Abstract peer review without the spreadsheets
Most scientific conferences still run peer review out of a shared spreadsheet and a mailbox. Here is what breaks, and what a proper review workflow looks like.
Every scientific conference reaches the same week. The call for papers has closed, four hundred abstracts have landed, and someone is about to open a spreadsheet.
Where the spreadsheet breaks
It rarely fails immediately. It fails at three specific points, and always in the same order.
1. Reviewer assignment
Matching abstracts to reviewers by topic is a bipartite matching problem with a workload constraint. Done by hand it produces two reviewers with sixty abstracts each and eleven with three, and nobody notices until the deadline.
2. Conflicts of interest
A reviewer from the submitting institution has to be excluded, and in a spreadsheet that exclusion lives in somebody's memory. The first time it is missed, the result is not a scheduling problem — it is a credibility problem.
3. Version control
Authors revise. Reviewers score the version they were sent. Once two people hold two different files, the committee is adjudicating on evidence that no longer exists.
What a real workflow does instead
- Rule-based assignment. Match on declared topics, cap each reviewer's load, and exclude institutional conflicts automatically.
- A rubric, not a number. Score against named criteria so a 7 from one reviewer means what a 7 from another does.
- Blind by default, multi-round when needed. Names hidden until the round closes, with a second round for the borderline band only.
- One record per abstract. Submission, revisions, scores, comments and the final decision on a single page with an audit trail.
None of this is exotic. It is what a conference committee already does in its head — written down, so it survives four hundred abstracts and a deadline.
ORBIT Conference does exactly this, and the same record carries through to the programme, the speaker portal and the proceedings. See how conference management works in ORBIT.