DM Automation Analytics: Which Numbers Matter and How to Find Your Top Engagers

Most dashboards answer a question nobody asked. Four numbers, read in order, tell you exactly which part of a campaign is broken.

By Anthony Casauria · · 8 min read

Share
Smiling woman in a cream quarter-zip sweater does a small fist pump at her laptop at a sunny cafe table.

Key takeaways

  • The DM automation analytics worth watching are trigger rate, delivery, click-through and conversions, read in that order so the first bad number shows what to fix.
  • Total DMs sent tracks your reach rather than your funnel, so treat the number as a measure of the post and not of the automation.
  • Log a conversion immediately after the result you are counting, not at the start of a flow, where it counts intent instead of outcome.
  • Top engagers are the people who keep turning up in your comments, not the ones who received the most DMs, so find them through comment-level reporting or contact tags.
In this article
  1. What DM automation analytics should actually tell you
  2. The numbers worth ignoring
  3. How to set up tracking that answers those questions
  4. How to find your top engagers on Instagram
  5. Reading your DM automation analytics without fooling yourself

Most DM automation analytics screens lead with the number that means least. DMs sent counts how many people commented, which measures your post, not your funnel. It climbs when a reel travels and drops when one doesn't, and it says nothing about whether the automation behind it is any good. The numbers worth your attention isolate a single point of failure, so that a bad week points at one specific thing instead of a vague feeling that engagement is down. There are four of them, they run in sequence, and each tests a different link in the chain. Below: the metric stack worth watching, the vanity numbers to skip, the setup that makes conversions countable at all, and how to pull out the people who keep showing up for you.

Short answer: watch four numbers in order: trigger rate (did people type the keyword), delivery (did the DM leave), click-through (did the message earn the tap), and conversions (did the offer close). Each isolates one failure. Top engagers come from comment-level reporting and contact tags, not from DM volume.

These numbers sit at the end of our guide to Instagram DM automation.

What DM automation analytics should actually tell you

A dashboard earns its place when every tile maps to a decision. Four tiles do, and they run in order, so the first one that looks wrong is the one to fix.

MetricThe question it answersWhat a bad number means
Trigger rateOf everyone who commented, how many typed your keyword?Keyword or caption problem
DMs deliveredDid the private reply leave, or was it held, queued or expired?Setup, permissions or timing problem
Click-throughOf delivered DMs, how many taps did the link get?Message or offer-mismatch problem
ConversionsOf the taps, how many became the result you wanted?Price, landing page or audience problem

The sequence does the work. Click-through only means something once delivery is healthy, and conversions only mean something once click-through is. HyperDM computes click-through as link taps per DM sent across the window you pick, and counts repeat taps in the raw link-clicks figure but not as separate people, which is why those two numbers drift apart on campaigns with a strong link. It also declines to crown a best automation until one has sent at least 5 DMs in that window. Copy the habit: a 100% click-through on three sends is three people.

The numbers worth ignoring

Some metrics move for reasons you don't control, and reacting to them costs you a week.

Total DMs sent is the main offender, since it tracks reach. Follower count is worse, because none of these mechanisms require a follow in the first place. Raw link clicks with no denominator attached will flatter any viral post. Per-hour send caps deserve the same scepticism: the widely repeated figure of 200 automated DMs per hour is not a limit Meta publishes for Instagram, and Meta's own rate limiting documentation describes app-level call budgets rather than a DM ceiling. Pacing like that is a choice your tool made, so read it as a setting. One subtler trap is worth knowing about: a link preview is not a tap. Instagram and other apps fetch your link once to draw the thumbnail, and a tool that counts that fetch as engagement will report clicks from nobody. HyperDM excludes preview fetches for exactly this reason. For the longer version of this argument, the funnel guide covers where DM funnels usually leak, and common DM automation mistakes covers the measurement errors that follow. For the reporting itself, HyperDM's link tracking and analytics feature shows what each tile is derived from.

How to set up tracking that answers those questions

Link click tracking in DMs and conversion counting both need deliberate setup. Four steps, in this order.

Step 1: Confirm link tracking is on before you launch

Tracking has to exist before the campaign runs, because taps that happened untracked cannot be recovered afterwards. In HyperDM, the links it adds for you are tracked by default: automation links, flow and broadcast link buttons, product buy links and email buttons. A workspace can switch to sending links exactly as typed, which disables tap counting for those links, so check the workspace setting rather than assuming. Worth knowing before you wonder why a campaign shows zero clicks: links to Amazon, Flipkart, Walmart, eBay, Etsy and AliExpress are always sent exactly as typed so they open in the retailer's app, which means their taps are never counted. If your whole offer is an affiliate link to one of those, plan to measure the sale at the other end.

