Dark social: what it is and how to measure what tracking can't see

Updated 2026-11-02

Dark social is any sharing that happens where analytics can't see it. A link in a text message, a mention in a Slack community, a recommendation in a group chat. The conversation is real and it can drive sales, but it leaves no trace in your reports.

The journalist Alexis Madrigal coined the term in 2012 for web traffic that arrived with no referrer. It's widened since then. Marketers now use it for almost any word of mouth that happens out of view.

What counts as dark social

The common examples are:

• Direct messages on LinkedIn, X, Instagram and other apps
• Messaging apps such as WhatsApp, Signal and iMessage
• Private Slack and Discord communities
• Email forwards
• Links copied into documents or internal wikis
• Private groups and closed forums

Some related channels leave little or no trail either, even though they aren't strictly social:

• Podcasts, where listeners can't click
• Video mentions that send people to search your name
• AI assistant recommendations
• Conversations at events, or in person

Why it shows up as Direct

When someone clicks a normal link, the browser usually tells the destination site where the click came from. That's the referrer. Analytics uses it to name the source.

Many private apps don't send a referrer. Then the visit arrives with no source, and analytics labels it Direct. The same thing happens when someone hears a name and types the URL.

So your Direct channel is a mix of loyal visitors, typed URLs and hidden sharing. Part of what looks like your brand's own pull is word of mouth from somewhere you can't see.

UTM tags help only when you control the link. Once someone copies a clean URL and pastes it into a chat, your tags are gone.

Why it matters

Buyers trust what peers tell them, and they often check with people they know before they buy. If your reports can't see those conversations, you get a skewed picture.

Here's how it can go wrong. Say a made-up company's founder posts useful advice on LinkedIn. People screenshot it and share it in private groups. A few weeks later, those people search the brand name, click a branded ad, and buy. Last-touch reports credit branded search. The founder's posts look like they do nothing, and the company cuts them. Three months later, branded search falls, and nobody can explain why.

That's the risk. You can defund a channel because it was never measured.

How to measure dark social

You can't track it directly, by definition. But you can catch its shadow from several angles.

1. Ask people

This is the most direct method. Add a how did you hear about us question to your signup form. When someone writes "my friend in a Slack group", you've recorded what no script could.

Compare the answers with tracked first visits. If 40 people say "podcast" and tracking has most of them as Direct, you've measured part of the gap. This is self-reported attribution.

2. Watch your Direct and branded search trends

Rises in Direct traffic and in searches for your brand name often follow word of mouth. Plot them against your activity. If Direct climbs the week after a podcast airs, that's a clue. It's only a clue, so check it against other data.

3. Look at which pages get Direct visits

People don't often type long URLs. If a deep blog post gets lots of Direct visits, someone is likely sharing that link somewhere private. Check your landing page reports for this pattern.

4. Listen in public

Search public communities for your brand. Reddit, X, LinkedIn comments and review sites give you a sample of what's being said in private spaces. Keep in mind that public talk is only part of it.

5. Ask sales and support

Your reps hear this every day. "My colleague told me about you" is a data point. Give your team a simple way to log it.

6. Use share-friendly links

Add a share button or short link for content you want passed around. You won't catch everything, and it can recover some of what you'd otherwise lose.

7. Run a test

To check whether a hidden channel works, change one thing and watch the results. Pause a sponsorship for a month. Or run it only in one region. Watch signups and self-reported answers. It's slower and takes some nerve, but it's how you measure effect.

What not to do

Don't try to track private messages. You can't, and you shouldn't try.

Don't treat all Direct as dark social. Plenty of it really is direct. Returning customers, bookmarks and typed URLs all land there.

Don't pick a neat percentage. You may see claims about how much sharing is dark. Your number will depend on your market. Measure your own.

Don't expect precision. Dark social gives you evidence, and you'll have to use judgment. If several signals point the same way, act on them.

How to get more of it

Measuring is one part. You also want people to talk about you.

• Make things worth passing on: templates, data and strong opinions.
• Write content that holds up without a click, such as complete LinkedIn posts.
• Show up in the communities where your buyers ask each other for advice.
• Give happy customers a simple way to recommend you.

See what tracking misses with Headstream

Headstream puts tracked and self-reported data side by side. Answers from a heard_about field, the identify call, the API or an import are sorted into groups like Podcast, AI assistant, and Friend or colleague.

Each person's answer sits next to their first tracked visit. A "hidden from tracking" stat shows how often people name a source that tracking missed. If you connect Stripe, in beta, you can see the revenue behind each group.

Headstream can't see inside private channels, and no analytics tool can. It can show you the size of the gap, so you can make decisions with it in mind. For more on this, read our guide to marketing attribution, or see what we can't see.