A reel with 5,000 likes and zero DM shares will lose to a reel with 1,000 likes and 50 DM shares. That's not a hypothetical — it's roughly how both Instagram and TikTok's ranking systems behave right now. If your content strategy still treats the like button as the finish line, you're optimizing for the signal these platforms care about least.
This shift didn't happen quietly. Instagram head Adam Mosseri confirmed in January 2025 that watch time, sends per reach (DM shares), and likes per reach are the three signals that matter most across every surface of the app — and by his own account, likes carry the least weight of the three. TikTok's public messaging has moved the same direction: shares and saves now outrank comments, which outrank likes, in how the For You Page decides what to test further.
We're not writing this to add another "algorithm changed again" post to the pile. We're writing it because most brands and creators haven't actually changed what they make in response — they've just added "share this!" to the end of captions and called it a strategy. That's not the same thing as making content built to be sent.
What Actually Changed, and Why It Makes Sense
The logic behind this shift is simpler than most breakdowns make it sound. A like costs the viewer nothing — a half-second tap, no real decision involved. A DM share requires someone to stop, think of a specific person, and choose to send them something. That's a much higher bar, and platforms read it as a much stronger signal that the content is genuinely worth distributing.
Mosseri has described it this way in public conversations: short-form video is deeply tied to how people connect with friends, and Instagram wants to reward content that starts those private exchanges, not just public applause. TikTok's framing is nearly identical — a share means someone valued the content enough to interrupt someone else's feed with it, while a like signals only mild, passive approval.
The practical result: content built for a broadcast audience ("look how good this is") now underperforms content built for one specific recipient ("send this to the friend who needs to see it"). That's a real shift in what "good content" means, not just a ranking tweak.
It's Not Just Instagram
This same pattern shows up on TikTok, where completion rate and watch time still carry the heaviest single weight, but shares and saves sit clearly above comments and likes in the signal hierarchy. TikTok has also confirmed that follower count isn't a direct ranking factor at all — every video gets tested against a small audience first, and how that audience shares, saves, and rewatches determines whether it goes further. A brand-new account with a genuinely shareable video can outreach an account with a much larger following posting content that only earns polite likes.
The mechanic differs slightly by platform, but the direction is the same everywhere: platforms are rewarding private, effortful engagement over public, passive engagement.
The Data Behind the Shift
The numbers back up what the anecdotal reporting suggests. Buffer's 2026 State of Social Media Engagement report, which analyzed more than 52 million posts across 10 platforms, found Instagram's median engagement rate fell from roughly 7.3% in 2024 to 5.4% in 2025 — a 26% year-over-year drop, the steepest of any platform in the dataset.
That drop isn't necessarily a sign creators are doing worse work. Buffer's own data lead has cautioned against reading a falling rate as decline on its own — it can just as easily reflect a platform steering its measurement of success toward different signals. And that's exactly what's happening here: Instagram has increasingly pushed creators toward views as the headline metric, while the traditional "likes and comments over followers" formula measures less and less of what the platform is actually optimizing for underneath.
One format-level finding from the same report is worth sitting with. Reels get roughly 36% more reach than carousels on Instagram, but carousels earn about 12% more engagement per impression. That's not a contradiction — it's two different jobs. Reels are the awareness engine, built to travel to people who don't follow you yet. To see how we build this discovery layer for video channels, see our blueprint on building a high-performance YouTube content system in 2026. Carousels reward the slower, more deliberate interaction — pausing, swiping, saving, sending — that platforms now weight more heavily per person reached.
| Signal | Instagram Weight (2026) | TikTok Weight (2026) | What It Actually Signals |
|---|---|---|---|
| Watch time / completion | Highest across all surfaces | Heaviest single factor (~40-50%) | Content held attention past the hook |
| DM shares / sends | Top 3, strongest for reach | Weighted well above likes | Viewer chose to spend social capital on a specific person |
| Saves | Elevated above comments | Strong intent signal | Content has lasting or reference value |
| Comments | Depth matters more than count | Mid-tier, sentiment-weighted | Real engagement, though easy to fake with short replies |
| Likes | Lowest of the primary signals | Weakest signal in the hierarchy | Passive approval, minimal effort or intent |
So What Do You Actually Make Differently?
This is where most advice stalls out at "make shareable content" without saying what that means in practice. Based on what's actually driving shares and saves across both platforms, a few patterns hold up:
- Content built for one person, not a crowd. Before you finish a script or a caption, the useful test is whether it's written for a specific person to send to a specific friend — "this is so you," not "this is impressive." Relatable, funny, or surprising content drives shares. Useful, reference-worthy content drives saves. Trying to do both at once usually does neither well.
