Most papers that get rejected are not rejected for weak ideas—they are rejected for weak execution. These seven tactics address the specific failure modes reviewers actually find, and you can apply them in the final week before submission without rewriting the paper.
The gap between a paper that gets accepted and a paper that gets rejected is rarely the underlying research. It is almost always the manuscript: how clearly the contribution is stated, how well the evidence supports the claims, how carefully the related work is positioned, how legible the figures are. The good news is that these are fixable.
Here are seven concrete ways to strengthen your manuscript in the final stretch before submission.
1. State your contribution as an action, not an adjective
The most common reason a reviewer loses confidence is a contribution they cannot identify. “We propose an efficient method” tells them nothing. “We reduce inference latency by 40% on benchmark X relative to the strongest published baseline” tells them exactly what is new and why it matters.
Go through your abstract and introduction. Every contribution should be a verb phrase: “we derive,” “we measure,” “we show,” “we prove,” “we validate.” Adjectives like “novel,” “efficient,” “robust,” and “comprehensive” are not contributions—they are adjectives. Cut them where you can; replace them with measurements where you cannot.
2. Map every claim to its evidence
Reviewers read results sections looking for the evidence behind each claim. Make this easy for them. For each claim in your paper, there should be a result, table, figure, or statistical test that supports it—and the connection should be explicit.
A simple test: print the claims section and the results section side by side. Draw a line from each claim to its supporting evidence. Any claim with no line is unsupported. Any line that crosses three paragraphs is a clarity problem. This is tedious and it works.
3. Cut the related work that does no work
Related Work sections accumulate citations over months of writing, and many of them become decorative—citations that fill space without advancing the argument. Reviewers can tell.
Go through every citation in Related Work and ask: what work is this citation doing? If the answer is “it shows I read the literature,” cut it. Keep only citations that (1) establish a gap your paper fills, (2) define a term or method you use, (3) provide a baseline you compare against, or (4) acknowledge a direct predecessor. Everything else is noise.
Use Google Scholar and Semantic Scholar to verify you are not missing the obvious direct predecessors, and a reference manager like Zotero to clean up the bibliography.
4. Make every figure standalone
Reviewers often skim figures before they read the body text. If your figures only make sense after reading three paragraphs of explanation, you have a problem.
For each figure, check:
- Axis labels with units
- Legend entries that match the body text terminology
- Captions that state the figure’s main point, not just describe what is plotted
- Legibility at the size the figure will appear in print, not on your monitor
If a reader skims only your figures, they should still understand what you found. Design figures for the final PDF format from the start.
5. Add a limitations section if you do not have one
Many authors omit a limitations section, fearing it weakens the paper. The opposite is true. Reviewers reward honesty about scope, and a paper that acknowledges its own limits reads as more trustworthy than one that pretends to have none.
A good limitations section states: what your method does not handle, what assumptions it depends on, and where the results might not generalize. Two paragraphs is usually enough. The absence of a limitations section reads as evasion and invites a skeptical reviewer to find the limits for you.
6. Tighten the abstract and introduction until they are mutually consistent
The abstract, introduction, and conclusion are the parts editors and reviewers read most carefully, and they are the parts most often inconsistent with each other. After months of revisions, the abstract can drift from the introduction, and the conclusion can drift from both.
In the final week, read all three back-to-back. The contribution should be stated identically in all three (in different words, but with the same scope and strength). The numbers should match. The terminology should match. Inconsistency between these sections is a clear signal to a reviewer that the paper was not finished.
7. Get an independent review before you submit
Self-review has a ceiling. You wrote the sentences, so you read into them what you meant, not what they say. The single most effective pre-submission tactic is to get an independent review from a reader who does not share your assumptions.
This can be a colleague, a mentor, or a structured external review that simulates the dimensions a peer reviewer will check. The goal is to surface the issues you cannot see yourself while you still have time to fix them.
If you are revising after reviewer feedback rather than submitting for the first time, the equivalent step is to run your response letter through an independent analysis before you submit it.
Why these tactics work
These seven tactics are not arbitrary. They map directly to the dimensions a peer reviewer evaluates: contribution clarity, claim–evidence fit, related work positioning, figure legibility, honesty about scope, internal consistency, and exposure to an independent reader. Strengthening any of them reduces the chance that a reviewer will find the avoidable problems that sink otherwise good work.
The papers that get accepted are not the ones with the best ideas. They are the ones where the manuscript made the idea legible, the evidence visible, and the contribution undeniable. Run a structured review on your manuscript before submission and close the gap between the paper you wrote and the paper the reviewer reads.
Frequently asked questions
What is the most effective way to strengthen a manuscript before submission? State your contribution as a concrete action rather than an adjective. “We reduce inference latency by 40% on benchmark X” is a contribution; “we propose an efficient method” is not. Reviewers cannot identify contributions they cannot state in one sentence.
How do I know if my claims are supported by my evidence? Map each claim to its supporting result, table, figure, or statistical test. Any claim with no supporting evidence is unsupported and will be flagged by reviewers. Print the claims and results sections side by side and draw the lines—if a line crosses three paragraphs, you have a clarity problem.
Should I include a limitations section in my manuscript? Yes. Reviewers reward honesty about scope and read the absence of a limitations section as evasion. A good limitations section states what your method does not handle, what assumptions it depends on, and where the results might not generalize—usually two paragraphs is enough.
How do I clean up my related work section? Go through every citation and ask what work it is doing. Keep citations that establish a gap your paper fills, define a term or method you use, provide a baseline you compare against, or acknowledge a direct predecessor. Cut everything else—decorative citations signal weak scholarship to reviewers.
Is an independent review before submission worth it? Yes. Self-review has a ceiling: you wrote the sentences, so you read into them what you meant, not what they say. An independent review—whether from a colleague or a structured tool—surfaces the issues you cannot see yourself while you still have time to fix them.