Welcome to Readers First, a monthly newsletter for INMA members that represents the public face of a broader association initiative on reader revenue innovation.
As Researcher-in-Residence for INMA, I will bring to you what I am seeing worldwide on reader revenue and the many corollary subjects that stem from this new focus: culture, digital subscriptions, engagement, battling churn, and the broader economics of content.
You are receiving this newsletter as an INMA member. If this is of high interest to you and you want to continue receiving this newsletter, click here and opt in for the next year.
Help me make this newsletter more useful for you. What do you like? What would you change? E-mail me at: firstname.lastname@example.org.
1. PAYWALL MYTHS. Freemium wins the popularity contest with the meter, but what if it’s wrong?
Nearly half, or 45%, of nationwide news media outlets choose the so-called freemium paywall, in which they classify what articles or features are free and which are paid. Only one in 10 outlets, or 11%, use a meter to count the number of articles a visitor reads to filter out the heavy users. One-third of publishers, or 32%, limit their paid content offer to an e-replica.
- These are findings of my survey of 129 Web sites charging for content in 33 Western markets that included Australia, Canada, all the European Union countries, Hong Kong, South Korea, Singapore, and the United States.
- From April to June 2018, my associates and I clicked through 397 news Web sites with the Top 15 highest brand reach in their respective countries to learn which publishers charge for content. One-third of the Web sites, or 32%, charged for something, at least for an e-replica. We tested the rules of the paywalls, too: whether they sampled free articles by the meter, or differentiated between free and paid, premium content. The survey was a part of my research on subscriptions for the Oxford’s Reuters Institute, of which the full results will be published later this year by the University of Oxford.
What if the majority of publishers is wrong? Contrary to what many think, it might be easier to predict subscriber behaviour by measuring the number of articles they read than by guessing topics that might convert them.
This is a result of an advanced data analysis of actual behaviours of would-be subscribers of two mid-sized news websites — in Canada and in Norway — that I partnered on with Cxense, a digital insights company.
Continue for the detailed results and an analysis of hybrid paywalls.
We followed behaviours of millions of news visitors over two weeks while training a machine learning model that predicted the likelihood of them becoming subscribers within the next two days. We looked for correlations with different temporal variables such as the number of visits, page views, or time spent on the site, and topics of pages that they visited.
We found both temporal variables and topic categories being positively correlated with the propensity to subscribe. Although all correlations were surprisingly weak, the relationship between how a visitor uses the site seems to be stronger signal than her interests.
In general, temporal variables are more universal than topics. Simply put, more page views are useful signals to any Web site whatever its content is about. In my study with Cxense, the converting topics actually varied across Web sites.
“Working with clients, we mostly use temporal information for estimating how likely a user is to subscribe: how often and how recently the user has visited,” confirmed Elena Volkova, data scientist at Cxense. “Even the simple number of page views correlates with whether or not the person will subscribe in the future.”
Interviewing news executives, I find the temporal variables often are the core of the prediction models that many leading publishers use. A good example is the Financial Times in the United Kingdom:
- The FT has actually developed an engagement metric used by both editors and marketers. They called it “RFV” for Recency (of the last visit), Frequency (of visits within the last 90 days), and Volume (of articles read in that time). The media company uses these variables to score all potential and current subscribers. For years, the RFV scores read as Recency, Frequency and (Monetary) Value were used to segment customers in industries with high transaction volumes such as direct mail shops.
- The FT found its version of the RFV correlated with the likelihood to subscribe (positively), to cancel the subscription (negatively), and with lifetime value of a customer (positively). As Tom Betts, FT’s chief data officer, put it: “Engaged readers equal healthy customer base, and that is the best predictor of future revenues.”
- If it’s usage that actually drives purchases, the old-fashioned meter suddenly feels like a simple, yet powerful way to segment users by behaviour, and show a subscription offer to the heaviest users.
The most sophisticated segmentation models analyse dozens of signals to determine the propensity to subscribe, and they often combine both temporal and interest-based variables.
- The Wall Street Journal has developed a sophisticated paywall that is adaptive to user attributes and adds some business considerations to the mix. The company calls it “a dynamic paywall.” It scores every visitor by their propensity and differentiates the experience on the site accordingly. For example, if one is unlikely to subscribe, one gets a sample article for free but with ads, so the publisher at least gets some ad revenue. In general, if one is likely to subscribe, the article is locked, and one gets an offer to subscribe. But if there is an ad campaign suffering from undelivered page views, the paywall unlocks the articles until the ads burn.
