Facebook has changed the way people consume news, which has transformed the underlying dynamics of the publishing industry. Yet the publishing business model has stayed the same. While advertising demand is increasing rapidly, the number of publishers and ad inventory has increased even faster. Bottom line: the amount of money publishers can make from their site with a pure advertising model continues to drop year-over-year.

The average revenue per user has been in decline for the last several years, a frightening trend for publishers. With the ad-driven media business model under pressure, publishers must come up with ways to increase the average revenue per user (ARPU) on their properties.

So, what can we do to monetize more effectively? Implementing a paywall can help publishers connect with a loyal audience and increase ARPU, which is more important than ever, since most online traffic comes from anonymous social feeds where the audience is often unaware of your brand.

However, paywalls don’t have the best reputation. In the past, paywalls didn’t generate as much revenue as advertising did because display ads were more lucrative than they are today. Publishers feared shutting out the audiences who view the ads, and that fear persists, especially as programmatic CPMs continue to decline. Several high-profile paywall failures also helped fan the flames of discontent. The U.K.’s biggest newspaper The Sun, for instance, put up a paywall only to take it down a few months later.

In the past, clunky paywall technology also posed a problem. Publishers spent a lot of time and money trying to build these walls, but their paywall strategies were often naive. Some blocked all of their content and simply assumed people would turn up and open their wallets. The loss of ad revenue greatly outweighed the value of those that did. However, paywall techniques and strategies have come a long way since the early days of digital publishing and are worth reconsidering.

Publishers are right to be concerned that their traffic will plummet if they put up a paywall. It might. To combat declining traffic, publishers should put up metered paywalls tailored to different consumers based on known data. In the metered wall model, the site doesn’t wall off any content outright, instead limiting the amount of content an individual consumer can access.

More flexible paywalls can allow publishers to generate more revenue. You don’t want the paywall to go up before the reader has gotten a taste for the content, but appetite for content varies from person to person. Some people may want 10 free articles before they are willing to pay; some will pay after two.

Even with the limited data that publishers have on anonymous site visitors, publishers can still know if the visitor came in through Facebook or Google. They can use this data to optimize the paywall to the user. For instance, a Facebook visitor isn’t going to be very amenable to a premium subscription offer right out of the gate. This audience is probably using Facebook to entertain themselves in between meetings, so a strict paywall probably won’t convert them.

Contrast that with Google, where a consumer searched for something specific and found it on your site. In this instance, a publisher might find success with a paywall that goes up after only an article or two.

Publishers can also play with the pricing model. In traditional print media, the price is printed on the cover of the publication. But digital publishers aren’t required to have fixed pricing. Instead, they can optimize pricing per audience member or per audience segment and then dynamically adjust through testing. For example, a publisher might try to infer data about an anonymous reader to determine how much to charge. While pricing can be difficult to determine — most publishers don’t have data for half of their audience — publishers can predict likely propensity to pay based on an analysis of what their audience is already doing.

Location data can also be useful in establishing if a paywall is appropriate. For instance, a U.S. site might decide not to put up the paywall for a visitor from Argentina because the reader may not be as likely to subscribe, given they are outside of the target market. Instead, publishers can use this data to drop the paywall and simply monetize those readers with ads.

These various types of data can help publishers determine the cadence of how and when to use a paywall and what to charge. There is no one-size-fits all paywall model, so publishers must test in order to arrive at their optimal settings.

In order to sustain their business moving forward, digital publishers must look for ways to diversify. Paywalls may have their challenges, but they also provide ample opportunities for publishers to increase revenues.