Instagram post ideas: 30 that earn a send, not just a like

Running out of Instagram post ideas is almost never the real problem. Most accounts that feel stuck have a notes app full of ideas and no way to tell which of them is worth the afternoon it takes to make. So they post the safe one, it does nothing, and the list gets longer.

The fix is not a longer list. It is a filter. Instagram publishes, in plain language, what it counts when it decides who sees a post, and that ranking order is the only filter worth using. This piece gives you thirty ideas sorted by the action each one is built to earn, the published benchmark to grade the result against, and a calculator to check whether the last one actually worked before you make the next one.

Why most Instagram post ideas fail before anyone sees them

Instagram is unusually direct about how it ranks. Its own explainer lists the most important predictions the system makes about a reel, in order: how likely you are to reshare it, watch it all the way through, like it, and go to the audio page.

Read that order again. Reshare is first. Like is third.

Almost every account chooses its ideas to earn the third one. "Will people like this?" is the question in the room when a post gets picked, and liking is the cheapest action on the platform. It costs nothing, it signals almost nothing, and it is the weakest of the four things Instagram is watching for.

A send is the expensive one. Sending a post to a friend means putting your own judgment behind it in a private conversation. Nobody sends a post that only says a nice thing about your brand. People send posts that settle an argument, save someone money, name something everyone noticed but nobody said, or are funny enough to be worth a small social risk.

So the filter is one question, asked of every idea before it gets made: who would send this, and to whom? If you cannot name the sender and the recipient, the idea is a like at best. That is not a reason to bin it, but it is a reason not to spend a day on it.

What should I post on Instagram when I have no ideas?

Sort your ideas into five buckets, and rotate them. Each bucket is built to earn a different action, and an account that only ever runs one bucket flattens inside a month: all teaching and you are a textbook nobody buys from, all selling and the audience stops arriving.

Teach one thing (earns saves)

A save is a private bookmark. It says "I will need this later", which is the strongest signal that your account is worth following rather than just worth watching.

1. The one mistake you see most often in your field, and the thirty second fix. Name the mistake in the first three words.
2. A before and after with the actual settings, measurements or numbers on screen. Not "we improved this", but what changed and by how much.
3. The thing customers always get wrong about your product, answered without defending yourself.
4. A checklist someone could screenshot and use tomorrow. Five items maximum, or nobody saves it.
5. The vocabulary of your industry, decoded. Every field has five words that quietly exclude beginners.
6. What you would do differently with a small budget versus a real one. Most advice assumes the real one.

Show the work (earns comments)

Process posts earn comments because they are unfinished, and an unfinished thing invites an opinion. Polished work invites silence.

7. The version you rejected, beside the one you shipped, with the reason.
8. A day of the job that nobody outside it has seen. The unglamorous half is the interesting half.
9. Your actual workspace, tools and setup, without tidying it first.
10. A mistake that cost you something, and what it cost. Specific, not "we learned so much".
11. The first version of your product beside the current one.
12. A job going wrong in real time, posted before you know how it ends. This one is uncomfortable and it works.

Take a position (earns shares)

A share is public. People share things that say something they already believe but had not put into words, which means a post with no edge cannot be shared, only liked.

13. The industry advice you think is wrong, and what you do instead.
14. A price you think is too high, including in your own category.
15. The thing everyone in your field does because everyone in your field does it.
16. Who your product is not for, stated plainly. This sells more than the reverse.
17. A prediction with a date on it. A prediction with no date is a horoscope.
18. The trade off nobody admits. Every product has one, and naming yours buys more trust than hiding it costs.

Involve the reader (earns replies)

Replies are private, they compound, and they feed the ranking signal Instagram calls your history of interacting with someone.

19. A question with a genuinely wrong answer. "What do you think?" gets nothing; "Which of these two is the mistake?" gets a queue.
20. Two options, one poll, and the promise that you will actually do the winner.
21. Ask for the worst experience anyone has had in your category. You will get a month of content in the replies.
22. A fill in the blank that only your audience could complete.
23. Post the question you get asked most, and let people answer it before you do.
24. Ask what you should make next, then publicly credit whoever suggested the one you make.

Sell plainly (earns revenue)

The bucket most accounts either avoid entirely or do badly by hiding the sell inside something else. One product, one reason, no apology.

25. One product, one specific problem it solves, one price. No lifestyle framing.
26. The customer review that changed your mind about your own product.
27. A demonstration with no music and no cuts, in one take.
28. What it costs and what it does not include. Naming the exclusions is what makes the price believable.
29. The comparison you would make if you were the customer, including where a competitor is better.
30. An offer with a real reason to end. A deadline with no reason behind it reads as a tactic and gets ignored.

How do I know if a post idea actually worked?

You compare it to a published median, not to your own last post and not to a number you feel good about. The number to compare it to is lower than most people expect.

Rival IQ put the all-industry median Instagram engagement rate by follower at 0.30% in its 2026 benchmark report, measured as total interactions divided by follower count, down from 0.36% a year earlier. That means an ordinary business account earns about three interactions per thousand followers, and it means the whole field is drifting down. A rate that held flat this year was a small win.

Published benchmarks disagree with each other, and the reason is arithmetic rather than performance. Socialinsider reports an average of 0.48% across 70 million posts, counting likes and comments only. Same platform, different formula, and neither number is comparable to the other.

