The relationship between circulation and readership

Circulation and readership are terms that are used freely in our industry, however a question that is often asked is why do the two not correlate like expected? Simon Baty explains their at times complex relationship.

People often ask about circulation and readership, and how at times it can seem like one is going up while the other is going down, The Newspaper Works research and insights manager Simon Baty says.

“It is a valid question, and in general you do expect them to move in the same direction,” Mr Baty said.

“However, there are various short term factors that can disguise the fact that they are trending in the same direction over the longer term.”

When referring to newspapers and other printed publications, circulation is the number of copies distributed on an average day, however readership is the number of people who read a newspaper or other publication each day.

“While you would expect the two would move together there is an obvious explanation as to why they don’t,” he said.

“One good example is the impact of promotions.  If a publisher lowers the cover price, then some people who would previously have read someone else’s copy might start buying the publication.  In this example, sales would be going up without a corresponding rise in readership.

“Nowadays, many publications are delivered not only in print, but also via websites, m.sites and apps, which complicates the relationship between print sales and readership.

“With several different ways to access similar content, you’d expect that pass-on readers might be more sensitive to the impact of promotional activity in any of those platforms, or changes in format and content in any of those platforms, which might prompt a switch partly or completely to a different platform.”

Another reason for the difference between the two is because the different sets of numbers are measured at different times.

“Audited circulation figures are an average across the most recent quarter.  Readership on the other hand is an average over a 12 month period.  You therefore tend to see changes coming through more quickly in circulation.” Mr Baty said.

If you would like to know more about the relationship between circulation and readership click here to view the comprehensive paper put together by Ipsos about the factors that influence both sets of numbers.

2 comments

  1. There is a long history of researchers attempting to understand the apparent lack of a systematic relationship between circulation and readership, e.g. Goerlich (1993) ‘The Relationship of Changes in Circulation to Changes in Total Audience” (Session Papers of the Worldwide Readership Research Symposium 6). While intuitively it seems reasonable to expect some such relationship to be discoverable, none has been reported – at least, not with the most widely employed readership measurement methodologies.

    This page links to an excellent overview of the many factors that may confound an analysis of the relationship between circulation and readership trends (http://www.newsmediaworks.com.au/readership-and-circulation-reasons-why-they-might-diverge/). It is possible however for such factors to be allowed for in experimental controls, but still no relationship emerges. In contrast, it is well-established that even only slight changes in readership measurement methodologies can have very significant and predictable effects upon the results.

    The fact is, readership claims are highly sensitive to many aspects of survey data-collection processes, including the measurement construct itself (i.e. what you set out to measure), many aspects of questionnaire design, method of contact, co-operation rates, interview protocols etc. These non-sampling effects can produce far greater statistical ‘noise’ in the results than sampling error. So it is simply wrong to say “The degree to which sample variation alone is responsible for period-on-period differences in the readership estimate for a given title can be demonstrated by what is known as significance testing. If the test shows the difference is significant, there is a strong possibility that it is a reflection of a real change in readership”. On the contrary, if a difference appears to be statistically significant yet contrary to circulation change, then it is most probably due to non-sampling effects.

    The great worry is that with print media currently in such a state of flux, non-sampling effects in the near future may systematically drive readership estimates in the opposite direction to not only circulation trends but also real life reading behaviour. It is crucially important that much more should be done to understand why readership does not correlate better with circulation change. Well-designed experiments may produce findings that enable readership measurement methodologies to be greatly improved and future-proofed.

  2. Very true Simon. There is also the issue of ‘diminishing returns’. People generally expect the Readers Per Copy (RPC) to remain static – hence the expectation that as circulation goes up or down, then so too should readership. This is easily debunked – albeit with an extreme example. Let’s say a newspaper in a city of 5 million has a circulation of 250,000 and a readership of 750,000. That is, the RPC is 3.0. But what if (say – again this is extreme but it demonstrates the point) the circulation goes up tenfold to 2.5m. Applying the RPC factor of 3.0 would imply a readership of 7.5m – 50% more than then population. There is a point at which as circulation increases, RPC MUST decrease (albeit readership can still rise – but not as quickly). The inverse also tends to apply.

Leave a comment