Which posts actually send DMs, and which just get likes

A reel with forty thousand views and nine DMs is telling you something specific, and it is not that the reel failed. Somewhere between the hook and the comment box, people decided not to ask. The numbers on a flow tell you where that happened.

In Instoly every flow reports four things: the comments it matched, the DMs it delivered, how many of those were read, and how many people clicked. We read them in that order, because each one narrows down what to fix.

What does each flow report?

• Comments matched. How many comments contained your keyword, which we would read as a measure of the ask rather than of the automation.
• DMs delivered. How many messages actually went out, which should track the matches closely.
• Read rate. How many of those DMs were opened.
• Click-through. How many people tapped the link or a button once they had read it.

The dashboard in Instoly adds the same picture across everything: replies sent, comments answered, click-through, and which flows are currently live. We look at that view first and drop into the per-flow numbers only when something there looks wrong.

What does a disappointing number actually mean?

Each of the four fails in its own way, and the fix is different every time.

• Few matches, plenty of comments. The ask was not clear, or the keyword was only in the caption. We would say the word out loud in the next one.
• Matches high, delivered lower. Something is stopping sends, and it is usually the messaging permission or an expired connection rather than the flow itself.
• Delivered high, read low. The DM landed in the Requests folder because they do not follow you, which is normal and worth expecting rather than fixing.
• Read high, clicked low. They opened it and were not convinced. This is the DM's fault, not the reel's, and it is the cheapest thing on this list to rewrite.

Why does one keyword per post matter here?

Because the numbers are only useful if they point at one thing. Use the same word on three posts and we can only show you one merged figure, which cannot tell you which reel did the work — the exact question you opened the page to ask.

We give each post its own word for that reason alone, and it is also what stops two live flows racing for the same comment. There is more on choosing those words in picking a keyword that does not misfire.

Which reporting comes with which plan?

Instagram analytics and Auto DM reports are on the free plan, so you are not flying blind at ₹0. The four per-flow numbers above work on every plan, free included: we did not think a creator should have to pay to find out whether their first flow is working.

There is also an AI insight report that reads your whole account and tells you what actually changed. That one is analysis rather than counting, and it spends an AI writing credit each time you ask for it, so we would run it monthly rather than after every post.

One more habit worth having: check the numbers a day after posting rather than an hour after. Comments arrive for as long as a reel keeps travelling, and we have watched people write off a post that went on to do most of its work overnight.

What would we do with a month of this?

1. Look at matches per post, not per follower. Two hundred matches on a small account beats a viral reel that nobody commented on.
2. Find your best two posts and copy their ask word for word into the next reel, including where in the video you said it.
3. Rewrite the DM on any flow where the read rate is healthy and the clicks are not, then leave it a fortnight and compare.
4. Retire keywords that misfire. A word catching compliments spends credits and teaches you nothing.
5. Keep the flows that work switched on. In our numbers an old reel that still pulls comments is the quietest, cheapest traffic anyone has.

Views tell you the algorithm liked it. Matches tell you a person wanted something.

Instoly counts all of this whether or not you look. The accounts that grow are the ones where somebody reads the four numbers once a week and changes one thing.