Sends outrank likes for reaching people who do not follow you. Instagram’s head named watch time, likes and sends as the top three ranking signals, and of those, a send is the one that costs the sender something. Forwarding a post to a friend spends social capital. Tapping a heart does not.
Saves sit slightly outside that list and still matter, particularly in search, where saves and shares are among the signals Instagram uses to rank results.
This article covers why the ordering exists, what each action actually signals, how to earn the ones that count, and which of these numbers to watch when you are deciding what to make next.
Key Takeaways
- The confirmed top three are watch time, likes and sends, as stated by Adam Mosseri in January 2025.
- Sends matter more than likes for non-follower reach, because the signal is costlier to give (socialpilot.co, 2026).
- Saves are a strong signal in search ranking, alongside shares and watch time.
- Measure per reach, not as raw counts. A post with 40 sends on 2,000 reach beats one with 90 sends on 50,000.
- Different actions need different asks. A post built to be saved is built differently from one built to be sent.
Why a send outranks a like
Think about what each action costs the person doing it.
A like costs nothing. It is a single tap, invisible to almost everyone, and people give them out of habit. As a signal of quality it is weak, which is why it is the weakest of the three confirmed signals.
A send costs something real. The sender has to decide a specific person wants this, then put their own judgement behind it. Nobody forwards mediocre content to a friend, because being the person who sends mediocre content has a cost.
A save costs a small amount of belief that the content will be useful later. It is a private action, so it carries no social risk, but it does carry an implicit claim of future value.
That ordering explains the ranking weight. Instagram wants to show content people act on at a cost, because that is harder to fake than a tap.
What each action tells you
| Action | What it signals | Strongest for | Earned by |
|---|---|---|---|
| Like | Mild approval, habitual | Almost nothing on its own | Being agreeable |
| Save | ”I will need this again” | Search ranking, evergreen reach | Reference value |
| Send | ”This person specifically needs this” | Reaching non-followers | Naming a situation precisely |
| Watch time | Attention actually spent | Overall distribution | Holding the middle |
| Comment | Investment, and a reply loop | Conversation depth | Asking something answerable |
The practical read: likes tell you a post was fine. Saves tell you it was useful. Sends tell you it was specific.
How to earn sends
A send happens when a viewer thinks of one named person while watching. That is the entire mechanism, and it is more controllable than it sounds.
Name a situation, not an audience. “If your Reels get 300 views and your saves are higher than your likes” produces sends. “Tips for creators” does not. The more precisely you describe a specific predicament, the more likely someone thinks of the exact friend in it.
Make it something people argue about or agree hard with. Both produce forwards, for different reasons.
Give it a reason to be useful to the recipient, not just interesting to the sender. A post someone forwards has to survive arriving with no context.
What does not work: asking for the send. “Share this with a friend” is the weakest line in the format, because it asks the viewer to do the work of deciding who. Describe the person instead, and the viewer does the matching themselves.
How to earn saves
Saves follow reference value, which is a different thing entirely.
Lists, steps, specifications, prices, measurements, scripts, checklists. Anything a person cannot act on immediately but will need later. The test is whether somebody could reasonably want this again in three weeks.
Carousels earn saves disproportionately because they hold more reference material per post, which is also why the format is worth keeping even as video dominates distribution. Instagram’s Edits app now builds them too, as the Edits carousel editor guide shows.
Measure per reach, not as counts
This is the part that changes decisions.
A raw send count rewards your biggest posts, which tells you nothing you did not know. Dividing by reach tells you which content earns the strongest signal from the people who actually saw it, which is the thing you can act on when choosing what to make next.
Run the same division on saves. Then look at the posts that score highest on both and ask what they have in common. That answer is usually more specific than any general advice about the format. Instagram ranking signals covers where these sit among everything else that is on the record, and why your Reels get no views covers what to check when the ratios are fine and the reach is not.
Where engagement stops telling you anything
Sends and saves measure how content travels. Neither tells you whether it produced anything.
A post with an exceptional send rate and no clicks reached a lot of people who enjoyed it and did nothing. That is a real outcome and sometimes it is the one you wanted. When it is not, the gap is on the other side of the tap, and no amount of engagement data closes it.
What closes it is knowing which post drove which click, and then which click drove which sale. Link in bio analytics covers the numbers worth tracking on that side.
A More.You page counts clicks per block with breakdowns by source, device and country, and attributes those clicks back to the post that drove them. Create your More.You page and find out whether the posts that travel are the posts that pay.
FAQ
Are shares more important than likes on Instagram?
For reaching people who do not follow you, yes. Sends are one of the three confirmed ranking signals alongside watch time and likes, and a send is the costliest of those actions for the person giving it, which makes it the strongest signal (socialpilot.co, 2026).
Do saves help Instagram reach?
Saves are among the engagement signals Instagram uses to rank search results, alongside shares and watch time. They also indicate reference value, which tends to produce reach over a longer period rather than in the first day.
Should I ask people to share my posts?
Describing the person who needs it works better than asking. “Send this to the friend who posts daily and gets nothing” outperforms “share with a friend”, because the viewer does not have to do the matching work themselves.
What is a good send rate?
There is no published benchmark. Measure sends against reach on your own posts and compare to your own median, since the figure moves heavily with audience size and format.
Do comments still matter for ranking?
Comments are not in the confirmed top three, which is watch time, likes and sends. They remain useful as a depth signal and as a conversation surface, particularly where a reply turns into a DM.
Sources
- Watch time, likes and sends as the top three ranking signals, as stated by Adam Mosseri, January 2025
- DM sends mattering more than likes for non-follower reach, and the Reels ranking order (socialpilot.co, 2026)
- Saves, shares and watch time among Instagram search ranking signals (keywordseverywhere.com, 2026)
- More.You product brief, docs/moreyou/01-product-brief.md, sections 2 and 3 (analytics, click breakdowns, post attribution)
