The Dark Social Threshold: Why Your Engagement Metrics Are Already Lying to You

<!-- AWAITING CCO REVIEW — do not publish until Pamela approves. -->

I need to tell you something that's going to make your quarterly report uncomfortable.

The engagement metrics you're presenting — the likes, the comments, the public interactions — are telling you an increasingly incomplete story. Not because the data is wrong. Because the behavior it's measuring is no longer the behavior that matters.

And the gap between what you can measure and what's actually happening is growing wider every quarter.

The Numbers That Made Me Stop

Let's start with what everyone knows: public engagement is falling off a cliff. Instagram engagement rates collapsed from 2.5-3% in 2020 to 0.5-0.9% by 2025, according to Socialinsider's benchmark data across 35 million posts. Comments dropped 16% year-over-year on Instagram, 24% on TikTok. Overall social media engagement fell roughly 24% YoY in 2025.

Every marketing publication is running the same headline: *social media engagement is down.*

But here's the number they're not putting in the headline: **TikTok shares grew 45% year-over-year.** Instagram shares grew 12%. Direct messages are climbing across every platform. Instagram DMs now achieve 90% open rates against email's 20%. DM reply rates hit 60% versus email's 1-5% click-through. DM-to-sale conversion rates range from 7% to 20% for targeted campaigns.

The engagement didn't disappear. It went underground.

The Dark Social Threshold: A Definition

I call this the **Dark Social Threshold** — the point at which your measurable public engagement metrics become less predictive of actual content impact than the invisible engagement happening in DMs, group chats, and private shares.

Crossing this threshold doesn't mean your content stopped working. It means your dashboard stopped being able to tell you whether it's working.

Think about what that means for someone managing social media. Your boss asks for the monthly numbers. Likes are down. Comments are down. Reach is flat or declining. By every visible metric, things look worse. But the actual impact — the shares, the saves, the DM conversations your content started — is invisible to standard analytics. You can't prove it. You can only feel it.

That's the threshold. And most social media managers are standing on the wrong side of it, optimizing for metrics that stopped reflecting reality two years ago.

Why This Is Happening: The Privacy Cascade

This isn't one thing. It's four things happening simultaneously, and each one accelerates the others.

**First: the social cost of public engagement went up.** When Instagram made likes permanently visible, attached to your name, broadcast to your entire network, it changed the calculus. Liking a brand's post became a statement. For a lot of people, the safer move was to stop liking things at all. They save instead. They DM instead. They forward to three close friends instead of reacting publicly.

**Second: feed saturation hit a breaking point.** One social media manager on Threads put it plainly: "I scrolled for the first time in a while and could not believe the frequency of ads. It was like 4:1." When your feed feels like a television ad break, passive scrolling replaces active participation. People consume. They don't react — at least not where you can see.

**Third: the intimacy preference kicked in hard.** According to 1440.io's research on business messaging, 75% of consumers now want to communicate with businesses the same way they message friends — through instant, conversational, private channels. Public comment threads feel performative. DMs feel real. And 72% of consumers say they're more likely to buy from brands that offer messaging. 73% will buy from a competitor if a brand doesn't respond.

**Fourth: the platforms are structurally deprioritizing public interaction.** LinkedIn is surfacing more discussion-based content. X deprioritizes engagement bait. Instagram's recommendation engine is optimizing for saves and shares — signals that most analytics dashboards still treat as secondary metrics, if they track them at all.

The sum of these forces? A wholesale migration of meaningful engagement from the public surface to the private channel.

What This Actually Looks Like in Practice

Let me paint a picture you'll recognize.

Someone sees your post. They don't like it. They send it to a friend in a DM. That friend sends it to another friend. They discuss it in a group chat for a week. One of them buys your product.

From your dashboard's perspective, that post generated: zero likes, zero comments, zero measurable impact. The sale looks like it came from nowhere. Your content looks like it produced nothing.

This is happening constantly. And it creates a perverse incentive: the metrics say your content isn't working, so you change the content. But the content *was* working — it was working in a place you couldn't see. Now you've optimized it into something that generates likes because those are visible — and you've killed the thing that was actually driving sales.

That's not a measurement problem. That's a strategy problem wearing a measurement costume.

The Silent Purchase Path

The commercially significant version of the dark social threshold is what I call the **silent purchase path.** It works like this:

Content → private share → private discussion → purchase decision → zero attribution.

If your attribution model only captures the first and last steps — the content post and the purchase — everything in the middle becomes invisible. The content gets no credit. The budget that funded it looks wasted. Next quarter, that budget gets cut, the content changes, and the purchase that was actually being driven by your social content stops happening.

You can't see the cause. You can only see the effect. And by the time the effect is clear, you've already broken the machine.

Meta's own business data confirms the scale: 150 million Instagram users message a business every single month. Those are active purchase-path conversations. Twenty percent of Instagram Stories posted by businesses generate at least one direct message. Stories aren't an awareness format — they're a DM initiation tool, and your analytics dashboard is barely tracking that function.

How to Measure What You Can't See

If public engagement metrics are no longer the right yardstick, what do you measure instead? Here's what I'm tracking:

**1. Saves and shares are the new likes.** Saves mean someone found your content valuable enough to return to. Shares — even when they're private and you can only see the aggregate — mean someone found it worth passing on. Both are stronger intent signals than a public like ever was. If your content isn't generating saves, it's not generating the kind of engagement that precedes purchase decisions, regardless of what your like count says.

**2. DM velocity is a leading indicator.** How many DMs are you getting? From how many unique accounts? What percentage convert to conversations? What percentage of conversations lead to a purchase or a sign-up? These are sales metrics hiding in a messaging inbox, and most brands aren't tracking them.

**3. Post-purchase attribution needs a dark social question.** Add one question to your checkout or post-purchase survey: "Did you see us on social media before buying?" Not "which ad?" Not "which campaign?" Just: did social content play any role in your decision. The answers will surprise you — and they'll give you a proxy for the invisible engagement that your dashboard can't capture.

**4. Response time is an engagement metric.** Responding to a DM within one minute increases conversion by 391%, according to LeadResponse's 2026 DM statistics. If you're measuring engagement but not measuring response time, you're measuring the wrong thing. Speed of reply is, functionally, a conversion rate.

The Threshold You Can't Un-Cross

Here's what I think about late at night: once you cross the dark social threshold, you can't go back. You can't un-know that your metrics are incomplete. You can't un-see the gap between what your dashboard reports and what's actually happening. Every quarterly report will feel like you're presenting half the story — because you are.

The social media managers who thrive over the next two years won't be the ones with the best engagement rates. They'll be the ones who can tell a story about impact that extends beyond the dashboard — who can connect public content to private conversation to purchase behavior, even when the attribution chain is broken in the middle.

The dark social threshold isn't a crisis. It's a filter. The managers who cross it and adapt their measurement framework will gain an advantage their competitors won't even be able to see. The ones who keep optimizing for public likes will watch their dashboards decline and conclude that social media doesn't work anymore.

They'll be wrong. But they won't have the data to prove it.

*The SMF Works Project runs social strategy for brands navigating the platform shifts most agencies haven't caught up to yet. If your engagement dashboard is telling you things are getting worse while your gut says otherwise — let's talk.*