GlossaryKeyword

What is a keyword?

A keyword is the word or phrase a person types into a search bar to find content on a topic, and on social media in 2026 it is the on-page signal the platform reads out of the caption, the title, the bio, the alt text, the on-screen text, and the audio to decide whether a post is about that topic and should appear when someone searches for it.

What is a keyword?

A keyword is the actual word or phrase someone types into a search bar when they are looking for something, written in plain language with no special syntax in front of it. On Google the keyword is what the page is built to rank for. On social media the keyword is what the algorithm reads out of the caption, the title, the bio, the alt text, and the on-screen text to work out what a post is about and which searches it should appear in.

The word came out of search engine optimisation in the late 1990s, when keyword tags inside the page's HTML were literally the way Google read what a page was about. Two decades later the meta-keyword tag has long since stopped mattering, but the underlying idea (use the words your reader is going to search for, inside the content itself) has spread out from Google into every search-enabled surface, including the search bar inside Instagram, TikTok, LinkedIn, YouTube, and Pinterest. The keyword is what those search bars match against.

Sprout Social's overview of social media SEO frames the same point from the working-team angle: a relevant keyword in the caption, the title, the description, and the on-screen text now does most of the discovery work that was the job of the hashtag stack in 2018. The platforms read those words, classify the post, and decide which searches it can appear in.

Keyword vs hashtag in 2026

Keywords and hashtags get talked about together because they do related work, and the easiest way to keep them straight is to look at what each one is and what each one is for. A hashtag is a word with the # symbol in front of it and no spaces inside it, which the platform reads as a topic label. A keyword is a word in ordinary writing, which the platform reads as part of the post's text and feeds into its search index. The two coexist on every platform; the share of the work each one does has shifted.

What a hashtag does

Tells the platform what category the post belongs to (a topic label, in metadata terms), feeds the hashtag search results, and anchors campaign and community conversations under a single tag. The signal is categorical: the post either has the tag or it does not.

What a keyword does

Tells the platform what the post is actually about, by appearing inside the caption, the title, the bio, the alt text, the on-screen text, or the spoken audio. The signal is textual: the platform's search index matches the words in the post against the words someone has typed into the search bar.

Why keywords now do more of the work

Instagram, TikTok, and LinkedIn all rebuilt their search systems between 2020 and 2023 to match plain-language queries against the text of the post rather than only against the tag stack. Social Media Today's write-up on Instagram adding keyword search documented the start of that shift in late 2020, and the platforms have widened the matchable text every year since.

How they work together

A small set of three to five relevant hashtags helps with categorisation; a caption that plainly says what the post is about helps with search. The combined signal beats either one alone. The mistake is treating them as substitutes: a stack of thirty hashtags with a one-emoji caption now performs worse than three hashtags plus a caption that explains what the video is.

For the longer story on the tag side of this pair, the hashtag entry covers the rules, the history, and the per-platform counts. The short version, on the keyword side, is that the words the caller types are now the words the platform looks for, and those words almost always live in the caption.

Where keywords live in a social media post

A single post has more keyword slots than most teams use. The list below covers the places the major platforms read words out of, ranked roughly by how much weight each slot carries when the search and recommendation systems try to work out what the post is about.

The caption

The single most important keyword slot on Instagram, TikTok, LinkedIn, and Threads, because the search system reads the caption text as the primary description of what the post is. The first 125 characters carry the most weight on Instagram (that is the part shown before the See More fold). The first sentence carries the most weight on LinkedIn for the same reason.

The title

On YouTube the title is the highest-weighted slot. On TikTok the caption acts as the title; on Pinterest the pin title is its own field and is read alongside the description. Front-load the primary keyword wherever the platform exposes a title field.

The bio and the name field

Instagram and TikTok index the words in the profile bio and (heavily) the words in the name field separately from the username. A name field that says "Brisbane pilates teacher" lands the account in the local pilates search results in a way that the brand name alone never does. The same pattern applies to the LinkedIn headline.

The alt text

Instagram, LinkedIn, and Pinterest let creators write alt text on each image. The text is meant for screen readers (accessibility comes first) but also feeds the platform's search index and, on Pinterest, the visual-search categorisation. A short, plain sentence describing the image earns the slot.

The on-screen text

Reels, TikTok, YouTube Shorts, and Pinterest video read the words burned onto the screen using OCR and feed those words into the same search index as the caption. A caption keyword that also appears on the screen carries more weight than the same keyword in only one place.

The spoken audio

TikTok and Instagram both transcribe the audio of every uploaded video and use the transcription as a search signal. Saying the keyword out loud in the first three seconds of the video does roughly the same work as putting it in the caption.

