Arc was supposed to be a key to The Washington Post’s future. It became a problem instead.

Several months ago, Brian Stelter wrote an article (gift link) for The Atlantic exploring how The Washington Post had lost its way. During the Trump years, the Post thrived under the ownership of Amazon founder Jeff Bezos, adding audience and staff as well as turning a profit. Since then, all three of those metrics have nose-dived. Bezos’ choice to turn things around, publisher Will Lewis, is beset by ethical problems that no one seems to want to deal with.

All those issues are explored in detail by Stelter, but there was one fact that stood out to me: The Post’s content-management system, Arc, which was supposed to be a money-maker, had instead turned out to be a drag on the bottom line. Stelter wrote:

In 2021, the Post’s total profit was about $60 million. In 2022, the paper began to dip into the red. [Then-publisher Fred] Ryan reassured people that the loss was expected because of the investments in the Post’s journalism and continued losses at Arc XP, the in-house content-management system that the Post expanded during Bezos’s and Ryan’s tenure (the software is now licensed to other companies). Arc needed to spend a lot of money to have a chance to make money in the future, the argument went, and according to two sources, it accounted for the majority of the Post’s losses in 2022 and 2023.

If Ryan was right, then there was nothing wrong with the Post that getting Arc under control wouldn’t fix. I was surprised, and I filed that factoid away for future use. Well, the future arrived this week, as the Post announced it was laying off about 25% of Arc’s staff — more than 50 people — in order to stem those losses.

What happened? Stories about the layoffs in The Wall Street Journal (gift link) and Axios don’t really make it clear. But it seems that what at one time had looked like a smart bet on the future went south in a serious way.

CMS’s are universally loathed, but Arc was billed as something different and better — simple and built in a modular manner to made it easier to add features. It’s fast. To this day, the Post’s mobile apps load much more quickly than The New York Times’. The Boston Globe is an Arc customer, and if you use its Arc-based apps (look for a white “B” against a black background), content loads more or less instantly.

When I was reporting on the Post for my 2018 book “The Return of the Moguls,” then-chief technologist Shailesh Prakash touted Arc as a key to the Post’s future success. Internally, the Post’s iteration of Arc featured the infamous “MartyBot” — an image of then-executive editor Marty Baron that popped up on a journalist’s screen as a reminder that a deadline was approaching. One of Arc’s customers was Mark Zusman, the editor and publisher of Willamette Week in Oregon. He told me by email:

They flew a team out here and within three months we were up and running. I was pleasantly surprised with how quickly it happened. Arc creates enormous functionality under the hood. I have a happy news team (talk about unusual) and the Post is rolling out improvements on a regular basis.

Prakash told me that he hoped Arc might help the Post become the hub of a news ecosystem that would benefit both the Post and news organizations that licensed the CMS:

I would love it if the platform we built for the Post was powering a lot of other media organizations. That would definitely break down the silos for content sharing, a lot of the silos for analytics, for personalization. The larger the scale the better you can do in some of those scenarios. But those are still aspirational at this point.

Well, Prakash is long gone, and is now vice president of news at Google. Baron has retired. And Arc has failed to deliver on its promise of becoming a revenue-generator for the Post as well as a way for the paper to establish itself as the center of a network of Arc-using news organizations.

I hope we find out what happened. I know that Arc is expensive — probably too expensive for it to be adopted by more than a handful of news clients. Still Axios reports that the CMS has more than 2,500 customers. Maybe the layoffs will allow for a reset that will lead to future growth. But the story of Arc sounds like one of opportunity that slipped away.