Why Your Instagram Reach Stalls Even When Likes Look Fine: Sends, Originality, and the Conversation After
A familiar frustration shows up in Insights. A post gets a decent number of likes, the comments look friendly, and yet the reach barely moves past your existing audience. No new followers of note. No new inquiries. It feels like the algorithm ignored a good post.
In many cases, it did not ignore it. It showed it to the audience it was built for, and that audience was your followers. Likes are good at one job. They are weaker at a different job, which is reaching people who have never seen you before. If you only read the like count, you will keep optimising for the wrong job.
This article is about that split, and about the step most algorithm guides skip: what happens after someone sends your post to a friend.
Connected reach and unconnected reach are two different jobs
Instagram does not have one audience for your post. It has, in plain terms, people who already follow you, and people who do not yet follow you. Content shown to followers is connected content. Content shown through recommendations, Explore and the Reels feed to people who do not follow you is unconnected content.
That distinction explains a lot of confusing results. A post can do well with followers and still fail to travel, because the signals that move it to strangers are not identical to the signals that keep it in front of followers.
Adam Mosseri, the head of Instagram, named three signals that matter most for ranking in January 2025, and those three are still cited in 2026 explainers: watch time, likes per reach, and sends per reach. His refinement is the part most summaries rush past. Likes matter slightly more for connected content. Sends matter slightly more for unconnected content.
Read that wording carefully, because it is the whole diagnosis. Not ten times more. Not a fixed multiplier. Slightly more, for a different audience. If a guide gives you a precise multiplier for sends over likes, treat it as that guide's invention, not as Instagram's rule. No multiplier has been published in the sources checked for this article.
The three signals, in plain language
Watch time is how long people keep watching. For a Reel, that includes average watch time and whether people watch to the end or replay it. A strong opening matters here, but not as a trick. If the first two seconds promise something the rest of the post does not deliver, people leave, and the signal falls. A hook earns the watch, and the watch only counts if the rest of the post holds it. Hook Engineering: How to Make People Stop Without Using Clickbait covers that craft in detail, so this article will not repeat it. The point here is simpler: a hook earns the watch, and the watch only counts if the rest of the post holds it.
Likes per reach is likes divided by accounts reached. It measures how many of the people who saw the post liked it, not how many likes you collected in total. A post with 500 likes from 50,000 people is not stronger on this signal than a post with 100 likes from 2,000 people. The rate matters, and it matters most for your existing followers.
Sends per reach is shares by direct message divided by accounts reached. When someone sends your post to a specific friend, Instagram reads that as a stronger sign that new people might also want it. A like can be a quick tap. A send is a choice to put your post in front of one named person. That is why sends carry weight for unconnected reach, and why a post can have many likes and still not travel. Nobody sent it anywhere.
Instagram Insights does not show sends per reach as a ready made number. You can work it out from two figures it does show: Shares divided by Accounts reached, as explained in the What sends per reach is and how to calculate it from Insights. Treat that result as a close stand in for Instagram's internal signal, not as an exact copy of it. The value is in comparing your own posts with each other, not in chasing a universal target that Instagram has not published.
A myth check, before you change your strategy
A lot of Instagram advice in 2026 sounds precise and is not. Here are four claims worth checking before you act on them.
Claim: Sends carry three to five times the weight of likes. What is on the record: Mosseri's wording was slightly more important for unconnected content. Some roundups repeat a larger multiplier, but the explainers checked for this article do not trace it to a published Instagram statement. What to do instead: Compare your own sends per reach across posts, and design for sends, without pretending you know the exact weight.
Claim: Original content gets a fixed percentage more distribution, for example 40 to 60 percent more. What is on the record: Instagram has said it tries to give the original creator more credit than someone who reposts the same content, and that reposts are less likely to be recommended. No fixed percentage was found in the announcements checked for this article. What to do instead: Publish your own version first, with your own footage, voice or edit, and do not rely on a repost to carry your reach.
Claim: One dated algorithm update in 2026 changed everything, so you need a new tactic this month. What is on the record: Mosseri's September 2026 video series explains ranking principles, including originality, recency, and showing small creators to people who do not follow them, as summarised in the Mosseri on originality, recency and small creators, September 2026 series. That is an explanation of how ranking works, not a single dated update with a new formula. What to do instead: Build around the stable signals, watch time, sends and originality, and test small changes, rather than rebuilding your strategy around every headline.
Claim: A high like count proves a post is working. What is on the record: Likes per reach matters most for connected reach. A post can satisfy followers and still give Instagram little reason to show it to strangers. What to do instead: Read likes, sends and watch time together, by format, and ask which audience each post actually reached.
None of this means likes are worthless. It means a like count alone cannot tell you whether a post travelled. If you want new reach, you need a signal that travels, and that signal is a send.
Diagnose your last 10 posts
You do not need a new tool for this. Open Insights, take your last 10 Reels or posts in one format, and answer three questions for each one.
First, what was the sends per reach? Divide Shares by Accounts reached. Sort the 10 posts by that rate, not by total likes. The top posts on this list are your sendable posts. Look at what they have in common. Is it a useful checklist, a relatable problem, a clear before and after, or a short explanation a follower would want a friend to see?
Second, was it original? Was the footage, image or idea yours, posted by you first, without another app's watermark, border, or a speed trick used to disguise a repost? Instagram has said it tries to replace identical reposts in recommendations with the original version, and that accounts built mainly on reposting other people's content are less likely to be recommended. If most of your last 10 posts are reposts with a logo added, that is a distribution problem, not a posting frequency problem.
Third, did the post give the receiver a next step? A friend who receives your Reel in a DM is not yet a lead. They are a curious stranger with one question: what do I do now? If the post ends with no clear next step, no reason to save it, reply, visit your profile, or start a conversation, the send ends there. That is the most common leak in this whole system, and it is entirely in your control.