The hashtags

A hashtag is its own slot, and the platform reads the words inside the tag (with the # stripped off) as additional keyword signal. Three to five tags that match the actual subject of the post is the working count in 2026 on Instagram, TikTok, and LinkedIn.

The closed captions and the description

YouTube reads the video description and the closed-caption track as two of its strongest keyword signals. A description that genuinely explains what the video covers, with the primary keyword in the first sentence and related keywords through the rest, still moves the needle on YouTube in 2026.

Hootsuite's walkthrough of Instagram SEO lays out the same slots on the Instagram side, with the caveat that only caption text is indexed for the in-app search (hashtags placed in the first comment, which used to be a clean-caption trick, no longer feed the search results in the way they once did).

Keyword search by platform in 2026

The job a keyword does is the same on every platform: tell the search index what the post is about. The way each platform reads keywords, and the slots that carry the most weight, is not. The breakdown below is the working version most social teams now plan around.

Instagram

Search reads against the caption first, the name and bio second, the alt text third, and the on-screen text and spoken audio on Reels alongside those. Instagram's own announcement of full keyword search in November 2020, covered in Social Media Today's write-up of Instagram's keyword search, was the moment the platform stopped relying on hashtags and account names alone and started indexing the words inside the post.

TikTok

The For You feed is driven by interest signals, but TikTok search has become a meaningful traffic source in its own right. The search index reads the caption, the on-screen text, the transcribed audio, the sounds, and (lightly) the hashtags. The platform now shows related-search chips under the search results, which double as a free keyword-research tool: type a seed phrase and read the chips for the most-searched neighbouring queries.

LinkedIn

People Search and Content Search both run against the visible text of the profile (headline, About, Experience, Skills, Education) and the posts themselves. The primary keyword belongs in the headline; the secondary keywords belong in the About section. Hootsuite's LinkedIn SEO guide covers the slot weighting in detail.

YouTube

Title and description carry the most weight, with the closed captions running close behind. YouTube search is the second-largest search engine on the web (it is owned by Google but ranks independently), so the keyword research that powers a YouTube video genuinely is SEO in the strict sense of the word: pick the phrase you want to win, put it in the title, expand on it in the description, and let the closed captions do the rest.

Pinterest

Pinterest is the most search-driven platform in the group; people arrive with an intent to find rather than a habit of scrolling. The pin title, the description, the alt text on the image, and the board name are all indexed. Hashtags stopped doing topical work on Pinterest in 2020 and the categorisation is done by the keywords in the text fields.

Threads, Bluesky, X

All three index the post body for search. Threads and X expose a single text field; Bluesky uses an open-source search index that reads the same way. Keywords in the body of the post are the relevant signal; hashtags add a categorical lift.

Across the six the shared pattern is that the search index reads the body of the post and not only the tag stack, and that the caption (or the title, where the platform has one) is the slot that carries the most weight. The implication for a creator is the cheap one: write captions that plainly say what the post is about.

Why Gen Z searches social before Google

The reason keywords now matter so much on social media is that social media is where a growing share of search now happens. Prabhakar Raghavan, then Google's senior vice president of Knowledge and Information, said at a Fortune conference in July 2022 that roughly 40 per cent of young people looking for somewhere to eat now go to TikTok or Instagram rather than Google Maps or Google Search. The quote, reported by TechCrunch's write-up of the Google internal data, was Google publicly acknowledging that the under-25 search habit has moved.

The behaviour has held since. Sprout Social's 2025 round-up of social search research puts the share of Gen Z who turn to social media first when they are looking something up at more than 40 per cent. The reasons people give in the surveys are consistent: video answers feel faster than a written page, the search results include real opinions from real people rather than affiliate listicles, and the format of a TikTok or a Reel matches the topic better for visual queries (recipes, makeup, travel, renovations, gym form).

Three things follow for anyone publishing on social. The keyword the searcher actually types matters as much as it does on Google, because the platforms are now matching against typed queries. The format of the answer matters more, because a video that opens with the answer beats a video that takes thirty seconds to arrive at it. And the old SEO discipline of writing in the language your reader uses (rather than the language the brand uses internally) ports over almost unchanged.

How to do keyword research for social media

Keyword research for social media uses the same mental model as keyword research for Google (find what the audience types, find what has enough volume to be worth writing for, find what is specific enough to actually rank for) but the source data is inside the social platforms themselves. The five sources below are the ones that consistently turn up the queries the audience is actually using.

