If you spend any time on the social-media-advice internet, you've seen the lists. "10 secrets to skyrocket your engagement!" "The best time to post on Instagram in 2026!" "Why your engagement rate is declining!"
Most of it is noise. Some of it is wrong. Almost none of it is specific to your account.
This post is an attempt at the opposite: a clear-eyed look at the small number of things that actually move the needle on Instagram engagement, why those things matter, and why most of the conventional advice misses them. I'll be honest where the answer is "it depends," and direct where the data is unambiguous.
A note on framing: when I say "engagement" in this post, I mean meaningful interaction — comments, saves, shares, sustained watch time on Reels. Likes are weakly correlated with the things you actually care about (reach, follower growth, conversion) and the algorithm has progressively de-emphasized them since 2022. If your analytics tool is still leading with likes, that's a smell.
The five things that actually drive engagement
Across thousands of posts and dozens of accounts I've looked at, five factors do almost all of the explanatory work. They are:
- Post format (Feed vs. Carousel vs. Reel)
- Publish time (day of week × hour of day)
- Caption features (length, structure, presence of a clear hook and CTA)
- Hashtag cohort (the specific set of hashtags, not the count)
- Asset choice (which images or videos you actually picked)
Everything else — emoji density, exclamation point count, whether you tag your location, whether you @-mention people in the caption — is in the noise. Those things can have small effects in specific niches, but they're never the difference between a campaign working and not working.
Let's take each of the five in order.
1. Format
The single biggest decision you make on each post is the format. And it's the decision most creators make least deliberately.
Carousels generally outperform single-image feed posts on engagement, sometimes by a factor of 2-4x. The mechanism is mechanical, not mysterious: a carousel gives the algorithm multiple chances to re-serve the post to the same viewer if they swiped through, and it gives the viewer a reason to spend longer on the post. Both signals feed back into reach.
Reels, separately, drive reach much more than they drive deep engagement. A Reel will often outperform a feed post on impressions by 5-10x while underperforming on comments per impression. If your goal is awareness, Reels are the answer. If your goal is community depth, Reels probably aren't.
This is where most generic advice falls down: it tells you "post more Reels!" without knowing whether your goal is reach or depth. The correct answer depends on what you're trying to do, and nobody can answer that for you in the abstract.
The right way to think about format isn't "which format is best." It's "which format is best for this campaign's goal, given this account's history." For a new product launch where the goal is reach, lean Reels. For a community-building campaign where the goal is conversation, lean carousels. For pure visual storytelling with established context, feed posts still have their place.
2. Publish time
There is a "best time to post on Instagram." There is not a single best time that applies to your account.
Generic best-time advice is calculated by averaging engagement across millions of accounts and finding the global peak. That global peak — typically late morning or early evening on weekdays — is the average optimal time for an average follower base. Yours is not the average follower base.
The way to find your time is to look at when your audience is most active and engaged, not when the global audience is. Instagram's native insights show this for accounts with enough followers. It will not match the generic advice. A B2B account targeting decision-makers in finance will see entirely different peaks than a lifestyle creator targeting young adults.
There's a second wrinkle worth knowing: the relationship between time-of-posting and engagement is not always monotonic. There are accounts where posting at 6am — when generic advice says you shouldn't — outperforms 9am because the early audience is small, devoted, and high-engaging, and the algorithm uses that early signal to decide whether to expand reach later. The only way to know whether you're one of those accounts is to look at your own data.
3. Caption features
There's a lot of advice about captions. Most of it is wrong about the right thing.
What matters about a caption, in order of impact:
- The first line. Instagram truncates after roughly 125 characters; everything after that is hidden behind a "more" link that fewer than half of viewers tap. If your first line doesn't earn the tap, the rest of the caption doesn't exist for most of your audience. This is the single most important micro-decision in a caption.
- Whether there's an actual CTA. "Tell me in the comments" is a CTA. "What do you think?" is a weaker CTA. A pure descriptive caption with no CTA at all gets fewer comments — predictably, mechanically, because you didn't ask.
