Publishers are trying to get more readers to visit them without waiting for Google to send them there. That’s one of the main themes in Digiday’s September 28 report from its recent Publishing Summit. Executives described putting more work into newsletters, subscriptions and the experience on their own sites as search referrals become less dependable.
Reuters offered an interesting example. According to Digiday, its paywall uses machine learning to decide when to ask a reader to subscribe and when to keep earning through advertising. Reuters executive Phil Andraos said the ad business was doing better than before the paywall, helped by subscribers returning more often. That’s Reuters’ account of its results, not proof that the same approach would work for a smaller publication.
The report also describes publishers exploring income from data, licensing and other products. AI licensing is still unsettled. The New York Times’ Adam Greenberg said questions about value and control make deals difficult. I wouldn’t plan a publication’s future around money those agreements might eventually bring in.
This interests me because I’m building a publication here. I want people to find an article through search or a social post. But after they’ve read it, what would make them come back? That seems like a more useful question than how to get another brief burst of clicks.
If someone enjoys the Ravens coverage, they should be able to find the next football piece. If they come for AI, the article should explain something they were wondering about and show where the information came from. A familiar writer can give someone a reason to return, but that depends on the work being worth their time.
A newsletter could help readers keep up without checking the site every day. It still needs something worth opening. Adding a signup box doesn’t make that happen by itself, and having an email address isn’t permission to fill someone’s inbox with things they didn’t ask for.
There’s a business question here, too. A sponsor would need to know who reads the publication and why the placement is useful. A paying subscriber would need a clear reason to pay. Those are different offers, and neither becomes viable just because a site publishes a lot of articles.
I want to use AI where it helps with this work, including research and making the archive easier to find. I’d be careful about rebuilding everything around whichever assistant sends the next wave of visitors. We could end up depending on its decisions just as heavily as publishers have depended on search.
I’ll keep paying attention to how people discover the site. I also want to know which stories bring them back and what they want to read next. Those answers would help me decide where to spend more time writing.