  1. Platform predictive search. Start typing a seed phrase into the Instagram, TikTok, LinkedIn, or YouTube search bar and read the suggestions that drop down. Each one is a query other people have typed; the order is roughly by volume. Five minutes of typing into a search bar produces a usable list of 20 to 30 phrases per topic.
  2. TikTok related-search chips.Run a search on TikTok and scroll to the bottom of the results. The chips labelled “Others searched for” are live data on adjacent queries and are the cleanest source of new phrases for the same topic.
  3. TikTok Creative Center. The free side of the TikTok Creative Center (under Inspiration, then Keyword Insights) shows search volume on advertising keywords by country. The data is biased toward advertiser-relevant terms but is still the closest thing to a TikTok-native keyword planner.
  4. YouTube auto-suggest. The same predictive drop-down that exists on the main YouTube search bar is the working keyword tool for video. Combine with the Search Insights panel inside YouTube Studio, which shows the actual phrases viewers typed before reaching the channel.
  5. Google Keyword Planner, for context only. Google's free Keyword Planner is still useful for checking whether a phrase has volume in absolute terms. The numbers do not map directly to TikTok or Instagram (the same phrase can be popular on social and quiet on Google, or the other way around), but the planner is the cheapest sanity check on whether the audience is using the phrase at all.

A workable cadence is to spend an hour per content pillar, once a quarter, pulling 10 to 20 phrases from the five sources above and dropping them into a shared sheet. The per-post decision then becomes which phrase to write about next rather than what to brainstorm.

How to use keywords without sounding like a robot

The trap with keyword work on social media is the same trap SEO writers fell into a decade ago: a caption written for the algorithm reads to the human like a brochure, and the human is the one deciding whether to watch, like, save, or share. The platforms know this and weight engagement signals far more heavily than keyword density, which means a caption that no one engages with stops being shown regardless of how many keywords it contains.

  1. Open with the answer. The keyword belongs in the first sentence because the first sentence is what shows before the See More fold and what the search index weights most heavily. Open with the answer to the question the keyword represents, then expand from there.
  2. Write in the words the audience uses. A caption that says “banana bread without gluten” will outrank a caption that says “coeliac-friendly baked good” because the first phrase is the one people search. The brand voice is a poor predictor of what the audience types.
  3. Say the keyword out loud. On TikTok and Reels the audio is transcribed and indexed. The keyword said in the first three seconds of the video does the same work as the keyword in the caption, and the combination of both is what wins.
  4. Put the keyword in the alt text. A short accessibility description of the image (one sentence, plain English) earns the slot. Avoid stuffing the alt text with multiple keywords; the field is small and screen readers will read it aloud verbatim.
  5. Use natural variants.The search index can stem (“running” matches “run”) but it cannot guess synonyms with perfect accuracy. Writing the same idea two or three ways across the caption catches the searcher who typed it slightly differently from the one you imagined.
  6. Test the post against the search bar. After publishing, type the primary keyword into the platform's search bar and check whether the post actually surfaces in the results. If it does not, the slots are not yet pulling weight (most often: the keyword is missing from the caption, the alt text, or the on-screen text).

Common keyword mistakes

  1. Putting the keyword only in the hashtags. A caption that says nothing and ends in #brisbanepilates #pilatesteacher #reformerpilates is the most common mistake on the list. The tags categorise the post; the caption is what the search index reads. Without the keyword in the caption the post will not rank for the phrase, no matter what the tags say.
  2. Keyword stuffing the caption.A caption that reads “best Sydney coffee, best coffee in Sydney, top Sydney coffee shops, Sydney coffee 2026” back to back triggers the same low-quality signal that flattened SEO-stuffed pages in the 2010s. Once per natural sentence is enough.
  3. Optimising for the brand's words instead of the audience's. A skincare brand calling its product line a “clean-beauty ritual” in every caption misses the fact that the searches that drive sales are “moisturiser for oily skin” and “cleanser for blackheads”. The audience does not search for the brand language.
  4. Treating Google volume as social volume. A phrase that has 90,000 Google searches a month can have almost no TikTok search volume, and a phrase with almost no Google volume can be a TikTok stream. The two intent profiles are different; check the platform itself before committing.
  5. Skipping the alt text and the on-screen text. Both slots are searchable on Instagram and TikTok, and both take less than thirty seconds to fill. Leaving them blank is leaving free reach on the floor.
  6. Forgetting the name field on the profile. The name field on Instagram and the headline on LinkedIn are the highest-weighted profile keywords by a wide margin and the slot teams most often waste on the brand name alone. A profile called “Lunar · Brisbane pilates teacher” outranks a profile called “Lunar” for the local query every time.
  7. Writing for the algorithm at the cost of the reader. The platforms weight engagement above keyword presence. A caption that contains the keyword but reads like spam will lose to a caption that contains the same keyword and reads like a person wrote it. Read the caption aloud; if it sounds like a brochure, rewrite it.
  8. Doing keyword research once and never updating it. Search behaviour on social moves quickly; the phrase that was a stream in March can be saturated by November. Refresh the keyword list once a quarter using the platforms' own predictive search.

