Several years ago, I was a journalist. Most recently, a technology reporter at The Associated Press, where I worked for 13 years. For most of that time, I covered the media business: movie studios and recording companies that were making the transition from DVDs and CDs to streaming.

I was inspired by this revolution, though I couldn’t help but notice that newspapers were slow to turn the corner. I wanted to see if I could make a difference.

I got into the JSK Fellowship at Stanford, and met a whole range of brilliant professors, advisers, fellows and students. Under their generous guidance and with a big assist from folks in the industry, I found myself as product manager of digital subscriptions at Bay Area News Group (BANG) in the fall of 2018. (You can read more about that journey here, here, here, here and here.)

A year later, Google, one of the companies I had covered critically as a reporter, awarded BANG funding to pursue a premium subscription tier that included four pillars: ad-free website access, location-aware news recommendations, VIP treatment at live events, and privileged or prioritized commenting.

Thanks to the help of a former Google engineer (whom I met separately from the funding), our corporate development team and our paywall vendor, we coded up ad-free and launched it on our checkout page with little fanfare in January 2020.

The thinking was to separate this element from other premium features that could boost engagement on their own, and to find discrete pricing for a feature that would cost money to deliver.

Our line in the sand was $1 more than the regular price — below which we risked giving up too much in ad revenue to be worthwhile — but wanted to see if people were willing to pay more.

We ran multiple tests to determine the best price for ad-free that maximized revenue, and it turned out to be $4 per month — a 28-33% increase over standard depending on which website you’re on.

About a fifth of new digital-only subscribers chooses to pay more for ad-free, many with no discounted trial, and many have upgraded their existing plans. Email campaigns to existing subscribers are seeing strong uptake and we have thousands of ad-free subscribers now.

Engagement is up amongst this group, consuming 1.2 pages more for every day they show up on the website than other subscribers. So far, retention — how long they keep paying after the trial period — is similar to other subscribers.

Let me say here that it’s relatively hard to raise prices, because the volume of starts goes down. Just like when we offer discounts, the volume of starts goes up. If done incorrectly, this can be a zero-sum game. But by offering something of unique value for a higher price, and making sure we cover our costs, we are setting up a cycle of product development that should help us succeed in the years ahead. It has a clear connection to supporting the business and paying for the work of journalists in the field.

Looking at the big picture, if one of every five people paid us 33% more, over the entire digital subscriber base, it would boost reader revenue by 7%.

Down the line the plan is to begin sunsetting our policy of allowing subscribers to use third-party ad-blockers.

Meanwhile, the lost ad revenue from premium subscribers is way below the increase in revenue. And we’ve already recouped the small amount we spent on development.

Of course, this story isn’t finished. To launch the minimum viable product, we made only the website ad-free. But there are many other places where consumers interact with our journalism: the mobile app, the digital replica e-Edition, and Google’s Accelerated Mobile Pages. All these platforms still have digital ads on them, and some paying ad-free subscribers have complained. We need to address this gap with more than just fine print terms and conditions. We also need to scale our solution to 30-some other websites in the chain, many of which have slightly different layouts. A lot of work remains. But essentially, we’ve got product-market fit.

So now let’s move onto the other pillars of the premium tier that we felt were worth exploring.

Location, Location, Location

A week after we launched ad-free subscriptions on MercuryNews.com, we launched a search function that allowed people to look for news happening near them or any location they chose. Our aim was to showcase the breadth of our coverage and double down on our most valuable and defining product: local news.

At launch, we hid this functionality beneath the “Local” menu tab, because in preparation for this, we learned that people who clicked this menu item were about 100X more likely to subscribe than others.

For half the audience, clicking the menu item would pop up a dialogue asking them for permission to access their location and a widget would emerge on the right side. If they gave permission, the widget would fill with stories from nearest to farthest away, otherwise they would have to enter a location.

The widget also allowed people to sign up for a personalized, location-aware newsletter that would send nearby story recommendations to their inboxes daily. The widget also directed them to a map where they could scan around the entire Bay Area and click on stories of interest.

This came after months of reporters and editors tagging stories in a collaboration with Bloom Labs, which created the tagging tool and visualization elements in WordPress.

There were several problems. We weren’t getting enough clicks to make this a sustainable feature. Just a quarter of people ended up sharing their location on the website when asked. And relying on an open source geolocation algorithm meant that even when they did share, it could take about 7 seconds to populate the widget with clickable stories, an eternity on the web. One user complained that we were interfering with his use of the “Local” tab, which uses neighborhood names instead of our new latitude-longitude coordinates. The hundreds of thousands of people exposed to this feature had resulted in just a handful who clicked.

With the “fail fast” motto in mind, we pivoted hard.

We decided to infer the user’s location using their IP address (even if this is incredibly fuzzy). Given the volume of tagged stories and their geographic spread, a visitor’s selection of recommended stories wasn’t likely to change much. We then served nearest-to-farthest recommendations by taking over the “Most Popular” section on home and article pages. We called the widget “News Nearby” and put a “miles away” label below each headline. Clicks went from a handful to the thousands.

This wasn’t the end of the tweaks. Some results were patently absurd, like the “nearby” Grammys story 479 miles away. So we limited its visibility to visitors in the San Francisco-Oakland-San Jose area and encouraged reporters to tag things nearer by.

We also loosened the constraints on the recommendations, allowing other cues to drive the rankings, with the distances out of order. Un-tagged stories showed up without a miles label. The jumbled approach actually performed better — to the point where the clicks on this widget now outperform a straight up “Most Popular” selection. We’ve now expanded this feature to East Bay Times and the Marin Independent Journal.

The “Local” menu item now just offers a link to the map page in the drop-down menu. We apply IP address-based location to new newsletter recipients or default them to downtown San Jose. Each newsletter comes with a link prompting the person to update their location if they’d like. We have more than 1,800 signups between two newsletters “News Nearby” and “Lifestyle News Nearby,” with open, read and click rates that are within striking distance of some of our top performers.