Step 2: Name the results you want counted

A conversion event is the result itself, named once and reused: a booked call, a paid deposit. In HyperDM these live under Settings, each name up to 50 characters and unique in the workspace, with up to 500 events including archived ones. If the result carries money, tick the value option and choose a currency at creation time, because that currency cannot be changed later without invalidating what has already been counted in it. Be specific when naming. "Lead" collapses three different outcomes into one unreadable total, while "Booked discovery call" and "Downloaded pricing PDF" still mean something six months on.

Step 3: Log the conversion at the moment it happens

Automation conversion tracking fails most often because the event gets logged in the wrong place, usually at the start of a flow, where it counts intent instead of outcome. Put the logging step immediately after the thing you are counting: after the booking question is answered, not before it is asked. In HyperDM this is a Log conversion action step inside a flow, where you pick the event and, for events that carry money, either a fixed amount or the value held by a number field at the moment the step runs. Each run counts the step once, so a loop or a re-delivery will not double-count. Set the flow live afterwards; a logging step with no event selected will block publishing until it is configured.

Step 4: Compare like-for-like windows

Pick a window length and keep it. Comparing a 30-day stretch against a 7-day one produces nonsense, and so does comparing a window with a viral post in it against one without. Both rates here count what happened inside the window, which means a tap can belong to a DM sent before the window opened, slightly inflating click-through at the start of any period that follows a busy one. Watch for partial history as well: features that began recording on a specific date will flag a window starting earlier rather than quietly reporting a low number. Read those notices instead of treating your first month as a baseline.

How to find your top engagers on Instagram

Top engagers are not the people who received the most DMs. They are the people who keep turning up in your comments, which is a different and far more useful list.

Comment-level reporting is where this lives. HyperDM's audience reporting counts unique commenters, comments and comments per commenter for the range you pick, then names top commenters with how many of your automations their comments matched. It also defines a superfan precisely: somebody who commented at least twice in a week, in at least two separate weeks of your range, with weeks starting Monday in your workspace timezone. That definition earns its keep by filtering out the person who left nine comments on one viral post and never came back. Those cards sit on Pro and Agency plans, though comments are recorded on every plan including Free, so the history is already there when you move up. On any plan, the alternative is tagging: tag contacts as they come through a flow, then filter the contact list by tag and combine conditions to build a segment you can message later.

Reading your DM automation analytics without fooling yourself

Good DM automation analytics are boring on purpose. Four numbers, read in order, each pointing at one fixable thing, plus the discipline to ignore everything that moves with reach rather than with your work. Set up link tracking before you need it. Name conversion events precisely enough that they still mean something next quarter, and log them where the result actually happens. Hold your comparison windows constant. For top engagers, use comment frequency across weeks instead of DM volume, because repeated commenting is the signal and a single viral thread is noise. Before you pick a plan, compare what each one records against what it merely displays: HyperDM's plans start free with 50 conversations a month, which is enough to get the tracking right before anything depends on it.

Frequently asked questions

ManyChat reports on flows, broadcasts and contact growth, and its contact records support tags and segments you can filter. Any tool built on Meta's API can report sends, clicks and contact-level activity. The meaningful differences between tools are comment-level reporting, how conversions are defined, and whether rates are shown with their denominators.
Partly. Instagram and Facebook send a read receipt when the person opens the chat, so a "seen" share is possible for DMs sent after read-receipt counting began in the tool. Treat it as a sanity check on delivery, not as engagement. A message can be seen and ignored, and that is what click-through is for.
Usually a denominator problem or a preview problem. Check whether repeat taps from the same person are counted individually, and whether link previews are excluded. Also check the window: taps landing inside it from DMs sent before it started will lift the rate for a few days after a busy period.
No for the basics. Sends, clicks, click-through and contact counts are standard on free and entry tiers across the category. Revenue attribution, folder-level totals, comment reporting and tap locations are where paid tiers usually start. Check whether a tool records that data on lower plans and simply locks the view, because that determines if you get history retroactively.
Until the automation has sent enough DMs for a rate to be stable, which is more than a handful. Five sends is a floor for noticing anything, and a few dozen is where click-through stops swinging wildly. Below that, read the activity log for individual failures instead of reading rates.
HyperDM

Run the automations. Add AI when you’re ready.

Comment-to-DM, keyword replies and follow gates — set them up, test them, then add optional AI answers from your real catalog. Free on one DM channel, 50 conversations a month, no card.