What 'good' looks like for hashtag strategy across platforms and tiers. Numbers drawn from public datasets and the patterns we measure across the user base.
The tier mix is the single most-actionable rule on this page. Most accounts default to either the lazy pattern (top 15 generic tags) or the spam pattern (30 mega tags). Both underperform the tiered 60/30/10 mix by margins large enough to show up in account-level analytics within two weeks. If you change one thing about your hashtag workflow, change to the tier mix.
Niche tag discovery is the bottleneck most hashtag strategies hit. The 100K-post ceiling for niche tags is a moving target, a tag with 80K posts today might cross 200K in three months and lose its niche advantage. A hashtag generator that re-runs the niche search per request avoids this drift; manual research surfaces yesterday's niche tags long after they've grown out of the tier.
The platform-specific count rules above are the single most-violated rules in hashtag strategy. The instinct from Instagram (more is better, fill the slot count) does not transfer to TikTok, X or LinkedIn. Each platform has its own count rule and the same set does not transfer. The fastest reach upgrade for accounts active on multiple platforms is to enforce the platform-specific counts: 8-15 on Instagram, 3-5 on TikTok, 1-2 on X, exactly 3 on LinkedIn.
The geo-tag lift for local businesses is the highest single-pattern ROI we measure on the platform. A plumber, a salon, or a restaurant using city + service tags (#bostonplumber, #austinsalon) earns 3-5× the discovery of the same business using only generic top tags. This is the pattern most local owners don't know about, and once they switch, the lift is immediate and durable.
The 30-day rotation cadence handles two problems at once: the spam-filter flag from identical sets, and the natural decay of tag popularity. Keeping the cadence on the calendar, first Monday of every month, re-run the generator, refresh the three saved sets, is the single most underrated hashtag habit.