Keyword List Randomizer
Randomise the order of a keyword list, or take a random sample of a given size, using a seed so results are reproducible.
Your input
Everything you type is processed in this browser tab. Nothing is uploaded, logged or stored.
Results
Fill in the fields and choose Run tool to see results here.
About this tool
Randomising a keyword list has two practical uses: sampling and fairness. Sampling lets you review fifty phrases out of two thousand without bias towards whatever happened to be at the top. Fairness matters when you are dividing work between people or rotating which keywords get attention first. This tool does both, reproducibly.
How it works
Your list is shuffled with a seeded pseudo-random algorithm, so the same seed always produces the same order. That reproducibility is the point: you can record the seed alongside a sample and regenerate the identical selection later, which matters when you need to justify a quality review or repeat an experiment. A sample size greater than zero truncates the shuffled list to that length.
- 1Paste the keyword list.
- 2Set a sample size, or leave it at zero to shuffle the whole list.
- 3Set a seed you can record, or keep the default.
- 4Generate and export the shuffled or sampled list.
Example
Input
4 keywords, sample size 2
Output
trail shoes · running shoes (seed: hashly)
When to use it
- Drawing an unbiased sample for manual quality review.
- Splitting a large list fairly between several writers.
- Rotating which keywords are worked on first each week.
- Creating reproducible test sets for comparing two approaches.
Tips
- Write the seed into your notes so the sample can be reproduced.
- For quality review, ten percent of a list usually surfaces systemic problems.
- Deduplicate first, otherwise a sample can contain the same keyword twice.
Limitations
The randomness is pseudo-random and deterministic, which is a feature here rather than a flaw, but it is not suitable for anything requiring cryptographic unpredictability. Sampling also gives no guarantee of representativeness — a random fifty may miss an entire theme.
Frequently asked questions
- Why is the shuffle not truly random?
- It is seeded so results are reproducible. Change the seed for a different order; use a new seed each time if you want variety.
- Can I sample more than the list length?
- No. The output is capped at the number of keywords you supplied.
- Does sampling remove duplicates?
- No. Deduplicate first if you want a sample of unique keywords.