- Length, but only in the right direction. Longer captions tend to outperform shorter ones on saves and shares (long captions get re-read and re-circulated), but only if they're actually saying something. A long caption that meanders has no edge over a short one.
- Tone match to the asset. A serious caption on a playful image creates dissonance the algorithm notices indirectly through reduced dwell time. The asset and the caption have to feel like they came from the same person on the same day.
What doesn't matter as much as people claim: emoji count, exclamation marks, sentence length variety, paragraph breaks. These have small effects at the margin but are nowhere near as consequential as the four above.
4. Hashtag cohort
Hashtag advice has been wrong, on average, for about five years.
The current state of the evidence: hashtag count (how many you use) matters less than hashtag fit (whether the ones you use are actually relevant to your content). A post with 5 tightly-fitting hashtags will reliably outperform a post with 30 hashtags scraped from a "popular hashtags in your niche" list, because Instagram has gotten very good at detecting hashtag spam and discounting the reach signal accordingly.
What matters is the cohort — the specific combination of hashtags you use. The same individual hashtag can appear in a high-performing post and a low-performing post depending on what it's combined with. The algorithm reads the cohort as a whole, not the tags individually.
The practical implication: if you have a set of hashtags that has worked for your account, keep using that cohort. Don't randomly shuffle in new tags every post chasing what's "trending." The algorithm rewards consistency in topical signal, not novelty.
5. Asset choice
The least-discussed driver, and arguably the most important.
Every account has a few images or videos that just work — for reasons that are partly aesthetic, partly compositional, partly about subject matter, partly about who's in the frame. You can usually tell which assets these are after the fact (they're the ones that overperformed), but it's much harder to tell in advance.
The shortcut: tag your assets when you upload them. Format (product shot, lifestyle, behind-the-scenes), subject (face, hands, product, environment), tone (warm, cool, candid, posed). Once you've published a few dozen posts, the patterns become visible. "Carousels that lead with a product shot underperform; carousels that lead with a face overperform." That kind of finding is invisible from raw metrics — it requires structured tags on the assets.
This is the thing none of the social media tools do well, and it's the thing that compounds the longest. A creator who has structured tags on three years of content can answer questions about their own performance that no analytics dashboard can.
What about all the other advice?
There's a long tail of conventional wisdom about Instagram engagement: post consistency, follower-to-following ratio, story stickers, the algorithm "punishing" you for not posting enough, etc.
Most of it is partially true in specific cases and wildly overstated in general. Posting consistency does help — but only because it gives the algorithm more data points and gives you more chances to learn. The "best practice" of posting daily is wrong for plenty of accounts; many B2B and creator accounts perform better at 3-4 posts a week with higher-quality assets than at 7 mediocre posts.
The deeper problem with all of this generic advice: it's been calculated against a global average that isn't yours. The right amount of consistency, the right format mix, the right caption length, the right hashtag count — these all depend on your specific account, your specific audience, and your specific campaign goal.
What to do instead
The honest answer: stop reading generic engagement advice and start reading your own data.
Specifically:
- Declare a goal up front for every campaign. Not "grow my account" — a specific target metric over a specific window.
- Tag your assets when you upload them, even if rough. Format, subject, tone. You'll thank yourself in six months.
- After each campaign ends, look at what drove the result — not just what the result was. Which format outperformed? Which posting times? Which caption styles? Which asset tags?
- Compare to your own historical baseline, not to industry averages. "Engagement rate up 12% vs. my last quarter" is meaningful. "Engagement rate 2.3% vs. industry 1.8%" is misleading.
If you do those four things — even crudely, even in a spreadsheet — you'll outperform 95% of accounts in your niche over a year, because almost nobody actually does them.
This is part of what we're building at Fohn. Every campaign starts with a goal. Every campaign ends with structured analysis of what drove the result — decomposed across format, time, captions, hashtags, and asset tags, compared to your own baseline. If that sounds like something you'd want, the waitlist is here.