- The scoring model is based on more than 60 signals such as whether a reader is visiting for the first time, whether she reads content from different sections, what is an operating system or a device she is using, and what is her physical location. To give an actual score, behavioural data may be cross-matched with other data, e.g. information about location may be enhanced by U.S. census data on average household income by postal codes, so the model would consider whether a visitor lives in a rich or a poor neighbourhood.
- How strong are the temporal variables such as a number of active days or a number of articles read? Karl Wells, general manager of the WSJ Membership, smiled: “Strong.”
- One can look at dynamic paywalls as the complex versions of the hybrids between freemium and metered models, and the hybrids as an attempt to combine two variables of different sorts — behaviours and interests.
Inspired to learn more? Read “Gannett, FT, Ekstra Bladet newsrooms are key to paid content models” and “The art of listening: lessons in analytics at Wall Street Journal.”
2. INMA CLINIC. Paywall check-up with Dr. Greg: Analysing the line of the Washington Post’s subscribers growth.
I am a passionate reader of charts featuring adoption curves of new products, and I love to investigate the reasons for their ups and downs. Thus I found The Washington Post’s growth story fascinating and puzzling (below is from a presentation by Miki Toliver-King, vice president/marketing, The Washington Post from June 2018).
- Look at 2013-2016: The company launched an online paywall with a meter set at 20 articles, then they go down to 10, and to five, and … nothing happens. The pulse is flat. Was the Post online a dead man walking when it was purchased by Jeff Bezos? One explanation might be that after converting its brand enthusiasts, the Post realised its Web site enjoyed almost no audience but fly-bys visiting from social networks and search engines. Then in just two years they managed to change that and sell a million subscriptions!
What happened? Find out about the Post’s key lead generator.
- 2016-2018: What has changed? Donald Trump won the election, raising eyebrows and bumping traffic to news sites. The Post regularly led with exclusive coverage grabbing attention, clicks and returning users.
Slowly, the early investments in the product such as mobile apps and the underlying tech infrastructure must have paid off. Today, the Web site is the newspaper’s main acquisition channel‚ selling every six of 10 subscriptions via on-site ads, buttons, and article locks.
The Washington Post used to heavily discounts its offers, but it changed — the Post claims less than one-third of new subscribers are enrolled on a discounted rate.
The biggest impact though had perhaps a successful effort by the Post to build habits of frequent reading by a big number of users: It raised frequent visitors with its dozens of newsletters. The proof? Sales spiked after it started to meter clicks from those e-mails.
According to Miki Tolver-King, vice president marketing at the Post, in just two years the newspaper has enrolled 4.6 million newsletter subscribers, and it has been driving lead generation since then.
3. LETTER FROM THE ROAD: What would you do if a tech billionaire bought your newspaper?
I am writing this month’s newsletter from Hong Kong where I have visited a new impressive newsroom of the South China Morning Post in Causeway Bay and lectured the Post’s digital team on user engagement. (I met the team right after the super typhoon Mangkhut struck Hong Kong. I followed their outstanding coverage from my hotel room where I was stranded for a day, feeling how the building swayed in the wind.)
Brian Rhoads, a managing editor of the South China Morning Post, showed me around a two-level open space on the 18th and 19th floors of the Times Square Tower. Regular seating, high standing desks, and collaborative desks for project teams, all shiny new and flex. Walls with dozens of screens featuring behaviours of the Post’s growing international audience. There is a hotel-style lounge for informal meetings, a gaming room, a yoga gym, a cafe and even a pub brewing signature beer. The Post’s office looks more like a headquarters of Google or Spotify rather than a 115-year-old newspaper.
No wonder: CEO Gary Liu is an ex-Google and ex-Spotify executive. The company’s new owner is Alibaba, an e-commerce giant, nicknamed the “Amazon of China.” With his investment in the famous Hong Kong newspaper, Alibaba’s chairman Jack Ma seems to have followed the steps of Jeff Bezos, a founder of Amazon, who bought himself The Washington Post.
As different as the two Posts are — in their roots, coverage and the political environment they operate within — they both have shifted their focus from local markets, respectively Hong Kong and Washington D.C., to the one global market of English-language news. They wish to be the guides to the world’s two greatest powers: the United States and China.
Jealous? Learn what is the dark side of having access to money, the best know-how, and a new shiny office.
