Welcome to the second edition of Readers First, a new monthly newsletter for INMA members on reader revenue innovation where I will share results of my original research, notes from visits to digital subscription leaders, reflections on talks at conferences, and my favourite readings. 

Meet other newsletter readers for an informal chat on subscriptions. Our first online meet-up is up on Friday, October 19, at 10:00 a.m. New York time/3:00 p.m. Oxford time. Register here.

SUBSCRIBE TO THIS NEWSLETTER

1. ALERT: 60% of visitors to news Web sites don’t come back the next week, so how can we hope they ever become subscribers?

Individual news Web sites may seem to have similar layout or content, but in fact they enjoy different patterns of traffic. For example:

  • The share of page views from direct visits can vary from 0% to 97%.
  • The average share though, among 471 news Web sites that I studied in detail, is just 19%.
  • The top 10% best performing sites enjoy 42%.

Some of the world’s leading news brands enjoy up to half of their traffic from social media. The average share in my sample though is 11% and the median is 8%. These numbers don’t include so called “dark social” or traffic to article pages without referrals that most likely comes from links shared with messengers and e-mails. 

Why it matters: Frequency of online visits and volume of pages viewed are the strongest signals of users’ likelihood to subscribe, as my other study — with Cxense — shows.

Oh, really? Show me the DATA.

Define “loyal visits” please! These are generated by users who visited the news Web site eight times out of the past 16 days, or every other day and more. On average, the sites in my sample had nearly a third, or 29%, of their page views generated with loyal visits. Theh top 10% best performing sites enjoyed 64% share.

Anything else? Yes, I was surprised not to find any relationship between size of the news site or scope of its content and loyalty of the users. I thought that bigger newsrooms, publishing more articles, would have it easier — but it seems they don’t. I thought the local news sites would in general be more engaging than those covering far-away capitals — but they are not.

In fact, I found:

  • The size of the news Web site’s traffic does not seem to be linked with loyal behaviours. Bigger Web sites, although more resourceful, happen to fail in attracting direct visitors. And smaller Web sites, despite all the challenges, happen to succeed. No fate here, just a hard work.
  • Scope of a news publication — local, national, international — does not seem to be linked with loyalty either. Some local news Web sites attract many more loyal visits than others, similar to national news and international news sites. The distribution of the data is normal.

Take-aways: Having found that usage drives purchase, we see usage is mostly driven by habitual and intentional visits directly to news home pages and apps. Visitors from social media are usually fly-bys: visit just article pages, bounce and don’t come back.

How do you drive direct visits to your site and app? Exchange ideas with other INMA members who read this newsletter: grzegorz.piechota@inma.org.

Inspired to explore the value of loyalty? 

Want to chat more about loyalty? Let’s meet online on Friday, October 19, at 10:00 a.m. New York time. Register now.

 

2. INMA CLINIC. Paywall check-up with Dr. Greg: What is the best paywall model?

After last month’s newsletter, I got a call from Riske Betten, digital manager at Amsterdam’s De Telegraaf. “What’s the best paywall model?” she asked. It is a great question. The better, though, is not what is best but when?

  • What does a paywall do? Primarily, in my opinion, it is a segmentation tool: It filters out readers with the highest likelihood to subscribe so we can better target our marketing efforts. Some paywalls segment readers by behaviours (e.g., a meter), others by interests (e.g., a freemium), or attitudes (e.g., memberships). The latest models are hybrids that combine multiple attributes. The most sophisticated paywalls digest hundreds of attributes to target the segments of one and each user.

  • So what paywall is the best? Simply put, the one that is predictive of sales and filters out large, profitable, stable, reachable segments for our marketing. Effectiveness of marketing surely depends on the quality of segmentation, but it depends even more on a product itself, a price, a user experience, a promotion — the whole mix.

  • Across the world, publishers switch from one model to another. In my opinion, this is not a sign that one paywall is significantly better than the other. It usually shows that a targeted segment of readers has been unresponsive or exhausted and the publisher needs to find another one. Every publisher hits this wall. In marketing, we call it a chasm — like this deep, narrow hole in the ice that a traveler in a snow desert needs to cross.

