Hashtags vs. Keywords: Which Actually Drives Discovery Now?

I've been testing this for months, and the answer is more complicated than the hot takes suggest. Hashtags aren't dead. Keywords aren't automatically winning. And the platform you're on changes everything about which strategy deserves your energy.

Let's actually compare these two discovery systems — how they work, where they're gaining or losing ground, and what a realistic content creator should do with this information in 2024.

First, What We're Really Comparing

Hashtags have always been a metadata layer. You tag your content with #gymtok or #sourdough and the platform files it under that label. Users browse those labels, or the algorithm uses them to categorize and distribute your post to relevant feeds.

Keyword-based discovery — what people increasingly call "in-app SEO" — is different. Here, the platform's search engine reads the actual text of your caption, your on-screen text, your spoken audio (via auto-captions), and sometimes even what appears visually in your images or video. When someone types a query into the search bar, the algorithm tries to surface content that genuinely answers that query, not just content that self-selected into a hashtag bucket.

These are not the same thing, and the distinction matters more than most people realize.

Instagram: The Platform Mid-Identity-Crisis

Instagram has been quietly shifting its search behavior for years. If you search for "vintage bookshelf styling" right now, you'll get a mix of Reels, posts, and Guides — and a meaningful chunk of them will have matched on caption keywords, not hashtag choices.

The platform's own guidance (buried in their creator documentation) now emphasizes "descriptive captions" over hashtag stacking. That's a significant shift from the era when 30 hashtags in the first comment was considered best practice.

What the data from creators actually shows:

• Hashtags still deliver some reach into curated hashtag feeds, but those feeds have declining engagement rates compared to a few years ago
• Reels are more likely to be discovered through topic-based search than static posts
• The explore page is increasingly personalized by behavior signals, not hashtag following

That said, hashtags aren't useless on Instagram. They still function as categorical signals to the algorithm, even if users rarely browse them directly. The change is that you no longer need 20 hashtags to communicate your topic — the algorithm can now read your caption to figure that out.

What actually works on Instagram right now: A caption that leads with natural language describing your content ("Here's how I style a small bedroom without buying anything new") performs better in search than a caption that leads with hashtags or says nothing at all. Then add 3–5 targeted hashtags as secondary signals, not primary strategy.

TikTok: Where Search Has Become Legitimately Important

TikTok's transformation into a search engine is the most significant platform shift of the last two years, and it's not hyperbole. A 2023 Adobe study found that 40% of Gen Z users prefer TikTok or Instagram over Google for search queries. A separate Google internal study (which leaked) showed 40% of young people using TikTok or Instagram as their entry point for finding places to eat.

This has real consequences for how content gets discovered.

TikTok's search function reads:

• The text overlay and captions you add to videos
• The audio (either your speech or auto-generated captions)
• Your video description
• Comments — including what you say in comment replies

Hashtags on TikTok still serve a function, but creators who have tested this systematically report that the keyword content of the video itself is the stronger ranking signal for search. A video where you verbally say "the best way to remove gel nails at home" and include that phrase as an on-screen text overlay will surface for that search query whether or not you used #gelnails or #nailcare.

The interesting asymmetry: hashtags on TikTok still influence For You Page distribution (the algorithmic feed), while keywords primarily influence search results. These are different distribution channels with different audiences and intent levels.

Someone scrolling their For You Page is in passive discovery mode. Someone who typed "how to remove gel nails at home" into TikTok search has explicit intent and is much more likely to follow through on whatever action your content suggests. For conversion-oriented creators — educators, product sellers, service providers — search traffic is arguably more valuable even at lower volume.

What actually works on TikTok right now: Treat the first three seconds of audio as SEO metadata. Say your topic out loud. Mirror that in on-screen text. Write a video description that reads like a natural search query ("How to remove gel nails at home without damage"). Then use 3–5 relevant hashtags. The creators winning in TikTok search are doing keyword research — actually typing queries into TikTok's search bar to see what autocompletes, just like they'd do for Google.

Pinterest: The Platform That's Been Doing This Longest

Pinterest is worth examining separately because it essentially started as a keyword-and-image search engine, not a social platform. Understanding how hashtags and keywords function there offers the clearest contrast.

For years, Pinterest hashtags were almost vestigial — clickable but not a primary driver of distribution. The real ranking signals were always in Pin titles, descriptions, and board names, which the algorithm indexed like a search engine would.

Pinterest's 2023 direction has reinforced this. The platform has made its "Trends" tool more prominent (showing search volume data to creators), added keyword targeting to organic distribution, and — importantly — uses visual search as an additional discovery layer. The algorithm can identify what's in an image and surface it for relevant searches, independent of text metadata entirely.

Hashtags on Pinterest now function primarily as clickable category links rather than distribution signals. They can help users self-select deeper into a topic, but they don't meaningfully boost your distribution reach the way they once did.

What actually works on Pinterest right now: Write Pin descriptions the way you'd write a search result snippet. Include the natural language someone would actually type ("small bedroom organization ideas for renters"), not just labels ("#bedroomorganization"). Use Pinterest Trends to identify genuine search volume before creating content. Board names matter more than people realize — they're part of how the platform categorizes your entire account.

The Underlying Shift Across All Three

There's a common thread: platforms are all investing in their search products because search signals intent, and intent makes advertising more valuable. Instagram Search, TikTok Search, and Pinterest's entire model all monetize better when they can match high-intent users to relevant content (and adjacent ads).

Hashtags, by contrast, were primarily a community-building and categorization tool. They thrived when platforms needed users to organize content for them, before machine learning could read captions, transcribe audio, and understand images.

That's not a knock on hashtags — it's just an honest read of why keyword-based discovery is becoming structurally more important. The platforms have better tools now.

The Honest Side-by-Side

What This Means Practically

The advice "just use keywords, hashtags are dead" is wrong. So is "just use hashtags, SEO doesn't matter for social." The realistic answer is that these tools serve different functions and you need both — just not in the ratios most creators currently use.

Most creators still spend more time choosing hashtags than writing search-optimized captions. Given where discovery is heading, that balance is probably inverted from what it should be.

A more useful framework: write your caption first as if it's answering a specific question someone would search for. Make sure your spoken audio (on video platforms) includes the natural language of your topic. Then add a small set of relevant hashtags as secondary signals — not because they'll carry your reach, but because they still provide categorical context the algorithm uses.

The creators who are growing fastest right now on all three platforms aren't the ones who cracked some hashtag formula. They're the ones who thought like a search engine before it became fashionable to do so — writing content that actually answers specific questions, and making that obvious in every text layer of their posts.

Hashtags got content discovered when platforms needed users to do that organizational work. Keywords let content get discovered even when a user doesn't know your account exists — they just know what problem they're trying to solve. That's the shift worth understanding.