Most authors fill in keywords in a few seconds at the last step of the submission form, with whatever comes to mind first. Yet keywords are your paper's search tags: databases classify your paper by them and readers find it through them. Poorly chosen keywords can make a good paper invisible in search results. This guide explains how to choose keywords that genuinely boost discoverability.
What do keywords do?
Keywords are the bridge between your paper and your target reader. When a researcher searches a term in Scopus, PubMed or Google Scholar, your keywords determine whether your paper appears in that result list. So the right keyword finds the right reader — and that reader is the one who will cite you.
How many should you have?
Most journals ask for 4–8 keywords; the range varies from 3 to 8 by journal. The binding number is whatever your target journal's author guidelines state. A good starting point is 4–6 specific terms.
The golden rule: don't repeat title words
This is the most common and most expensive mistake. Databases already index the title; repeating words from the title as keywords adds no new visibility and wastes valuable space.
Instead, use your keywords to widen your indexing footprint: if your title says "machine learning", use the keyword slots for complementary terms like "supervised classification" or "feature selection".
Balance specificity
The strongest keyword set mixes different levels of breadth:
- Narrow terms — the specific name of your method, model or sample. They find you in targeted searches.
- Broad terms — the general field or application area. They find you in exploratory searches.
Using both together makes you visible to the reader "searching for exactly this" and the reader "browsing this field."
Controlled vocabularies
Where possible, use standard terms. Depending on your field:
- MeSH (Medical Subject Headings) — medicine and health sciences. Because PubMed indexes papers using MeSH, many journals accept only MeSH terms.
- IEEE terms — engineering and computer science.
- Scopus subject terms — general use.
Choosing from a controlled list ensures your paper lands in the right category and appears to everyone searching the same term.
What to avoid
- General/vague words — terms like "study", "analysis", "model", "method" add no specificity.
- Overly long phrases — unless it's an established expression in your field, keep each keyword to 4 words or fewer.
- Meaningless abbreviations — unrecognized abbreviations no reader will search for.
How to find them
A practical method: ask yourself, "what would someone searching for this paper type into the box?" Then look at the keywords of similar, recent papers in your target field — you'll see which terms are standard and position your paper alongside them.
Keywords don't work alone: together with the title and abstract, they form your paper's discoverability trio. Use core terms consistently across all three; but instead of repeating the same words in the title and keywords, widen coverage with complementary terms.
Complete visibility with the right journal
The right keywords make your paper findable; but the paper delivers its real impact when it's in a journal your target reader scans. Peerfect's free journal recommendation engine puts journals matching your paper's scope in front of you; when you combine the right journal with the right keywords, your paper appears in front of exactly the reader searching for it.
In short: keywords aren't a formality — they're your paper's search engine. Choose 4–8 specific terms, don't repeat title words, balance narrow and broad terms, draw on controlled lists like MeSH where possible, and avoid vague words. This few-minute careful selection directly affects the reads and citations your paper earns for years.
Sources: How to create keywords for a research paper — Editage, How to Choose Keywords for Your Research Paper — MDPI Blog, Choosing Effective Keywords — AJE. The exact number and format vary by your target journal's author guidelines.