    • Here’s a great example from Norway on crossing the chasm: Aftenposten launched an online paywall with a meter in 2013, and after a little more than a year it approached the chasm — the sales stopped to grow. In the next three years, it has added premium articles available only to subscribers and tweaked the meter for non-premium articles several times. The reality: No paywall lasts forever untouched!

Walk me through the Aftenposten’s decision to switch from a meter to a hybrid — STEP BY STEP.

  • The launch of a paywall is not the end of a subscription project but its start. Digital market leaders such as Aftenposten closely analyse the data and relentlessly optimise their product, content, and marketing. Over time, they target different segments — be it fans of politics before the elections or young job seekers. They experiment by closing different content categories such as magazine features, culture, investigative stories, etc.  It is the process of testing and learning that brought Aftenposten to 108,000 digital-only subscribers by spring this year — not one particular paywall model. 
  • There is no one paywall that fits all publishers at all stages of their development. Many abandoned the early models, and some switched several times such as Poland’s Gazeta Wyborcza. It started with a freemium in 2012, switched to a meter in 2014, and moved on to a hybrid in 2016. By the end of last year, Wyborcza enjoyed 133,000 digital-only subscribers. As a proverb says, what brought you here won’t get you there.

  • My evidence-based advice: Start with a meter, then tighten it or switch to a premium model. If predictiveness of sales is any guide, based on my original research with Cxense, I would advise a rational publisher to start with segmentation based on the most predictive variables such as usage. And then, after the most profitable segments are exhausted, she would switch to less predictive variables such as interests. Another route, if resources are not scarce, is to jump directly to a hybrid or a dynamic paywall that builds segments based on multiple variables.

  • Remember: All models can be good enough for you — be it a meter, a freemium, or even a hard paywall. Since 2010, The Times of London has famously run one of the strictest paywalls in the world, and it has just announced it had 255,000 digital-only subscribers at the end of June 2018. It is the superb marketing mix and skill that made the company successful — not the sophistication of their paywall.

Tell me about your “chasm” moment: How quickly did you reach it after the paywall’s launch? What did you do to cross it? Do you have a question you’d like me to address in the next INMA clinic? Talk to me: grzegorz.piechota@inma.org.

Dive deep or surf far away:

Want to chat more about paywall models? Let’s meet online on Friday, October 19, at 10:00 a.m. New York time. Register now. 

 

3. LETTER FROM THE ROAD: What would you do if your editor was a robot?

Before a reporter writes another article about her city’s traffic, she can ask a computer whether the effort is worth it.

If that sounds like fiction, it is because it reminds you the Douglas Adams’ classic The Hitchhiker’s Guide to the Galaxy and asking the Deep Thought for the answer to life, the universe, and everything. And yet, for a reporter on a deadline, an idea for an article may be the Ultimate Question.

The answer the reporter gets though may be no less puzzling than in a book: 42. What does it mean? Is that enough to hit the home page or bury her story?

I am afraid this no longer is a matter of the future. A number of publishers have shown during INMA Media Innovation Week in Amsterdam in late September that they had started to use machine learning algorithms to analyse its content to understand better which articles engage readers and how. The output is often a score, a number like “42.”

There is just a step for this technology from serving as an advisory tool for editors to automating editorial tasks: selecting stories to home pages or newsletters and predicting engagement for stories that have not been written yet. Like walking on a moon, it would be a small step for a tech passionate but perhaps a huge leap for the mankind in newsrooms.

Robots excel in translating numbers into dull reports, saving reporters’ time but not replacing their jobs. Editors? They are endangered species.

This is part of a broader trend to rely on algorithms to curate media content for engagement. The news publishers’ struggles with Facebook have shown that algorithms can outcompete human editors in selecting stories to news feeds if engagement is the metric. At Netflix, algorithms have already changed the ideation and production of new shows.

Having listened to discussions in Amsterdam, a number of future use cases came to my mind: Based on user segmentation by interests, the algorithm can easily predict which topics are more likely to convert visitors to subscribers or to engage those who subscribed. Knowing reporters’ beats and past performance, the robot can assign topics to the authors who are most likely to succeed in delivering the engaging stories. The machine can select the stories for the home page and then track performance, learning and relentlessly optimising. How far away are those?