For the surrounding context this entry sits inside, the hashtag entry covers the topic-label side of the same pair, the caption entry covers the writing the keywords sit inside, the bio entry covers the highest-weighted profile keyword slot, and the algorithm entry covers the ranking systems that decide what a keyword-rich post actually reaches.

The matching tools on this site cover the per-platform side of writing around a keyword: the Instagram caption generator builds the caption around the keyword you want to be found for, the LinkedIn headline generator works the same way for the highest-weighted slot on a LinkedIn profile, and the character counter keeps the caption inside each platform's limit so the keyword does not get cut off above the See More fold.

Keyword FAQ

What is a keyword in simple terms?

A keyword is the actual word or phrase a person types into a search bar to find something, and (on social media in 2026) it is the on-page signal that decides whether the platform thinks a post, profile, or video is about that topic. On Instagram, TikTok, LinkedIn, YouTube, and Pinterest the search bar reads the keywords out of the caption, the title, the bio, the alt text, the on-screen text, and the spoken audio rather than only the hashtags, which is why a caption that plainly says what the post is about now does more for discovery than a stack of tags.

What is the difference between a keyword and a hashtag?

A hashtag is a word with # at the front and no spaces inside it, used as a topic label the platform reads as metadata. A keyword is just a word, written in plain English inside the caption, the title, the bio, or the on-screen text, used by the platform's search and recommendation systems to work out what the post is actually about. Instagram, TikTok, and LinkedIn now read both, but the working consensus from Instagram's own ranking explainer through to Hootsuite and Sprout Social is that keywords in the caption now do more of the heavy lifting for search and discovery than the tag stack does.

Where should I put keywords in an Instagram post?

The caption first, because the caption is what Instagram search matches against most directly. Then the alt text on the image (which doubles as accessibility metadata and a search signal), then the on-screen text on a Reel (so the OCR pipeline can read it), then the spoken audio (which the auto-captions transcribe). The bio and the name field on the profile carry the two highest-weighted keyword slots on the account itself, which is why a freelance pilates teacher in Brisbane is better off writing "Pilates teacher, Brisbane" in the name field than putting the studio's brand name in the same slot.

Do keywords matter on TikTok in 2026?

Yes, and they matter more on TikTok than on almost anywhere else, because TikTok search has grown into a meaningful share of how Gen Z looks things up. A Google senior vice president said in 2022 that roughly 40 per cent of young people now turn to TikTok or Instagram instead of Google Maps or Search when they are looking for a place to eat, and that share has held or grown since. TikTok reads keywords out of the caption, the on-screen text, the spoken audio (it transcribes the speech), and the sounds and stickers used in the video, then matches them against the user's typed query. A video about "the best ramen in Sydney" needs those words said aloud, written on the screen, and present in the caption if it wants to win the search.

How many keywords should a LinkedIn profile have?

Treat the LinkedIn profile as five separate keyword slots: the headline (120 characters), the About section (2,600 characters), the current job title, the Experience entries, and the Skills list. Put the primary keyword (the one you most want to be found for) in the headline and the first 220 characters of the About section, and let the secondary keywords scatter naturally across the Experience entries and the Skills list. Hootsuite's LinkedIn SEO write-up frames the goal as having the keyword appear in three or four places across the profile rather than stuffed into one section, which lines up with how LinkedIn's own People Search ranks results.

Should I still do SEO-style keyword research for social media?

Lightly, yes, with the source moved off Google. The mechanics are the same (find what people type, find which terms have search volume, find which terms have low enough competition to actually rank for), but the data lives inside the social platforms themselves now: the predictive search suggestions when you start typing into Instagram, TikTok, or LinkedIn search; the related-search chips at the bottom of TikTok search results; the YouTube auto-suggest; and the topic clusters in TikTok's Creative Center. Pull a list of 10 to 20 phrases per content pillar from those sources, then write the posts around the phrases.

EziBreezy GlossaryMore terms
Keep Learning
  1. No. 01Glossary

    Hashtag

    A hashtag is a word or phrase preceded by the # symbol, written with no spaces, that labels a social media post by topic so the platform can group it with other posts using the same tag and surface it through search and recommendation feeds.

  2. No. 02Glossary

    Social media caption

    A social media caption is the text written next to a post that gives the photo, video, or carousel its meaning, helps the algorithm understand the topic, and gives the audience a clear next step.

  3. No. 03Glossary

    Social media bio

    A social media bio is the short block of text on a profile that says who you are, what you do, and where people should click next, written within a strict per-platform character limit.

  4. No. 04Glossary

    Social media algorithm

    A social media algorithm is a ranking and recommendation system that uses signals from people, posts, accounts, and context to decide which content appears for each user and in what order.

Trusted by creators and agencies who refuse to settle

WDYM
Jargon?