Does a preprint hurt my chances of being published?

Most major journals in the life sciences and physics accept papers that were preprinted, and their preprint policies are listed on their sites. Check your target journal before posting, but preprinting and publishing are usually compatible.

Should I preprint if the paper isn't strong?

Posting will not rescue a weak paper. The citation advantage in these studies is tangled up with the fact that people preprint their better work. Fix the paper first. Then make it findable.

Is this just SEO for scientists?

Partly. The mechanics overlap with SEO, but the target now includes AI answer engines, which quote and summarize. That raises the payoff for a clear title, a quotable abstract, and clean metadata.

Your preprint is live. That is not the same as being read

By Boris Gorelik · August 7, 2026

TL;DR: Papers with a preprint do collect more citations and more online attention, roughly 36% more citations and 49% more attention in one large analysis. But those studies are observational, and stronger work tends to get preprinted in the first place, so posting by itself is not a citation machine. The thing that turns a deposited file into a read one is findability: a title and abstract that read like a quotable answer, correct metadata and identifiers, and a clear link path from your profile and your posts to the preprint.

Two numbers are worth a sticky note. In a study of 74,239 papers, the ones that had gone up as a bioRxiv preprint had, on average, 36% more citations and a 49% higher Altmetric Attention Score than the ones that had not. That is Fu and Hughey, “Releasing a preprint is associated with more attention and citations for the peer-reviewed article”, in eLife (2019), across 39 journals.

It is tempting to read that as a recipe: post the preprint, wait, collect the citations. That reading is wrong, and this whole post is about why. Posting a preprint is not the same as being read.

Do preprints really get more citations?

Yes, the association is real, and it is not one lonely dataset. On top of Fu and Hughey, Fraser and colleagues ran a separate large analysis in Quantitative Science Studies (2020) and also found a citation and attention advantage for bioRxiv preprints. In Fu and Hughey the effect held independent of subfield, number of authors, and even the journal’s access model and impact factor. So the number is not a fluke, and it is not just an artifact of where the paper eventually landed.

Does posting the preprint cause the extra citations?

This is where the sticky note gets a caveat written under it. Both studies are observational. Nobody ran the experiment where you take the same paper, flip a coin, and preprint one copy but not the other. So the 36% is a correlation, and correlations this size usually carry passengers.

The obvious passenger is self-selection. Authors do not preprint at random. People tend to post the work they are most confident about, the results they most want seen, the papers from labs that already have a mailing list and a following. The eLife peer reviews raise exactly this point: the advantage may reflect that preprinted papers are more interesting, or come from more visible groups, not that the act of posting did the lifting. Posting a weak paper does not summon citations.

Deposited, discoverable, and read are three different states

Here is the distinction the citation numbers quietly skip over. There are three things people collapse into one:

Deposited. The file exists at a stable URL with a DOI. Congratulations, you and the server know it’s there.
Discoverable. A stranger who isn’t searching your name can actually surface it, in Google Scholar, in a plain web search, in an AI answer.
Read. A human, or the model answering that human, opens it, uses it, and cites or links it.

Depositing gets you the first one and none of the others for free. Most of the “preprint advantage” you actually want lives in the jump from deposited to discoverable. I wrote about the repository version of this in Deposit it and they won’t come: a study of 217,589 records where a few got almost all the search traffic and a long tail got essentially none. Uploading put every one of them in the repository. It did not put them in front of anyone.

What actually closes the gap?

Findability, and it is more boring and more doable than it sounds. Three things carry most of it.

Write the title and abstract as a quotable answer. Not a clever label, not reviewer-facing hedging. If a reader, or an AI search tool, asks the question your paper answers, the first sentence of your abstract should read like the answer, in the words a stranger would actually type. Answer engines lift sentences that already sound like answers.

Get the metadata and identifiers right. A registered DOI, correct author names and ORCIDs, the right field tags, a version that links cleanly to the published article once it appears. Broken or missing metadata is how a paper becomes invisible to the exact machines now doing the finding.

Build the link path. Discovery follows links. Point to the preprint from your profile pages, your lab site, and the posts where you talk about the work, so there’s an actual trail from where people already are to where the paper lives. A preprint with no inbound links is a house with no road to it.

Where Loud Camel fits

This is the part we work on. Loud Camel watches for places where your paper is the real answer to a question someone is asking right now, and drafts a response that points back to the preprint in plain language, with the link and the framing a searcher would use. You approve every draft. Nothing is sent on its own. The goal is narrow: shorten the distance from deposited to read, not to game a metric.

Posting the preprint is the floor. Whether anyone reads it is a separate job, and that job starts the moment the file is live, not the moment it gets cited. I could be wrong about how much findability moves the needle for any single paper. I’m fairly sure that leaving it to chance isn’t a plan.