pass to fudge the numbers as you describe above.

    1. That’s a good point, Ian.

  5. Zachary M. Seward Wednesday, May 1, 2013

    It’s worth noting that, in its quarterly earnings reports, the Times does reveal its actual number of digital subscribers: 676,000 at the end of March http://qz.com/78178

  6. David LaFontaine Wednesday, May 1, 2013

    This is a cross-platform problem – as others have pointed out here, phantom ad impressions and overcounting is hardly confined to newspapers (viz: the mind-blowing numbers that came out of the RadiumOne study, claiming upwards of $400 mill/yr in ad fraud). The response is a dramatic increase in demand for reliable metrics on digital platforms.

    One solution may be the nature of the mobile platform itself. One of the key consumer behaviors we’ve noticed is that users, particularly Millennials, are quite reluctant to share their devices. That “personal & intimate relationship to device” trend, if it continues, will allow us to see just how many people actually view a particular piece of content … which may be good or bad news, depending on how much you rely on claiming that each individual subscription = 5 separate viewers.

    In the meantime – maybe something akin to how Nate Silver aggregated & averaged political polling on FiveThirtyEight? Take comScore, Neilsen, Adobe (Omniture), Experian, etc., throw them all in a hopper & discount the outliers?

  7. First off circulation is a figure that should be completely disregarded. It’s not a valid measure measurement. In fact, most reps selling newspaper ads (unless they are really from the dark ages) don’t even use it any more because it is not comparable to other forms of media use. The best measurement is an integrated audience, it’s the total number of unduplicated, print, digital and niche users. It is as close (not perfect by a long shot) as a comparison with radio listenership and TV viewership. Using something like Scarborough, you can measure the media across and get a comparison. Again it’s not perfect because when building a schedule it counts a full page ad with the same weight as a classified ad – everything is cume of the day in the paper – but the other media are measured in a quarter-hour setting.
    The other reason why circulation should be discounted is that the old ABC now MMA or AMM (whatever) has changed their counting rules and sampling is now allowed to be counted – so when they decide to blanket an area with free copies to try and build subscriptions, newspaper can count those in their Publishers statement or the Audit. They also used to be able to could paid for copies. To be valid and counted a consumer would have to pay at least 50% of the listed price. Now they don’t have to charge anything to be counted.

  8. Circulation is a measurement that shouldn’t be used anymore for several reasons
    We live in a media comparison world and circulation says nothing about the audience – just how many papers got sold. the only people who should care are the bean counters.
    The measurement they should use is integrated audience. It’s the unduplicated reach of their print, digital and niche products. It’s not a perfect measurement, but it at least allows for somewhat of a comparison between radio listenership and TV viewership. It’s a measure of consumption – not a measurement of how many papers were sold.
    The other reason circulation shouldn’t be used is because of all the changes ABC, MMA, AMM or whatever they call themselves have made.
    They now allow sample copies to be counted – so when a newspaper blankets an area and dumps a ton of paper on people’s driveways – those copies can be counted.
    They also changed the free copy rules. They used to be only able to count copies where a consumer paid at least 50% of the stated price. Now consumers just have to pay something, anything. One word about readers per copy – some one intimated that newspapers just make up the number and then come up with circulation. It’s not how it’s measured. You take your measured readership and divide it by your circulation number to get readers per copy. yes it’s a measure made up of 2 unreliable factors, but it’s better than just arbitrarily coming up with a readers per copy number.

    1. The assumption that a reader is a reader is at least as flawed as the metrics being used to measure audience. Beyond the lust for eyeballs lies the real prize: engagement by specific demographic tied to propensity to act.

      Arguing over how copies are counted, or page views are measured, is like arguing over which premium ice cream comes in the best package. Who cares about the box when it’s clearly the content that matters.

      Someone like Nielsen needs to develop measurement of trust, engagement and efficacy metrics that describe how likely an ad message is to result in a desired response, be it call to action or brand awareness, with a specific audience.

      The marketing, which is the actual value, always seems to get brushed aside in the measurement discussion. Which is a shame, because one can’t be solved without the other.

  9. francine hardaway Wednesday, May 1, 2013

    The problem with data in the media industry is that the people whose careers started in media don’t know how to interpret data, and the people who know how to interpret data don’t understand enough about the media industry to be helpful. This will get fixed. The same thing is currently true in ad tech, where advertisers have craptons of data and still don’t know what they are buying when they do a media plan, especially when they go to programmatic. Right now, the use of data isn’t really helping either side, because it’s the proverbial garbage in and garbage out, with thousands of intermediaries on the Lumascape making sure it’s all as obfuscated as possible.

    I’ve just finished my second semester of teaching the Business and Future of Journalism while consulting for two ad tech companies, and I believe we’re still in the middle of the shift to digital. Eventually, all the intermediaries will converge or disappear, and we will be left with the advertiser, the publisher, and data that can be queried correctly. For that to happen, we have to wait for some media-focused applications of “big data.”

  10. Waa! Get off my lawn!

Comments have been disabled for this post