Inspired tactics: A news subscription series – Part 2, Tech turns readers into subscribers
By Bihag Karnani
As a senior product manager at Google, Bihag Karnani oversees the development of Reader Revenue Manager – a digital subscription service designed specifically for media companies. On average, he meets with approximately 50 to 70 publishers during the course of a year.
Part 1 of this four-part series on the lessons of technology companies regarding the increase of subscriptions looked at first party data and the first 100 days. This part is looking at what occurs prior to the conversion; the moment when a reader decides if he/she will pay.
News media organisations treat the paywall as a barrier. Technology companies treat the paywall as an evaluation mechanism. Do we request payment? At what point? How much? Which type of offer should we provide? Should we request an e-mail instead of a payment method? Can we allow the user to proceed without requesting payment and try again at a later date?
The distinction between evaluating a single decision or continuing to evaluate decisions is rapidly becoming the distinction between a subscription based business model which increases in size and one that does not.
Trials: Opt-in vs. opt-out
Each news media organisation must choose strategically the length of their free trials and whether they require a credit card in advance. Two types of trials exist: opt-out trials (subscriber must actively cancel his/her subscription before being charged); and opt-in trials (subscriber must take action to subscribe after the end of the trial). The results differ significantly depending upon whether a credit card is requested in advance.
Open View Partners has followed many SaaS (Software As A Service) companies using both opt-out and opt-in trials. An opt-out trial produces a conversion from trial to paid subscription of 40-50 percent, however, opt-out trials also create a reduction in sign-ups of 50-80 percent.
This is due to a very large friction point created by requiring a credit card to start a trial. An opt-in trial produces a conversion from trial to paid subscription of 15-25 percent; however, opt-in trials produce approximately 35 percent more sign-ups. Only about 12 percent of SaaS companies currently using free trials require a credit card to begin a trial. The other 88 percent have concluded that it is more important to capture users in their funnel than to maximize their conversion rates.
Opt-out models are suitable for higher end B2B subscription products. Because the cost per contract is large enough that losing a small number of potential subscribers is negligible, you should limit the number of tire kickers accessing your content.
An example of an opt-out model is The Wall Street Journal’s “4 euros a month for one year” or the Financial Times’ “$1 for four weeks,” both of which require a credit card.
Opt-in models are generally best for increasing the upper funnel volume of local dailies. Because there are just too few readers who will even look at your content, it is better to capture all interested readers regardless of the conversion rate. In other words, getting a smaller percentage of a larger total population is more desirable than the opposite.
A new trend in SaaS pricing is the “reverse trial.” This means giving each new customer complete premium access for 14 – 30 days and then automatically downgrading them to a permanent free version with less functionality.
The idea behind this method is that users have become accustomed to using the premium version during the trial period, and therefore miss its capabilities once they lose access. According to Open View, reverse trials convert at a 7 – 21 percent rate – a reasonable middle ground between offering free versions of services/products as well as charging for full access to services/products.
Paywall placement & frequency
After determining whether or not to use a trial model, the publisher must now determine the placement and frequency of the paywall. The honest truth is that most publishers have not yet realised that the simplest “meter,” such as “Three Free Articles – then you hit the paywall every time and on every Reader,” allows for significantly greater revenue to be left on the table.
Two variables play the largest role in determining where and how often to place a paywall: placement and frequency. Both can be determined without machine learning.
The Boston Globe was very open about experimenting with different meter placements. At a WAN IFRA Digital Subscription Summit, Esfand Pourmand, then Chief Product Officer at The Boston Globe, discussed his company’s meter testing. Most newspapers offer readers a certain amount of free content on a monthly basis.
The Boston Globe extended that window from 45 to 60 days, and reduced the amount of free content within that window to two.
By extending the window and limiting the number of free articles available, readers develop a habit of reading your paper before hitting the paywall. Data showed that extending the window and limiting the free articles available led to a higher conversion rate compared to a tighter monthly meter.
One of the Globe’s other moves was to shut down those work-arounds. Accessing articles through search, social media referrals, or incognito mode – whatever methods readers used to circumvent the pay-wall. This approach only works if the prompts occur when they’re supposed to.
The second variable is the type of content that will be gated. Every story is not created equal. News breaking now is ubiquitous. Readers can find it anywhere; so there is little incentive to subscribe.
Exclusives, opinions, analytical pieces, and beat reporting that readers can’t access from competitors are the real conversion tools. The best publishers create a premium experience for the content that is exclusive to them and leave open the commodity content.
The easiest form of this is an editor-made call. Editors determine whether certain stories qualify as “premium” and therefore eligible for gating while others do not.
The more advanced form adds basic behavioural cues such as; a reader who has visited your site five times this month and has read three opinion pieces is much further along in converting than a reader who arrived once via a google search. Both types receive a different paywall experience.
It’s not the specifics of any particular tool that are worth taking away. Rather it’s the underlying philosophy. The paywall is not a wall – it is a succession of conversion opportunities.
