↖︎ Vishal Singh
Data Stories · Media & Agenda-Setting

Nine Thousand Bylines

HuffPost was built on free labor: an open platform where anyone — senators, celebrities, your neighbor — could blog. The author fields on 190,000 headlines record that machine at full throttle in 2014, and then record it being switched off.

Author
Vishal Singh
NYU Stern School of Business
Published
July 2026
Data
News Category Dataset (R. Misra)
Author fields · 2012 – 2017 dense window
Who this data represents Author fields on headlines published on HuffPost.com, 2012–2017 (the dense-archive window; the archive turns sparse after May 2018, so this article makes no claims beyond 2017). Bylines measure credited output, not employment or pay — most of the names in this data were never paid.
9,003
distinct bylines in 2014 alone — the open-platform high-water mark
19% → 42%
share of bylined output written by the top 50 authors, 2014 → 2016
72% → 96%
share of headlines carrying any byline at all, 2012 → 2017
1,789
headlines by the single most prolific writer, 2015–2017 — about two a day

Every era of digital media has a signature labor model. HuffPost's founding bet was the unpaid contributor platform: open the CMS to the crowd, harvest the content and the search traffic, pay in exposure. At its peak the model was extraordinary — in 2014, more than nine thousand distinct names appeared in the site's bylines, most of them writing a handful of posts and vanishing.

The author fields in the archive let us watch that bet being unwound, quarter by quarter, well before the company's formal announcement. In January 2018 HuffPost publicly shut the contributor platform. The data says the retreat began three years earlier.

Figure 1 · The contributor machine, opening and closing
Distinct bylines per year (bars) and share of headlines carrying any byline (line), 2012 – 2017
The peak is 2014; the retreat starts immediately after. Distinct bylines fall from 9,003 to under 5,000 by 2016 while total output holds steady near 30K headlines a year — the same newsroom, produced by half as many names. Meanwhile anonymous/wire output nearly disappears: by 2017, 96% of headlines carry a name.

The re-concentration

Fewer names writing the same volume means concentration, and the concentration is dramatic. In 2014 — the platform's most open year — the fifty most prolific authors accounted for just 19% of bylined output; the long tail of contributors wrote the site. By 2016 the top fifty wrote 42%, and the number holds in 2017. In two years, HuffPost quietly transformed from a crowd platform with a newsroom attached into a newsroom with a shrinking crowd attached.

Figure 2 · Share of output written by the top 50 authors
Percent of bylined headlines each year, 2012 – 2017
The inflection lands in 2015–2016 — the same years the politics desk took over the front page. Professionalization and politicization arrived together: election-cycle news is staff work, not contributor work.

Two headlines a day

Who is the new machine? The staff core. The most prolific byline in the window produced 1,789 headlines in three years — roughly two every day, weekends included — and the rest of the top ten run at similar industrial cadence. This is the aggregation desk: trending-news rewrites, celebrity items, viral embeds, each a few hundred words, filed relentlessly. The crowd of nine thousand was replaced by a dozen people with very fast keyboards.

Figure 3 · The industrial core
Most prolific bylines, total headlines 2012 – 2017 · hover for active years
“Reuters, Reuters” is the wire feed, credited like an author — a reminder that byline data encodes CMS conventions as well as people. Every other name on the list is (or was) a HuffPost staff writer or editor.

The archive can't say whether the trade was wise — it records output, not revenue or trust. But it captures, with unusual resolution, a structural fact about digital media in the 2010s: the open-platform era didn't end with an announcement in 2018. It ended in the byline fields, in 2015, one departing contributor at a time.

Data & method

Source: News Category Dataset v3 (Rishabh Misra), HuffPost headlines with an `authors` field, 2012–2017 (190,270 headlines, 82–96% bylined depending on year). Distinct bylines count exact author-string values per year; multi-author strings (e.g. “A, B”) count as one byline. Top-50 share is computed within bylined output per year.

Caveats

The archive is dense only through May 2018 — nothing here extends past 2017. A byline is a CMS field: wire feeds (“Reuters, Reuters”) and package bylines appear alongside people, and contributors who posted without bylines are invisible. Counts measure credited output, not headcount, employment, or compensation (contributors were unpaid — that was the model).

Reuse & citation

Data: CC BY 4.0 — attribute Rishabh Misra, arXiv:2209.11429. Headlines © HuffPost. Article text and figures: CC BY 4.0.

Singh, V. (2026). “Nine Thousand Bylines.” vishalsingh.org Data Stories. Data: News Category Dataset (Misra 2022, arXiv:2209.11429), CC BY 4.0.