- Reference value that survives a second look. Step-by-step breakdowns, comparison content, before-and-afters, and anything framed as "save this for later" earns saves at a much higher rate than pure entertainment. If someone can point to a specific reason to come back to your post next week, that's a save-worthy structure.
- A hook that works with the sound off. None of this matters if the first two to three seconds don't hold attention, since watch time still gates everything else. About half of video is watched muted by default, so the opening needs to communicate value visually before it can earn the share or save. For more on how to manage these audience drops, read our guide on watch time vs completion rate in 2026 video performance.
- Genuine conversation, not just replies for volume. Both platforms now weight comment depth and DM interaction with your existing audience more heavily than raw counts. A few real threaded conversations do more for your ranking than fifty short emoji comments.
This is the exact pattern we build into every Social Media Management engagement. To see how this topic focus affects your account reach, read our guide on the TikTok cross-niche penalty and reach retention.
Building This Into an Actual Workflow
Here's roughly how we approach this shift when we're building content calendars for clients:
- Audit for save-and-send potential before production, not after. Every concept gets one question before it's greenlit: would the target audience actually send this to a specific friend, or save it for later? If the honest answer is "maybe someone likes it," it gets reworked before it gets filmed.
- Separate reach content from relationship content deliberately. Reels and short-form video get built as the discovery layer — hook-first, sound-off-friendly, structured for a stranger's first three seconds. Carousels and longer-form posts get built as the depth layer, for people already following, where saves and comment depth matter more than raw reach.
- Track sends and saves in reporting, not just likes and follower count. Both Instagram and TikTok expose this data in native analytics. If a client's monthly report still leads with likes, it's measuring the wrong thing.
- Iterate on what actually gets shared, not what gets the most surface-level praise. A post with modest likes but disproportionate saves or sends is a signal to make more like it.
If your team's current reporting still centers on likes and follower count, that's usually the first thing worth fixing — not because those numbers are meaningless, but because they're no longer the ones the platforms are actually rewarding. This is exactly the kind of workflow our Social Media Management service handles end-to-end, from the content calendar through the analytics that actually tell you what's working.
Summary
Likes aren't gone, and they're not irrelevant — they're just no longer the metric that gets you in front of people who don't already follow you. Instagram and TikTok have both said this plainly, and the engagement data backs it up: platforms are measuring intent now, not applause. A DM share means someone thought about a specific person and chose to interrupt their day for your content.
Track the signals that actually matter instead of the ones that are easiest to screenshot for a report. Ready to find out whether your current content is actually built for this? Get in touch and we'll walk through what's working and what isn't.
FAQs
Are likes still important on Instagram in 2026?
Likes still contribute to relationship-strength signals and public social proof, but they're the weakest of the primary ranking signals. They help retain an audience you already have more than they help you reach a new one.
What counts as a "DM share" on Instagram?
A DM share (also called a "send") is when someone uses the send/share icon to deliver your post or reel directly to another Instagram account through direct message, rather than reposting it publicly.
Does TikTok weight shares the same way as Instagram?
Similarly, though not identically. TikTok's heaviest single ranking factor is still completion rate and watch time, with shares and saves ranked above comments and likes. Instagram treats DM shares as one of its top three signals across all surfaces.
How do I see how many DM shares or sends my posts are getting?
Both Instagram and TikTok surface this in native analytics — check "Sends" under Instagram Insights for individual posts and reels, and "Shares" within TikTok's video analytics.
Why did my Instagram engagement rate drop even though my content hasn't changed?
Industry-wide data shows Instagram's median engagement rate fell significantly year-over-year as the platform shifted its emphasis toward views and private signals like sends and saves. A falling engagement rate calculated the old way doesn't necessarily mean your content is underperforming.
Should I stop posting Reels and focus on carousels instead?
No — they serve different purposes. Reels typically reach more people, making them better for discovery. Carousels tend to earn higher engagement per person reached, making them stronger for deepening relationships with an audience you already have.
Can I just ask people to share my posts in the caption?
Asking doesn't hurt, but it won't compensate for content that isn't genuinely worth sending. Platforms are reading actual sharing behavior, not caption text — the content itself has to earn the send.
What's the difference between a save and a share for ranking purposes?
A save signals a viewer wants to return to the content later, which is why reference-style content performs well. A share signals a viewer thought of a specific person and wanted to send it to them, which tends to favor relatable, funny, or surprising content. Both outrank likes.