Trapit and the Content-Discovery Startup Cohort

Edited on Jun 23, 2026

Between 2009 and 2015, a whole generation of content-discovery startups tried to solve a single problem: too much information, too little time, and no good way for any individual reader to find the things they actually wanted to read. Trapit, Prismatic, News360, Zite, Pulse, Flipboard, Summly, Circa. Most of them are gone. The problem they tried to solve did not disappear. The companies did.

The case file matters because the failure pattern repeats. Every two or three years a new generation of content-discovery startups raises money on the same pitch. Most of them fail for the same reasons. Understanding the 2009-2015 cohort is the cheapest way to predict whether any future content-discovery bet has structural legs or is repeating the original mistakes.

The Trapit Story

Trapit launched in 2011 out of SRI International, the same research institution that produced Siri. The product used machine-learning techniques to surface articles based on user interest signals rather than explicit feeds. The user experience was strong. The early reviews were positive. The company raised funding from major venture firms and was widely regarded as one of the more sophisticated technical approaches to content discovery.

Trapit pivoted to enterprise content curation around 2014, focusing on B2B social-selling use cases. Salesforce acquired the company in 2015 for an undisclosed amount and integrated the technology into the Sales Cloud platform. The consumer product was retired. The B2B integration ran for several years before being deprioritized inside the broader Salesforce content suite. Trapit no longer exists as a standalone brand.

The Cohort

Prismatic — Launched 2012 with a strong machine-learning team and Y Combinator backing. Beautiful product, sophisticated personalization. Shut down December 2015 citing inability to compete with platform-based content distribution.

News360 — Launched 2010, raised $14 million, focused on news aggregation with personalization. Acquired by an Australian media company in 2019. Still operates in reduced form.

Zite — Launched 2010, acquired by CNN in 2011 for $20 million, sold to Flipboard in 2015 for $60 million. Shut down by Flipboard after acquisition. Content and team folded into the parent.

Pulse — Launched 2010, acquired by LinkedIn in 2013 for $90 million. Rebranded as LinkedIn Pulse, eventually folded into LinkedIn's broader content distribution. The brand no longer exists.

Flipboard — Launched 2010, the most enduring of the cohort. Still operating in 2026 with a refocus on creator-driven content and federated social via the fediverse. Survived but did not produce the venture returns the early investors expected.

Summly — Launched by then-15-year-old Nick D'Aloisio in 2011, acquired by Yahoo in 2013 for $30 million, shut down within a year of acquisition. The technology was reportedly licensed back to SRI.

Circa — Launched 2012, raised $5.7 million for “atomic” news writing. Shut down in 2015 citing inability to find a sustainable business model. The team was hired by Sinclair Broadcasting.

Why the Cohort Failed

The 2009-2015 content-discovery startups failed for four overlapping reasons, and the reasons are worth knowing because they recur.

One. Platform distribution beat aggregation. Facebook's News Feed, Twitter's Home Timeline, and the early algorithmic surfaces inside Google's ecosystem captured the consumer-attention layer before the standalone aggregators could scale. Once the platforms made discovery a free byproduct of being on the platform, paying users to install a separate discovery app became a much harder sell.

Two. The monetization gap was structural. Content-discovery products generate user attention but not enough first-party data or commerce intent to support direct advertising at scale. The cohort tried to monetize through paid subscriptions, sponsored content, native ads, and B2B pivots. None of the approaches produced the unit economics the venture-scale investors required.

Three. The publisher relationships were brittle. Most of the cohort depended on RSS feeds, publisher partnerships, or scraping arrangements that publishers could revoke at any time. As publishers built their own apps and consolidated traffic on their own properties, the licensing layer that the discovery startups depended on contracted.

Four. The user behavior was social, not algorithmic. The 2009-2015 hypothesis was that machine-learning recommendation would beat human curation at scale. The 2026 reality is that the dominant content discovery surfaces — TikTok For You, Instagram Explore, Twitter/X For You, YouTube Recommended — are algorithmic, but they are algorithmic inside social platforms with massive first-party engagement data. Standalone algorithmic discovery never had that data and could not produce comparable recommendation quality.

What Replaced the Category

The category did not die. It migrated. Content discovery in 2026 lives inside the major platforms — TikTok For You is the most successful algorithmic discovery surface ever built, Instagram Explore handles visual-content discovery, YouTube Recommended drives more video discovery than any other product, and X's For You algorithm runs the closest thing to real-time text content discovery. Each of these is doing what Trapit, Prismatic, and the others were trying to do — but inside a platform with enough user engagement data to make the recommendations actually work.

The lesson generalizes. Standalone content-discovery apps face structural disadvantages against in-platform algorithmic discovery, and the structural disadvantage compounds with every quarter the platforms accumulate more engagement data. Future content-discovery bets either need to be platform-native (already inside a major social or media surface) or need to identify a discovery problem that the major platforms structurally cannot solve.