circa news — when a good idea can't find a business model

The structural problem with digital news

In 2012, when Circa launched, the basic unit of digital news was still the article — a form inherited from print without modification. An article had a headline, a byline, a few hundred words, and a publication timestamp. It was written once, published once, and froze. When something new happened in an ongoing story — a new arrest in a criminal case, a policy update, a correction to a previously reported fact — journalists wrote new articles, often repeating 80% of the context for readers arriving late. The reader who had followed the story for weeks got the same recap as someone reading about it for the first time.

The inefficiency was systemic and visible. Anyone following the Brexit negotiations, a corporate fraud case, or a long-running political scandal was reading enormous amounts of repeated context in each new article. The information architecture of the internet had evolved to surface individual articles optimally — but the articles themselves were still designed as if they'd be read in print.

Circa's founding team — Matt Galligan, Ben Huh, and editor David Cohn — identified this as a structural product problem, not a presentation problem. The solution wasn't a better article format. It was a different fundamental unit.

The Decision

David Cohn called it "object-oriented journalism." The concept was borrowed directly from software engineering: instead of treating a news story as a document, treat it as a composition of atomic units — facts, statistics, quotes, images, and timestamps. Each atom was verified, tagged, reusable, and could belong to multiple stories simultaneously. When a story developed, the journalist updated the atoms. A user who had "followed" that story received only the delta — the new information — not a full re-read of context they'd already absorbed.

The product logic was genuine. In software, you don't rewrite a function every time you call it — you update the function and every caller benefits. Circa applied that principle to news: update the fact, and every story referencing that fact reflects the update. A statistic about unemployment that appears in ten different stories is one object, not ten copies.

By 2016, internet users were the primary news source for 73% of the US population. The mobile news market was enormous and growing. Circa's product had real technical novelty and real user benefit — faster reading, no redundant context, notifications that delivered only new information. The "follow a story" feature was a genuine innovation: subscribers could track complex ongoing stories and receive push notifications when something substantive changed, rather than manually checking for updates.

Circa launched in 2012. It shut down in 2015. Sinclair Broadcasting acquired it and relaunched it in 2016. It shut down again in 2017. The idea was sound. The business model was fatally mismatched.

The Failure

Circa's atomised format had a self-defeating property embedded in its design: time-on-site was lower. The product was efficient by design — users got only what was new and left. For a news product running on advertising, where revenue is tied to impressions and session depth, efficiency is not a virtue. The more successfully Circa delivered on its core promise — give users only what's new, eliminate redundant context — the worse the advertising economics became.

This is the structural tension: Circa's product philosophy optimised for user time. Advertising-based news products must optimise for advertiser time — meaning they need users on the platform long enough to see and engage with ads. These objectives are not compatible. A news product that efficiently surfaces only the new information will have shorter sessions than one that requires you to re-read context. Every design decision that made Circa better for users made it worse for advertisers.

The "follow a story" feature compounded this problem. When users followed a story, their engagement pattern shifted from browsing (visiting the app to explore news) to checking (visiting the app to see if the specific stories they followed had updates). Browsers generate more pageviews. Checkers generate fewer, more targeted interactions. A checking behaviour pattern is ideal for a subscription product or a B2B alert service; it's problematic for an ad-supported consumer product.

The failure was visible in the product architecture before a single line of code was written. Object-oriented journalism, by design, produces shorter sessions and more targeted engagement than article-based journalism. Any PM or business model analyst who mapped the product experience to the monetisation model would have identified the collision.

What Worked

The atomised content model was technically successful. Circa's newsroom built genuine competency in decomposing complex, evolving stories into reusable fact objects. The "follow a story" notification system worked as a product — users who received a Circa notification got genuinely new information, not a repackaged summary of what they already knew. The UX was clean and the product was fast.

The mobile-first approach was correct for the era. In 2012, most news publishers were still treating mobile as a secondary screen — a smaller version of the desktop experience. Circa built for mobile as the primary context, with notifications as the primary engagement mechanism. That instinct was right; the news industry spent the subsequent decade trying to rebuild for mobile after failing to build for it from the start.

What a PM Should Take From This

Circa is one of the clearest examples in product history of a genuine innovation that failed not because of execution but because the business model was incompatible with the product's design from the beginning. The product did what it promised. The promise didn't generate the economics required to sustain the business.

The lesson is not "build for monetisation, not for users." It's more specific: model what the product optimises for, then check whether the monetisation model requires a different optimisation. Circa optimised for reading efficiency. Advertising optimises for engagement depth. The mismatch was structural, not fixable by sales strategy or product iteration.

The thought experiment at the modelling stage — before building — would have looked like this: if users get only what's new, how long is an average session? If average sessions are short, what monetisation models work with short sessions? Subscriptions work with short sessions because they're paid upfront. B2B alert services work because they're sold on relevance, not volume. CPM advertising does not work with short sessions. Building a CPM-dependent revenue model on top of an efficiency-first product is not a go-to-market problem — it's an architecture problem.

The alternate history is plausible. Circa with a subscription model might have worked — a small number of users who valued genuinely efficient news consumption would pay for it, and their short sessions would be fine because the revenue doesn't depend on session length. Circa as a B2B news intelligence service — selling the atom feed and story-tracking infrastructure to corporate clients, law firms, or fund managers who needed precise, timely information on specific topics — might have worked, and actually was a direction Sinclair eventually attempted before giving up.

The anti-pattern this case exposes: building a product that solves a genuine problem without simultaneously designing the business model around the product's actual operating characteristics. The business model is not an afterthought to the product — it's a constraint that should be designed in from the beginning.