Digital Publishing Reader Revenue Top Stories
5 mins read

“Our North Star is building habit”: How The Atlantic uses data for more effective subscription marketing

Getting your Trinity Audio player ready...

As the competition for reader revenue hots up, the ability to not only establish the triggers that drive people to pay, but to use them effectively will be of paramount importance for publishers looking to get ahead with subscriptions.

On Thursday, Sam Rosen, The Atlantic’s SVP of Growth joined Mather Economics’ Matt Lindsay and World Newsmedia Network’s Martha Williams for a FIPP Insider Webinar looking at data-driven pricing strategies that grow subscription revenue.

Here are some of the key areas Rosen discussed regarding The Atlantic’s shift to subscriptions, and how they evolved their strategy.

A shifting strategy

The Atlantic was comparatively late to (re-)join the paywall party, only launching their offering in September 2019. Prior to that, they had what Rosen described as a “traditional subscription model,” with the options of a print, print and digital, and a “modest” digital-only package which gave readers access to a PDF of the monthly magazine alongside some perks. “We knew that we had to shift our strategy to get ahead, because no one had any idea what was going to happen in the advertising market,” Rosen explained.

Despite originally announcing their intention to introduce a paywall in 2017, The Atlantic ended up delaying the launch, reportedly to better survey its audience about what products they would pay for. Owner Laurene Powell Jobs had also requested delays to allow more time for hiring and platform development, according to the Wall Street Journal.

Rosen added that the feeling internally was that the legacy systems the publisher had in place wouldn’t allow them to move as quickly. “We spent a period of time working with and around our legacy systems and partners…and tried to modernise our technology infrastructure as much as possible,” he said. 

Using data to model pricing

The Atlantic set its pricing for the paywall based on a Van Westendorp Price Sensitivity Meter study, which is a market technique for determining consumer price preferences, with the assistance of digital business platform Piano. This actually resulted in some of the paywall pricing being double that of the previous subscription offering. 

“For digital only, we had been offering it at $25, but that’s now $50; for print and digital, it had been at $35, but it’s now at $60,” Rosen explained. “We asked what kind of features people would have interest in, but it was mostly looking at how we think about annual and monthly pricing.”

Setting those numbers was a delicate balance, with the team keen to avoid either a “race to the bottom” with cheap subscriptions, or maximising the order value with expensive packages that far fewer people would take up. “We certainly didn’t want to reduce price, because even though that would have got us hundreds of thousands more subscribers, we know a lot from the data,” said Rosen.

“One of the nice things about being a little bit later to the game is that we can learn from everyone else’s opportunities for learning – I won’t call them mistakes! So our strategic context was that we wanted to optimise for revenue and subscribers equally.”

One of the biggest goals for the paywall launch was to set out a path to price parity, where the subscribers to the legacy product were paying the same amount as new subscribers. To do this, The Atlantic worked with the team at Mather Economics on a dynamic pricing model.

“We chose to do it by cohorts and segments, rather than by individual subscribers with thousands of different price points,” said Rosen, explaining that the small team and technology meant that data-driven pricing needed to be kept relatively simple. “We simplified it, and we looked at tiers of individuals; what they’re paying, how long they’ve been with us, and whether they’re on auto renew or not.”

A data-driven launch

Armed with the research, The Atlantic’s paywall launched with “a fairly sophisticated, complicated model” projecting paywall stops and a conversion rate by nudges at the bottom of the screen indicating how many free articles the reader had left.

The response to the initial launch blew the team’s expectations out of the water. “By multiple digits, we exceeded our expectations,” said Rosen. “We settled fairly quickly into a business as usual state, but were still exceeding our model expectations probably by double on a monthly basis.”

But even with extensive research, there were still surprises after launch, “The uptake of our [$100 per year] premium bundle was well below what we saw from a survey standpoint,” admitted Rosen. “That said, we wanted to keep it anyway because that meant we could increase demand for the print and digital, which was a higher [margin product than the digital-only product], and actually be more profitable…then we were comfortable with a lack of significant adoption.”

Monthly and annual price points

One noticeable feature of The Atlantic’s current subscription offerings is that there is no monthly option, or significant discounting. Monthly options were certainly raised during the research phase, but Rosen explained that the team decided to go in a different direction for now as it instinctively didn’t feel right.

“My philosophy with qualitative data is that it’s good to get signals, but it shouldn’t be too prescriptive in terms of your strategy,” he explained. “We saw from [the research] that there was a massive price disparity between what the monthly and annual people said they would pay, with monthly saying $10 a month sounds perfectly reasonable, but annually no more than $50 a year.”

Rosen believes that the monthly price point comes as a reference from Netflix and other entertainment, and that the willingness to pay annually is the one to take more seriously. He emphasised that, as long as the demand is sustained for annual packaging, that is where the focus will be, particularly as the team turn their attention to retention efforts towards the latter half of 2020.

Overall, the tests so far have been really successful, and The Atlantic will continue for a while with the dynamic pricing model to slowly bring everyone in line with the new prices. “We know the demand curve is not going to last forever, so the question is when, not if we will launch monthly and trial pricing to grab demand,” Rosen explained. “But for the moment, we’re bowled over by the support that we’re seeing, and the demand for people paying for journalism.”

Building habit

For now, The Atlantic is monitoring user behaviour and learning what behaviours indicate a willingness to pay in the future. “We see repeat gateway hitting is a really important indicator of subscription propensity, and even reading two articles in a single session is most correlated with subscribing in a single day,” said Rosen.

“Our North Star is habit, and getting people back to the site multiple days in a 30 day period, because it means they’re not only engaging with the journalism, but that they’ll also be more likely to hit the paywall,” he concluded. 

Watch the full webinar here:


FIPP are continuing their series of Insider webinars during May with a session on Tuesday 19th on Sustainable revenue: How European premium publishers turned to e-commerce; and Thursday 21st with Dow Jones on opportunities and challenges for publishers in lockdown.