To grow, the two Posts have gone on a recruitment spree, and the Hong Kong one has added 100 people to the newsroom in a year, to a total of around 350, as well as an additional 100 to product and tech teams. The Posts have been redesigning the tech infrastructure of the companies well beyond the CMS, for example, investing in data science and automation. They relaunched the old and launched new digital products. The South China Morning Post launched three digital-only services, such as Abacus on the Chinese fast-growing tech sector that I recommend you check. The two Posts have aggressively expanded their reach, with the online audience of the South China Morning Post having grown five-fold since last year, according to the internal data.
SCMP’s Liu and Rhoads talk about nothing but growth. Lucky them. Their Post, similarly to the one in Washington, enjoys what so many other publishers don’t. The sole owner is one of the richest tech oligarchs in the world, with whom comes a clarity about the business purpose and a strategic focus, money to invest, know-how in tech, and — perhaps the most precious — time, or a longer runway to reinvent before they turn profit.
What can we do, the mortals, with very limited resources, a shorter runway, and no access to the secret sauce of Alibaba or Amazon? Is that even possible to have a growth mindset without all the means of the two Posts?
Interestingly, with all their might, the Post’s management struggles, too. The story of the new shiny newsroom has a dark side that the CEO shares publicly. People actually hated leaving their desks that they had sit at for 10 to 15 years, and they complained about the newly introduced flex seating. How to think about the growth where you no longer can find your coffee mug?
In fact, Liu found that people were resistant to whatever change in the Post he wished to see. “I heard more than once: I’ve seen it before; it didn’t work before; wake me up when it’s over,” Liu said in an interview about the early struggles he has faced. “At the end of the day, I can’t place fault on anyone who has that perception because I think from their perspective, there have been multiple false starts, not just at the Post but probably across the industry.”
This way the South China Morning Post’s CEO learned that transparency, communication, and trust between the newsroom and the business, and not just money or tech, were the primary means for the digital transformation he imagined.
Share your story: What do you struggle most within your business’ transformation project: the scope, resources, time, culture, or all of it? What would you do if a tech billionaire bought your news organisation? E-mail me at email@example.com.
- Alan Soon, “18 months in, the SCMP is on a very different footing. This is what CEO Gary Liu did to turn things around,” The Splice, July 3, 2018.
- Carol Dweck, “What Having a ‘Growth Mindset’ Actually Means,” Harvard Business Review, January 13, 2016.
- David Rogers, “Customers, competition—not technology—should lead digital transformation,” INMA World Congress Blog, June 4, 2018.
4. WHAT AM I READING? Try “Driving Digital Strategy” by Sunil Gupta and “How to Succeed in the Relationship Economy” by Matt Lindsay.
Who should get a discount? “Driving Digital Strategy,” a new book by Harvard professor Sunil Gupta, provides many insights about digital strategy. But I would like to point you to his evidence-based insights about the tactics. Companies in many subscription industries, be it newspapers or cable TV, offer huge discounts to attract new customers. Digital news subscriptions are no exception.
Professor Gupta found the discounted offers may be effective in the short term, but they are also costly in the long term. They have a negative impact on the perception of the brand and on lifetime value of the acquired customer. He suggests focusing on referrals from existing customers, and on word-of-mouth instead, as these tactics are proven to attract more valuable and more loyal customers.
The gift of a greeting. Churn is an elephant in the room at many publishing houses. Marketers focus on acquiring new subscribers, but their efforts are wasted when the existing ones leave. Luckily, publishers actually can identify the users who are most likely to stop their subscription before it happens and take care of them. Matt Lindsay of Mather Economics, one of the authors of a great book on subscriptions, “How to Succeed in the Relationship Economy,” dived deeply into the incentives that publishers grant to customers to save their accounts. Surprisingly, he found inexpensive incentives such a greeting card may be much more effective than expensive gift cards. It really works like in our private relationships, doesn’t it?
About this newsletter
Today’s newsletter is the first step in a year-long reader revenue and media subscriptions initiative by INMA, outlined here.
We are sharing today’s newsletter with all INMA members. To continue receiving this newsletter, click here and opt in for the next 12 editions.
This newsletter was written by Grzegorz (Greg) Piechota, Researcher-In-Residence at INMA, based in Oxford, England. Greg is a digital strategy expert, an author and a fellow at the University of Oxford and Harvard University who studies digital disruption and transformation at news media companies worldwide.
E-mail Greg at firstname.lastname@example.org with thoughts, suggestions, and questions.