Hashtags still work, just not for what most people think. These days they're a topic and search signal, not a reach lever. Stacking 25 of them on a post won't earn you distribution, and dropping them entirely won't sink you either.
TLDR: Use 3 to 5 specific hashtags per post as a filing label, then put your real discovery effort into the words in your caption, your on-screen text, and your audio. That's what in-app search and the recommendation feeds read in 2026.
What changed: feeds stopped browsing tags
Every major platform now works on interest-based distribution. Your post gets shown to people the system predicts will watch, save, or reply, based on the content itself and early engagement. Nobody sits in a hashtag feed waiting for new posts anymore.
Instagram has said publicly that hashtags don't boost reach, and recommends a small number of relevant ones. TikTok's For You page has always been behaviour-driven. LinkedIn quietly stopped surfacing hashtag follows. The pattern is the same everywhere.
So hashtags have shifted jobs:
- Classification. They tell the platform what your post is about, which nudges who it gets tested on.
- Search. People do tap and search tags, especially on Instagram, TikTok, and Pinterest.
- Campaign tracking. A branded tag still collects entries, UGC, and event content in one place.
Where hashtags still earn their spot in 2026
| Platform | Real value | How many |
|---|---|---|
| Moderate. Search plus topic signal | 3 to 5 | |
| TikTok | Low for reach, real for classification and search | 2 to 3 |
| Low. Mostly topic labelling | 1 to 3 | |
| X | Low. Can reduce reach if spammy | 0 to 2 |
| Minimal. Keywords in the title and description matter far more | 0 to 2 | |
| YouTube | Low, but the first 3 show above the title | 2 to 3 |
| Threads | Only one topic tag is allowed anyway | 1 |
| Almost none | 0 to 2 |
If you take one thing from that table: nowhere on it is a case for 30 hashtags.
How to pick tags that actually do something
Pick for specificity, not volume. A tag used on 40 million posts tells the platform nothing and puts you in a feed you can't compete in. A tag used on 15,000 posts describes you.
A workable mix for a single post:
- One niche tag that names exactly what you do (#sourdoughbakerysydney, not #food).
- Two topic tags at mid-size, roughly 10k to 500k posts, that match the post's actual subject.
- One community or format tag your audience genuinely follows (#studiovlog, #marketingops).
- Your branded tag, every time, so your archive stays collectable.
Skip: banned or spam-adjacent tags, engagement-bait tags (#follow4follow, #likeforlike), tags that don't describe the post, and copy-pasted blocks reused on every upload. Repeating an identical 30-tag block is the clearest low-effort signal you can send.
Do the discovery work in the caption instead
This is the part that replaced hashtags. In-app search now reads your actual words.
- Front-load keywords in the first line of the caption, phrased the way someone would search it.
- Put text on screen in Reels, Shorts, and TikToks. It gets read by the platform and by muted viewers.
- Say the keyword out loud. Auto-captions turn your audio into indexable text.
- Fill in alt text on image posts.
- Name your files and cover frames descriptively before uploading.
If you want the deeper version of this, our guide to optimising for social search on TikTok and Instagram goes platform by platform. And if reach is the actual problem, hashtags are rarely the cause, so start with what actually moves engagement.
Test it on your own account in 30 days
General advice is a starting point. Your data settles it.
- Pick one platform and one content format so you're comparing like with like.
- For two weeks, post with 3 to 5 specific tags. For the next two weeks, post with none.
- Keep everything else stable: same posting times, same format, similar hooks.
- Compare reach from non-followers and search or explore impressions, not total likes.
Most accounts see a difference small enough to be noise, which is the useful finding: it frees you to stop agonising over tag lists. If one platform does show a real lift, keep tags there and drop them elsewhere.
Mydrop makes this test less painful. You can schedule both variants across your content calendar in one pass, keep hashtag sets saved so you're not retyping them, and read the non-follower and search breakdowns per post in Performances instead of exporting three dashboards. Its AI caption tools will also suggest tags scoped to the post rather than a generic block.
Your next step
Open your last 10 posts. Count how many hashtags you used, then check which of those posts actually got non-follower reach. If there's no relationship, cut your tag list to five and spend the saved 10 minutes rewriting your first caption line as something a person would type into search.
That trade is where the return is now.
Google review
Trustpilot review
Product Hunt comment