Speakers in Amsterdam seemed to believe in data-informed journalism and not data-driven. The past, they repeated, tells you little about performance of new topics and story ideas. The newsroom still needs humans with “a nose for news.”

Meanwhile, the automation changes many industries. Algorithms have already been effectively managing drivers at Uber, matching them to riders and suggesting the best route. Amazon surrendered its warehouse workers to algorithmic bosses guiding and optimising the process of packing parcels. Now, Amazon has begun automating retail team jobs — forecasting demand, ordering inventory, and negotiating prices with vendors!

Studies find new robot managers inhumanly efficient but also exploitative. Over time, the workers suffer from the relentless pace and micro-management. They get frustrated when they don’t understand the robot’s decisions or feel treated unfairly. And they long for interactions with human managers for motivation, guidance, feedback, and conflict resolutions.

As we see, algorithmic decision-making cannot be optimised for revenue or reader engagement alone. The journalists’ and editors’ satisfaction should be measured and taken into the account, too. It’s not only fair, it is also smart — studies find people work actually harder when they are happy.

Is your boss a robot already? Have you automated jobs in the newsroom or in the marketing yourself? Share your story: grzegorz.piechota@inma.org.

Inspired readings:

Want to chat more about automation of editorial and marketing tasks? Let’s meet online on Friday, October 19, at 10:00 a.m. New York time. Register now. 

 

4. WHAT I AM READING? Try Smart Business by Ming Zeng and Subscribed by Tien Tzuo.

The future of business strategy: networks, data, and automation. Having explored the media and tech scene in China this September, I read a new book by Alibaba Group’s former chairman Ming Zeng, Smart Business. The book is smarter than an unimaginative title suggests. 

E-commerce giant Alibaba is often coined as “Amazon of China,” although Dr. Zeng would disagree: Alibaba is not a retailer in the traditional sense, as it doesn’t source or keep stock, and it outsources all logistics. It is more like Amazon, eBay, Google, PayPal, FedEx, and Citibank combined — it’s a giant marketplace that empowers all sorts of businesses with machine learning, mobile Internet, and cloud computing.

According to Dr. Zeng, a formula for Alibaba’s success in disrupting multiple industries is approaching them as networks of businesses and customers, and data intelligence. Data about interactions in the network is the primary asset, as it feeds machine learning algorithms that automate business decisions for all. Automation allows the seamless coordination of the giant network — and lets it scale.

The book discusses algorithms and APIs, but it is a business strategy book, not a technology primer. Dr. Zeng presents case studies from China that you might have never heard about but I’m sure you’ll find inspiring. One remark: This is not a critical account of the Alibaba’s rise nor a biography of its founder (for those you might wish to check Duncan Clark’s “Alibaba: The House That Jack Ma Built”).

From product- to customer-centricity. Product cultures are built around assembly lines: Stay in your lane, do your job, then pass it on to the next person — produce, assemble, distribute, sell. That does not work in subscription business, explains Tien Tzuo, CEO of Zuora, a subscription management software company, in his new book Subscribed, for which I attended the London premiere last week.

The objective in subscription businesses is about making sure the customer continues to succeed with your service over time and translating that ongoing value into revenue. According to the book: “The same organisational structures that worked in the past are now hindering the future. Organisation is the biggest thing holding us back — just like in the horror movie, the scare phone call is coming from inside the house.”

As a subscription professional, you may find the Tien Tzuo’s book a bit rudimentary for yourself, as it surveys different industries adopting recurring payments for half of the book and lays out the fundamentals of the business model in the other half. You may though find it a useful textbook for your colleagues, your board, shareholders, or business partners, whom you may wish to familiarise with a big shift across industries from product to services, from ownership to usage, and from single to recurring payments. 

And now we have come to an end of my second newsletter. Thanks for reading! I would love to learn what you think about this newsletter. What do you like? What would you change? E-mail me at grzegorz.piechota@inma.org.

See you next month in Miami at the INMA Consumer Engagement Summit. I’ll be there from Wednesday through Friday, November 7-9. DM me on Twitter or LinkedIn if you would like to meet me.

This newsletter is a public face of a year-long reader revenue and media subscriptions initiative by INMA, outlined here.