The metered model deserves a closer look. One the dirtiest little secrets of the online media business is the actual number of truly loyal readers — as opposed to fly-bys. No one really wants to know (let alone let anyone else know). Using a broad brush, about half of the audience is composed of casual users dropping by less than 3 times a month, or sent by search engines; 25% come more than 10 times a month. Over the years, as audience segmentation increased, media buyers (and publishers) selected the simplistic counting of Unique Visitors (UVs) as the metric of choice. In the meantime, all forms of Search Engine Optimization (SEO) and Search Engine Marketing (SEM) outfits have further elevated the collecting UVs as the primary goal for online publishers. Along with that practice came cheating. In order to inflate their UV numbers, many large news sites now rely on third party services such games that have nothing to do with their core business.
This distortion contributed to the erosion in advertising prices. Media buyers might by cynical, but they are not stupid. They know that a growing percentage of audiences is composed of accidental visitors with no brand loyalty whatsoever and who offer no attractive demographics. Combined to the “unlimited supply” factor inherent to the internet business, the result is a downward spiral for ad prices. These are important factors to keep in mind while considering paid-for systems.
News organization have implemented such systems in different gradations. At the far end of the spectrum, we have the Times of London: no access to the site without first paying. That’s is the riskiest option. The site ends up losing 90% of its audience (and the related advertising revenue) but hopes to offset the loss by gathering enough online subscribers. Without the promotional booster of free contents, this is a challenge – to say the least.
Others choose to give some of the site for free and put the most valuable contents — sometimes the digital version of the print edition — behind a paywall. This doesn’t always make economical sense as many readers are happy enough with the free content part. Editorially speaking, this leads to the creation of two categories: cheap fodder available for free (often created by junior staffers), and more “noble” content produced by the most senior members of the newsroom who also feed the print version. This works fine for a brand associated with significant added value, or specialized (such as business news), or one that dominates its own market. The most successful paywall implementation has been the Wall Street Journal: it now has more than 1m paid subscribers, but it took 10 years to get there.
The third option involves a metered system. The principle is simple: once you’ve seen a certain number of stories in a given period of time, you need to become a paid subscriber to keep viewing the site. Some newspapers have been quite successful at deploying such a metered system.
For example, the Financial Times has set the cursor to 10 stories per month before hitting the paywall, after which the reader is asked to pay between € 4.99 and €7.49 (about $7.30 and $11) per month, depending on the package deal. A high price for really premium content. So far, FT.com has 3.4m registered users of which 224,000 have been converted to paid-for contents (+8% for Q1 2011). This translate into €20m to €25m extra revenue, only from subscribers (the service has been launched in October 2007). Currently, digital revenue (both ads and subscriptions) accounts for 30% of the FT’s revenue; according to FT execs, it is expected to reach 50% in 2013.
For the meter, finding the right setting is far from trivial. The trick is to decide how many free stories will be allowed before hitting the paywall, and how much to charge thereafter. In New York, three weeks ago, I spoke with Gordon Crovitz. With Steven Brill, Gordon co-founded Press+, which creates bespoke metered system for online medias. Press+ provides a complete set of e-commerce tools for publishers, from the access mechanism to the transaction system. It works with passes (daily, weekly), subscriptions plans (monthly or annual), topical packages, bundles and ancillary products.
Determining the right formula is usually done through A/B testing. Crovitz and Brill explain: the publisher will test two or three levels of free access (5, 10, 15 stories per month) and the same number of prices ($5 to $10 or maybe $15 a month). A few months of testing will determine the right formula. Typical ingredients are: the type of content, surrounding competition and possible alternative for the customers, the publisher’s willingness to bundle digital and print products. Metering can also be attractive for out-of-market audiences: an Australian newspaper will be free for its domestic audience but will charge overseas readers consuming more than 10 stories a month.
Another factor is the site’s advertising structure. The amount of inventory sold to advertisers varies widely. In the US market, the “sell-trough” ratio is about 60%, but it can go as low as 30% on some markets. This means the media can sustain some loss in page views due to the implementation of the metered system without losing ad revenue. An online media with a sell-trough rate of 55% can allow a 45% decrease in page views before eroding its ad revenue. According to Press+, traffic losses from implementing a meter are modest, ranging from 0% to 20% as counted in page views, and 0% to 7% in UVs.
Let’s try back-of-the-envelope calculations. A site gets 5m UVs and 100m page views per month; its yearly ARPU (Average Revenue Per User) coming from advertising is $3. This results in a yearly revenue of $15m. Now suppose only 20% of its audience reads more than 15 stories a month and one out of ten such readers are willing to pay $10 a month. The additional revenue will be: 5m UVs x 20% hitting the paywall x 10% willing to pay $100/year (discount included) = $10m in additional income — without depleting its advertising revenue. Actually, experience shows advertisers are now paying roughly 30% more for readers reached behind a paywall. All this before the 20% cut taken by Press+.
Naturally, as the saying goes, YMMV (Your Mileage May Vary), actual results will depend on many factors, one of them being how the pricing system is set (the simpler, the better). Again, a rigorous test of all hypotheses is critical. Metered systems are the opposite of the one-size-fits-all.