Do this diagnosis once, honestly, before you post anything new. Most accounts find the problem in one of those three places, not in all three at once. Fix the one that is actually broken.
Make a post worth sending to one specific friend
People do not send generic content. They send content to a specific person, for a specific reason. A useful test before you publish is to finish this sentence: "I would send this to [one named type of friend] because [one clear reason]."
If you cannot finish that sentence, the post is probably too broad. Narrow it until you can.
Four kinds of posts tend to earn sends, and each one works for a different reason.
A useful shortcut earns a send because it saves the friend time. A short checklist, a script, a template idea, or a step by step explanation that fits in one Reel is easy to forward with a short note like "this is useful for you".
A relatable problem earns a send because it names something the friend has already felt. "This is exactly what you said last week" is a powerful reason to forward a post. The post does not need to be funny. It needs to be specific enough that the right friend recognises themselves in it.
A clear opinion earns a send because it starts a conversation. A calm, specific point of view on a common practice, with a reason behind it, gives the sender something to discuss. A vague motivational line does not, because there is nothing to agree or disagree with.
A proof post earns a send because it is hard to dismiss. A short demonstration, a real result explained with its limits, or a before and after that shows the work, not just the outcome, gives the friend a reason to trust the sender's taste. Keep any result honest and specific, and do not present one example as a promise for everyone.
In each case, the post still needs a clean opening and a payoff that matches it. The send test does not replace a good hook. It tells you whether the hook is attached to something worth forwarding.
The originality checklist, before you post
Originality is not only about avoiding a penalty. It is about giving Instagram a clear reason to credit you, and giving a new viewer a clear reason to remember you rather than the account you copied.
Before you post, check these six points.
- Is the footage, photo or core idea yours, or have you added a real edit, voiceover, or point of view that changes it?
- Is there any other app's watermark on it? If so, export a clean version before you post.
- Is the video low resolution, muted, bordered, or mostly text on screen? Instagram has said those formats are less likely to be recommended, so fix the format rather than blaming the idea.
- Does the post look and sound like the rest of your feed, or does it look like a brand ad dropped into a creator's grid? A consistent voice helps both viewers and ranking systems recognise the original.
- Did you post your version first, rather than reposting someone else's version and adding your logo?
- If you are using a creator's video on a brand account, have you agreed the usage separately? Reach and usage rights are different questions, and one does not settle the other.
None of these checks require a new app. They require a pause before publishing, and that pause is cheaper than a month of reposts that never travel.
After the send: the conversation is the conversion
This is the step most algorithm guides leave out, and it is where service businesses either win or waste the reach they just earned.
When someone sends your post, the receiver often replies to the sender, not to you. That is normal. Your job is to make the next step easy for both of them.
First, reply fast when the conversation does reach you. A DM reply, a comment reply, or a Story reply within the same day keeps the thread warm. A slow reply does not only lose the lead. It teaches the sender that forwarding your content leads nowhere.
Second, give the receiver something worth saving. A post that ends with a useful summary, a checklist, or a clear question is more likely to be saved and revisited than a post that ends with "follow for more". Saves are a sign that the content has lasting value, and they give the receiver a reason to come back when they are ready.
Third, make your profile answer the three questions a new visitor asks in seconds: who do you help, what do you do, and how do they start? If a stranger arrives from a forwarded Reel and finds a vague bio, no clear offer, and no booking route, the send has done its job and your profile has undone it. Getting Views But No Clients? How to Build a Content-to-Client System for Service Businesses covers that bridge in full, and it is the right next read if your profile is the weak link.
Fourth, turn the conversation toward an inquiry without a hard pitch. A simple, honest next step works better than a sales script in a first DM. Ask what they are working on, offer one useful observation, and invite a short call only if the conversation points that way. 30.8 Million Views Taught Me This: Virality Does Not Automatically Build Trust makes the same point from the other side: reach can make people notice you without giving them a reason to trust you. A send is reach with a recommendation attached. Trust is still earned in the reply.
What to test this week
Do not change everything at once. Change one thing, and read the result properly.
Pick your best performing topic from the diagnosis above, and publish two versions of it in the same format. The first version is your normal post. The second version is the same idea, narrowed to one specific friend, with one clear payoff and one clear next step for the receiver, such as a question to reply to, a checklist to save, or a profile link to check.
After a few days, compare the two posts on sends per reach, average watch time, and profile visits, not on total likes alone. If the narrowed version earns more sends per reach, you have found a sendable angle for that topic. Repeat it with a new topic next week. If it does not, check whether the problem was originality, watch time, or the next step, rather than assuming the topic is dead.
That test is small, but it is honest. It uses your own numbers, your own audience, and a signal that actually travels, instead of a borrowed benchmark that was never publ
ished.
The point to carry forward
Likes tell you whether your followers liked a post. Sends tell you whether anyone thought it was worth passing on. Watch time tells you whether it held attention. Originality tells Instagram who to credit.
A service business does not need all four to be perfect. It needs the right combination for the audience it actually wants: original posts, built for one specific friend to send, with a next step that turns a private conversation into a real inquiry.
Start with your last 10 posts. Find the one with the highest sends per reach, and ask what made it sendable. Then make the next post a little more like that one, and a little less like a post that only asks for a l
ike.
If you want that system built around your own services, explore the Social Media Strategy service, the Organic Growth Systems service, and the Conversion Psychology service, or use the booking route on the site to start a conversation about which topics your buyers actually forward. Discovery is changing on other platforms too, as Buyers Are Asking AI First explains for AI answers, but on Instagram, the send is still the simplest test of whether your content is worth passing on.