Each opportunity takes on a different shape, depending upon what the reader is doing at the time. For most publishers, the first step is to begin treating the pay-wall as multiple rules rather than one universal rule and begin customising the experiences based on which moments and which content convert.
Urgency, anchoring, and discounted on-ramps
If a user has experienced the conversion moment, then how the user receives the offer becomes relevant. As noted earlier, Mailchimp has found that simply including countdown timers in emails increases click-through rates by 10-20 percent. Similarly, according to Duolingo’s Chief Product Officer Cem Kansu, “writing ‘Last Chance’ in the subject line causes click-through rates to explode”. There is no easier or higher impact lever than urgency when it comes to conversion copy.
In terms of news publishing, this creates opportunities for time-limited campaigns. Examples include; “The election day rate expires at Midnight on Tuesday,” “The new year subscription drive ends 15 January.”
Most major publishers run some variation of these types of limited-time promotions – the FT, The New York Times, The Washington Post, and The Wall Street Journal all utilise limited-time introductory pricing, but many regional and local publishers fail to take advantage of these opportunities for increased conversion lift.
A second option is utilising introductory pricing.
Le Monde has been particularly vocal about creating a well-structured progressive pricing plan. The initial pricing plan includes a first-month price point of €1 and subsequent monthly prices of €9.90 during the first-year, €13 in year-two, and €17.90 thereafter. In a WAN-IFRA post published last year, Flavia Barbosa Ferreira (Le Monde’s head of marketing retention) broke-down her team’s process for developing the pricing plan.
The first iteration of the pricing plan resulted in strong acquisition numbers, however also resulted in high churn at the upgrade points.
To address this issue, Ferreira’s team worked with a data analyst specialising in subscriptions to build a churn model using 13 months of historical data and then utilised this model to predict retention levels over a period of 47 months (see image below).
The resulting redesign of the pricing plan struck a balance between long-term revenue growth and customer retention. Le Monde reported an increase in subscribers of 10 percent, revenue growth of 12 percent, and increasing ARPUs (average revenue per subscriber) since 2020 – reaching approximately €12 by March 2025.
When discussing her team’s efforts, Ferreira made note of an important takeaway: “Rather than forcing everyone to fit into our single model, we allowed users to select their own value level – and they remained.”
The first party data foundation
As previously discussed, all of these concepts rely on having knowledge about the reader themselves. A meter is only as intelligent as the amount of data it uses to make its determinations. Le Monde’s churn model required 13 months of historical data regarding their subscribers. Even The Boston Globe’s decision regarding when to show windows and which content to display would improve with additional reader behavior-based signals.
This is why the registration wall needs to be placed before the paywall and not after it. Anonymous visitors provide virtually no useful information. Registered readers – regardless of whether or not they have paid – provide email addresses, reading histories, topics of interest and a method for follow-up contact.
The tenfold conversion lift associated with known readers compared to anonymous readers that were referenced in part 1 compound with each conversion mechanism presented here. Improved trial periods. Smart paywall placements. Targeted offers.
Every time you make a change, it’s an experiment
There are so many ways to look at how Duolingo makes changes. They run over 300 different experiments every quarter. When The Boston Globe chose the 60 day trial period, they did it by writing about it in two different articles.
For years, Le Monde has been making adjustments to their product offerings as new information became available.
The shift in the way that publishers think isn’t big but is significant. Each time you place a paywall, that is an experiment. Each time you extend or shorten the trial period, that is an experiment. Every time you create a new offer, that is an experiment. Every time you write some copy to try to get someone to buy sooner, that is an experiment.
Those publishers who make each of these choices without thinking of them as hypotheses to be tested and improved upon, will likely lose market share to those who consider each choice as a hypothesis to be tested.
There are also many tools available today that allow small newsroom staffs to conduct paywall testing. Tools such as Statsig, GrowthBook, Optimizely, etc., provide newsroom staffs with the ability to set up testing on aspects of their paywalls such as copy used on buttons, colour of buttons, order in which plans are presented and various other options for testing paywalls. The increase in conversion rates resulting from simple tests (in terms of percentage) can range anywhere from 10 to 30 percent depending on the ease of the test.
What to do
This week: Audit your current paywall and trial process. What are your rules? Are there multiple segments receiving different versions of your rules? If your answer is one rule; everyone receives the same version of your rules then you should take note that you are using the minimum standard as a reference point.
This quarter: Run a single paywall experiment within this quarter. Either lengthen or shorten your meter window. Also, either gate your high end exclusive content tighter and open up your commodity content further. Compare your conversions against your previous baseline.
This year: Treat every conversion decision as if it were an experiment. Trial lengths. Time when the paywall goes live. Copy used in offers. Price tiers. Establish a regular schedule for structured experimentation and establish a team responsible for tracking the results of the experimentation.
The publishers that win out in the subscription wars going forward will be those who no longer treat conversion decisions individually, but rather continuously.
About the author: Bihag Karnani is a Senior Product Manager at Google, where he leads Reader Revenue Manager (RRM), Google’s subscription platform for news publishers, and drives user personalisation strategy for AI Mode and AI Overviews. He works with publishers globally on subscription strategy.
