THE Network: A Specialty-First Academic Network Built for Ten Right People

Roughly a dozen university partners have asked me some version of the same question this year, and it's never about salary. It's about rooms. Where do the people in my subfield actually talk to each other? Not the discipline — the subfield. A chair hunting for someone who works on computational phonology doesn't need a million profiles. She needs ten, and a way to reach them before the advert goes live.

THE Network is AcademicJobs.com's answer to that question, and the narrowness of it is the point. You name your specialty, join the rooms built around it, and get found by people who already understand what the words in your profile mean. No advertising. No sales pitches in your inbox. No feed designed to hold your attention rather than answer your question.

What the platform actually does

Three functions sit under the platform's "connect, communicate, collaborate" line. The first is a profile that names a specialty rather than a job title — coastal sediment transport rather than researcher — because a search that runs on job titles returns everyone and finds nobody. The second is discipline-specific rooms, so a marine biologist isn't wading through posts written for data analysts. The third is a consortium feature that lets members assemble a group around a shared question.

The smallest feature carries the most psychological weight. Members post milestones — a paper accepted, a difficult course delivered, a grant submitted — and colleagues tap "kudos", a mechanic lifted from the fitness app Strava. Gallup has long found that only about one in three US workers strongly agrees they received recognition or praise in the previous seven days, and that people who don't feel adequately recognised are twice as likely to say they'll quit within a year. For a postdoc who has spent two years hearing nothing back, a tapped kudos registers as somebody outside the group noticed.

AcademicJobs.com's founder, Jarrod Kanizay, built hiring infrastructure before he built a community: he launched the job board UniJobs in 2006, grew it across 40 countries, and sold it to TES Global, the parent of Times Higher Education, in 2015. Joining the network is deliberately short — a profile, a specialty, and a signal of the kinds of opportunity you want to see.

The problem a specialty room is trying to solve

Academia runs on closed circles. A journal editor knows who she calls for review. A conference chair knows who gets the keynote. Those invitations circulate among known names, and the people outside the circle are not less capable, only unlisted. A standing profile that names a specialty gives a researcher a way onto the list without needing an invitation first.

The labour market sharpens the case. The US Bureau of Labor Statistics projects postsecondary teaching jobs to grow about 8 percent from 2023 to 2033, with roughly 118,900 openings a year, and most of those are replacements rather than new seats. In a market that size, a search committee can still miss the strongest person in the discipline because that person was never looking.

Openness also varies sharply by country, which is why a network built for one national system rarely travels well. Data compiled by the BBVA Foundation and Ivie from SciVal-Scopus records put the share of publications with international co-authors in 2024 at 62.3 percent for the United Kingdom and 51.8 percent for Spain, against 38.2 percent for the United States, 24.3 percent for India and 18.1 percent for China. Volume tells a different story than percentage: the United States still leads the raw count, with China second.

More collaboration is not automatically better collaboration

Here is where I'd push back on any pitch that treats connection as a good in itself. The evidence is more interesting than that.

A 2019 paper in Nature by Lingfei Wu, Dashun Wang and James Evans analysed more than 65 million papers, patents and software products published between 1954 and 2014. Teams of fewer than five people tended to produce work that disrupted a field; larger teams tended to develop what already existed. The same individual was more disruptive in a small team than in a large one, which means the effect is about group structure rather than about who joins it. Both sizes are necessary. Neither is a virtue on its own.

Citation data points the same way with a blunter instrument. A 2026 study in the Journal of Informetrics by Hwiyoung Lee, Chelsea Zhang, Manuel González Canché, Yoosun Park and Ram Cnaan examined 417,201 English-language articles published in 2014 across eight disciplines. Solo-authored papers consistently drew fewer citations, with the gap between solo and multi-authored work sitting at 0.1 to 0.2 standard deviations. In the social sciences, the returns to co-authorship followed an inverted U: past a certain team size the benefit flattened, then fell away.

International co-authorship also has an awkward edge that the sales material usually skips. Mike Thelwall and colleagues, writing in the Journal of the Association for Information Science and Technology in 2024, matched 148,977 UK-based articles against their post-publication review scores from the 2021 Research Excellence Framework. International partners raised the odds of a higher quality score in 27 of 34 Units of Assessment. But collaborations with lower research-expenditure economies tended to be judged lower quality, which the authors suggest may reflect a failure to understand Global South research goals rather than a failure of the research. A network that only makes it easier to find the well-funded names will reproduce that pattern.

What conference research says about rooms

The nearest thing we have to evidence on rooms is the literature on conferences, and it is mixed in a useful way. In one study of Gordon Research Conferences in the biological sciences, Sen Chai and Richard Freeman found that attendees with no prior collaboration together produced about 9 percent more joint publications after attending than matched non-attendees, and that junior researchers gained the most while being the least likely to attend.

A natural experiment makes the mechanism clearer. When the 2012 American Political Science Association annual meeting was cancelled because of Hurricane Isaac, researchers who had planned to attend became about 16 percent less likely to form a collaboration with another potential attendee, with the strongest effect between people who were not already in the same city. Take the room away and the ties that would have formed mostly don't.

Then there is the case study that reads like the best-case version of what the network is attempting. Researchers studying the International Milk Genomics Consortium, publishing in PLOS ONE, tracked 311 attendees who published in the field. After a first symposium, 64.3 percent had more co-authors, 88.7 percent published at a higher rate, and 74.3 percent saw higher per-article impact. The meeting had no membership dues and no gatekeeping, which is precisely what made it a fair test of the meeting rather than of who could afford to attend.

The sceptics deserve a hearing too. An analysis of the DBLP computer science database, covering more than a million papers and over 640,000 authors, concluded that at best 4.61 percent of collaborations could have emerged from people meeting at a conference, and that conference environments do not significantly promote collaboration among participants. Location alone does little. Structure and follow-through do the work.

The early-career researcher carries the most weight here

Nature's 2023 survey of postdoctoral researchers drew 3,838 respondents from 93 countries. Sixty-one percent were working outside their country of origin. Fifty-five percent said they were satisfied with their current postdoc, five points lower than in 2020. One in four reported experiencing discrimination or harassment in a current or previous postdoc.

Those numbers describe a group for whom the department corridor is not a sufficient network. A postdoc who arrives in a new country on a two-year contract needs collaborators with complementary methods before the tenure clock starts, and needs them faster than a conference calendar allows. The same logic runs through our guide to thriving in a postdoctoral role, which treats collaboration as pipeline work rather than something that begins after the thesis is finished.

What it won't do

Two limits are worth stating before anyone signs up. Every niche network faces a cold start: a room is only useful if enough of the right people have walked into it, and whether the computational linguists you need are already there is a question only your own field can answer. The second limit is subtler. The network matches people; it does not manage them. A profile fixes your discoverability, not the internal dynamics of a team that has never agreed on who owns the data.

You can join at academicjobs.com/the-network, and it is worth reading the launch material with clear eyes: it concedes that the network won't find you a job on its own and won't make a search committee move any faster. That honesty is the right kind. A directory sold as a deliverable is where these things usually go wrong.

For hiring teams, the practical step costs almost nothing and can be done this month. Add one line to the candidate pack and the induction pack: the name of the specialty room your new hire should join, written in their own field's words, before they arrive on campus. The people who execute this are rarely the ones on the panel. They are the research office administrator who maintains the induction file, the department manager who chases the IT account, the faculty HR coordinator who notices the new starter is the only person in the building working on that question. Give them the line to add, and the room starts filling itself.

Photo by Christina @ wocintechchat.com M on Unsplash