It is time to announce a winner, ladies and gentlemen! As we head into the last quarter of 2012, we now seem to have a runaway winner in the battle of the paywall models – the metered model.

A recent Newspapers Association of America study of 156 newspapers revealed that 87% of the newspapers profiled had adopted the metered model. And since the study broke, there have been further announcements. See here and here. True, the study is restricted to the USA, and it is only about a tenth of the 1,400 odd daily newspapers published in the US. But it is enough to convince even the biggest critic of the metered model that it is time to announce it as the decisive winner.

Why did the metered model decisively win?

Firstly let us look at the various options available to newspapers for their digital editions / sites.

Paywall Models

Paywall Models

1 Hard Paywall

The Hard Paywall is an endangered species. With a hard paywall where you need to pay to get access to any content on the website, it is a given that your online readership will fall sharply as happened to Newsday and The Times. And when that happens, it can pretty much erode your advertising revenue.

The only case in which a hard paywall is justified is for a monopoly local paper. The risk here is much lower – the reader does not have any other option for local news, and it is unlikely that there is a large online advertising component which runs the risk of a drop.

2 Some Free, Some Paid

The WSJ’s model of keeping some articles free and some paid is referred to as Freemium / Editor’s Choice (an altogether better phrase) by Sandy MacLeod in an article where she discusses the various business models that newspapers can use to monetize content.

The Freemium model rests on the well-founded belief that what some people will pay for is not what most people will like to read. So there will be articles in niches such as Tax Law which is arcane enough that most people will not want to read it, but useful for people in law, accounting that they would not mind paying for it. Conversely there may be articles such as developments in the Eurozone, which most people will want to read about, but since the article doesn’t offer much of a value-add to other sites, it might be hard to justify putting it behind a paywall. Hence you might as well keep this free to bring in the traffic that you monetize through selling advertising.

3 Combo Model

Sandy MacLeod in her article also alludes to a combo model – used by Boston Globe which publishes boston.com (which was their online site originally) as a free site and bostonglobe.com as the paid site. I am not sure if there are any other markets where such a two-track model exists. (With just 23,000 digital subscribers by end of Q2 2012, this doesn’t seem to have worked out the way they planned.) 

4 The Metered Model

In this model, the reader is allowed to read a set number of articles, typically between 10 and 25 every month. Beyond this, the reader has to sign up for subscription.

The Metered Model’s success rests on the simple reason that implementing it doesn’t lead to a drop in online viewership, at least not a significant one anyway. This protects the Publisher from a sharp fall in online advertising. In addition, the Publisher now benefits from the incremental subscription revenues that now begin to flow.

Thus the Metered Model has much less material side-effects than the others, and this in short is the reason for its runaway success.

Why doesn’t readership drop significantly with the Metered Model?

For one, the limit of 20 stories (or less) is sufficiently large to not restrict casual readers (who reach the site thanks to google search or a facebook link) from accessing the site. And even if you are not a casual visitor, if you do access the site through multiple devices such as smartphones, tablets and laptops, it will be a lot of stories before you are required to pay.

Thus, it is largely the heavy-duty users of the site (who typically tend to be a smaller sub-set of the reader base – The Dow Jones Local Media Group found that 3/4ths of visits came from just 1/3rd of their audiences suggesting the Pareto Principle at work) who will need to subscribe. And typically, if you are consuming more than 2-3 stories a day on the site, it is highly likely that you see value in the site and don’t mind subscribing.

Thus by charging only those readers who most value the content on offer (and are thus unlikely to drop off), the metered model is able to ‘price-discriminate’ effectively between its consumer segments.

Over the last two years, several examples of successful implementation of the Metered Model have emerged. Some of these include –

–       Ellie Behling writes about how the Concord Monitor and Augusta Chronicle have implemented the Metered Model without any sharp drops. In fact in the case of the Monitor there is been a slight uptick (Article dated 25 Aug ’11)

–       Rob O’Regan writes about the experiences of Lubbock Avalanche-Journal and Lancasteronline where the traffic has now crept back

–       Joe Pompeo writes about how NYT’s traffic has remained steady at 48 million monthly unique visitors (UVs) a year following the implementation of the Metered Model, though page views (PVs) have suffered slightly.

