"Use low-competition hashtags" is correct but incomplete advice. A tag with four posts on it is technically uncontested and completely useless — nobody is looking at it. The target is the narrow band where competition is low but human search behaviour still exists.
The three tests
Does anyone use it? Open the tag. If there are recent posts from real accounts, people are there. If the most recent post is from eight months ago, it is a dead tag.
Would someone type it? Read it aloud. If it sounds like a phrase a person searching for your content would use, it passes. If it sounds like a tool generated it, it fails, however low the competition.
Is your post actually about this? Ranking on an irrelevant tag brings viewers who bounce, which teaches the platform your content disappoints its audience.
Four reliable sources
Modifier stacking. Take your core term and add a qualifier — the skill level, the constraint, the equipment, the location, the season. Each qualifier cuts the pool sharply while keeping intent intact.
Related tags on the platform. Every tag page suggests related tags. Two or three clicks in from a broad tag is usually where the useful specificity lives.
Your own comments. The exact words your audience uses to describe what you make are the best tags you will ever find, and they are sitting in your replies already.
Adjacent communities. Forums, subreddits and Discord servers in your niche use vocabulary that has not yet migrated to social tags. That gap is opportunity.
Build a ladder, not a list
The most durable structure is a ladder: two or three tags you can realistically top, three or four you can place in, and one or two you almost certainly cannot. As your account grows, the whole ladder shifts upward — the tags you could not touch last year become the ones you place in.
Re-check quarterly
Tag pools grow. A tag that was winnable eighteen months ago may now be crowded. Re-running your core set every quarter takes twenty minutes and keeps the ladder honest.