What this means for a small account: pick one definition, write it down, and never change it mid-year. A trend line built on a moving formula is measuring your spreadsheet, not your posts.

There are also two different rates, and they answer different questions. Divide by followers and you learn whether your audience still cares, which is the number every public benchmark uses. Divide by reach and you learn whether the post earned its distribution, which is the number Instagram Insights shows you.

A post with a strong reach rate and a weak follower rate travelled well past the people who already follow you. That is usually your best post of the month, and an account tracking only the follower rate will file it as a failure.

Here is the calculator. Take every figure off one post's insights screen, not off three different weeks.

Which format should each idea belong in?

The idea is not the post. Choosing the wrong container is where most decent ideas quietly die, because a position needs motion and a teaching post needs a reader who can go back a slide.

A word on the shape of the examples in this post

Every example post above is square. That is worth saying out loud, because square stopped being the feed's default shape some years ago: Instagram now supports a range of crops per placement, and Meta publishes which ratios each placement takes. Most brands design for the tall end of that range and treat the square version as the fallback, because a taller post occupies more of the screen on the way past.

The reason the square examples still teach the idea is the thing worth taking from them: in every one, the headline, the product and the one action sit inside the middle of the frame. Everything decorative sits at the edges. That is what a layout designed to survive a crop looks like, and it is the difference between one asset you can post in three places and three you have to make separately.

The pattern underneath the table: reels buy reach, carousels buy saves, single images buy the fast reaction, stories buy replies. Pick the container after you have the idea, never before, and never because it is the format you are most comfortable making.

Instagram also publishes the production choices that get a reel shown to fewer people: low-resolution or watermarked reels, reels that are muted or contain borders, reels that are majority text, or reels that have already been posted on Instagram.

Every item on that list is a production decision, not a content one, which makes them the cheapest fixes available. The watermark is the one that costs most accounts the most reach: exporting a video out of another app with its logo burned in is a self-inflicted demotion on an otherwise good idea. The same principle applies to paid placements, where the length of an Instagram story ad changes what the format can carry.

How many posts a week does this actually take?

Five, if each one has a different job. Volume is not the lever, and posting seven versions of the same bucket is worse than posting three that rotate.

The load bearing day is the one with no post on it. Reading the comments on Tuesday's position, in full, is what produces next Tuesday's position. Accounts that skip that step end up generating ideas from nothing, which is exactly the state that made the calendar feel like a treadmill in the first place.

That habit is also the only reliable way to write in your audience's own words rather than your own. It is the same discipline behind turning customer feedback into hooks: the language that converts is language somebody else already used.

How Selzee turns your own comments into next month's ideas

An account never really runs out of ideas. It runs out of time to read the evidence it already has, which is sitting in comments, replies, reviews and the results of every post it has ever published.

That is the loop Selzee runs. It reads your comments, reviews and post results, finds the questions and objections that keep coming back, and turns them into concepts and briefs sorted into the five buckets above, so the next month of posts is built from what your audience actually said rather than from a prompt. Then it reads the results back and tells you which answers earned the sends.

It does not post for you and it does not invent an audience insight you never had. It reads the evidence, proposes what to make, and learns from what shipped. If you want the paid equivalent of the same loop, the short form video ads system covers it end to end.

Instagram post ideas: FAQ

What is a good engagement rate on Instagram?

Grade against the published median rather than a round number. Rival IQ's 2026 benchmark report puts the all-industry median at 0.30% by follower, down from 0.36%. Below that you are behind the field, at or above it you are normal, and at 0.60% or better you are running at double the median. Smaller accounts sit higher on this measure than large ones, so read your own trend first and the median second. You can check any single post with the Instagram engagement rate calculator.

How often should I post on Instagram?

Often enough to rotate all five buckets, which usually means about five posts a week. Beyond that, extra volume tends to come from repeating whichever bucket is easiest to make, and an account running one bucket on repeat loses more to boredom than it gains from frequency.

Do hashtags still help a post get found?

They are not where the leverage is. The predictions Instagram publishes are all about what a viewer does with the post: reshares it, finishes it, likes it, taps the audio. Nothing in that list is a hashtag. Use a few accurate ones, then spend the remaining energy on the first three seconds.

Should I repost my best performing post?

Not as a reel. Instagram lists reels that have already been posted on Instagram among the things it makes less visible. Remake the idea instead, in a different bucket or a different format, which usually improves it anyway because you now know which part of it worked.

What should I post if I am just starting and have almost no followers?

The same five buckets, weighted towards positions and teaching. With a small audience the follower rate is volatile and close to meaningless, so track the rate by reach instead: it tells you whether a post travelled, which is the only thing that matters before you have an audience to retain.

Where do I get ideas when the comments are quiet?

Read the comments on other people's posts in your category, and read your own customer service messages. A quiet account is not a quiet market. The questions people will not ask you publicly are the ones they ask in a support ticket, and those make the strongest posts because nobody else is answering them.

The short version

The list is not the hard part. Sorting ideas by the action they are built to earn, choosing the container after the idea, and reading the result against a published median rather than a feeling is what turns a content calendar into something that compounds.

The post layouts above come from the creative ad design dataset by Contralabs, used under CC BY 4.0.

Start with one bucket you have been avoiding, run it for two weeks, and check every post against the median. If you want the reading half of that loop done for you, from your own comments and reviews rather than from a prompt, see Selzee run on your own account.