Press+ which manages digital paywalls  (typically the Metered Model) for many smaller US papers revealed last January that a study of two dozen publications where the metered model had been operational for ‘several months’ had seen page views impacted by 0 to 20% while UVs dropped from 0 to 7%. It said that no paper had seen a decline in ad revenue.

Why advertising doesn’t get impacted with the Metered Model

With negligible impact on traffic, there is virtually no impact on ad revenue as well.

In fact, Frederic Filoux in his wonderful Monday Note article states that advertisers are paying higher for audience behind paywalls, as much as 30% he says. I couldn’t find any independent corroboration of this however.

The higher prices paid by advertisers for readers behind paywalls could be due to Publishers’ enhanced ability to

1)    offer better targeted ads / innovations (thanks to the data and insight that they have on their loyal readers) which brings in more premium advertising. See an advertiser’s view on this here.

2)    convince advertisers that the readers who have registered are more premium (because they can pay and are comfortable with the digital medium) and are more engaged with the brand (as they are more loyal)

I would also argue here, without any supporting evidence however, that perhaps charging for content is also a way of signaling to the market that your content / website is of higher quality (as it is paid for), and thus attracts more premium audiences.

Filoux also illustrates how the low Sell-Through Rate (STR) of the newspaper websites (average STR of about 60%, which means about 40% of the inventory on the website is unsold) works in the newspapers’ favour when the site goes pay. The fall of up to 20% in PVs we saw earlier thus does not affect advertising sales as this is fundamentally the unsold inventory that is getting reduced.

As Filoux says “An online media with a sell-through rate of 55% can allow a 45% decline in page views before eroding its ad revenue.”

Generalizing another Paywall Principle

In a previous blog post, we discussed when it made sense for a publisher to institute a paywall, and when it didn’t make sense.

In this post, I propose to suggest a principle, call it Paywall Principle #2, that suggests that as long as a newspaper website has

1)    a small proportion of readers accounting for a large proportion of articles read (as seen earlier with the Dow Jones Local Media Group)

2)    a sizeable proportion of its advertising inventory unsold

it can adopt the Metered Model without any sizeable impact in its UVs and advertising revenue. It really doesn’t matter whether the content is high-class or not (it may matter in the number of online subscribers though). All it matters for the UVs and advertising revenue to hold are the above two conditions being valid. 

Summarizing my arguments

Nature of Paywall Qualifiers Implications
Hard Paywall
  • Monopoly market
  • Limited advertising inventory or negligible ad revenue

Both of the above are seen typically seen with hyperlocal / community papers

–       Number of subscribers will be very limited for general interest papers

–       Online advertising may collapse as seen with The Times

Some Free, Some Paid
  • Paper has to have exclusivity or dominance in a few niches (for content that you want to price) – sports content such as Packer Insider from the Milwaukee Journal-Sentinel, selected business content from WSJ etc that are behind paywalls
–       Will need to invest in  editorial resources who can judge which story needs to be behind paywalls

–       Reader dis-satisfaction is likely to be higher in this model

Metered Model
  • Online readership of articles should ideally follow some variant of the Pareto Principle
  • Unsold advertising inventory – the higher the better
–       Higher chances of revenue leakage – readers can find ways to game the system, accessing the website through multiple devices to delay hitting the paywall

So there you have it! An explanation of why the Metered Model won out!

As I get additional data, I shall update this blog post. One data point that I am looking for is the market share of various paywalls, and seeing what is the Metered Model’s share in it. I would be very surprised if it was lesser than 90% (perhaps even 95%).

I am also looking for examples of  Hard Paywalls beyond The Times of UK. If you do know of any such publications / websites, do send me the links. Also, if there are any cases (I am sure there are) of Hard Paywalls migrating to the Metered Model (Is the Australian Financial Review one such?) do send me those as well!