Dark social
Dark social is sharing that analytics tools can't see, like a link pasted into Slack, a WhatsApp group, a DM or a private community. When a RevOps lead drops your guide into a Slack group, the visits it sends land in GA4 as direct traffic.
What is dark social?
Dark social is any sharing your analytics can't see. A link sent in a Slack channel, a WhatsApp group, a LinkedIn DM, a text message or an email to a colleague usually arrives without referrer data. So GA4 can't tell where the visit came from and files it as direct traffic.
Alexis Madrigal gave it the name in a 2012 article for The Atlantic. He noticed that a big share of the site's traffic came from links people passed along privately, and none of it showed up in his reports as social.
Here's what it looks like in B2B SaaS. Let's say you sell sales engagement software. Your head of growth writes a teardown of cold email reply rates. A RevOps lead pastes it into a private Slack community with 3,000 members, and 40 of them read it. Two weeks later, one of those readers asks for a demo. Your CRM says "direct". The buyer would tell you "someone shared your post in a Slack group".
Why dark social matters for B2B SaaS
B2B buying happens in groups, and groups talk in private channels. A champion forwards your pricing page to their CFO in Slack. A founder asks five peers in a WhatsApp group which tool they use. None of that shows up as a source in your reports.
That creates a real risk for your budget. If you only fund what your attribution tool can see, you'll keep moving money toward paid search and retargeting, which get credit for the last click. The posts and podcasts that started the conversation get cut, because their visits look like nothing came from them.
Here's why I think every SaaS team should plan for it:
- Word of mouth is still the strongest pitch: A recommendation from a peer in a private group carries more weight than any ad. Dark social is where that recommendation happens now.
- Your reports undercount your best content: The guides and opinion posts people share are the same ones your dashboards show as low traffic. Without a way to see dark social, you might stop writing them.
- It points you to the right communities: Once buyers tell you which groups and newsletters they heard about you in, you know where to show up next.
How to work with dark social
You can't track private chats, and you shouldn't want to. You can make your content easy to share and then ask buyers what happened. Here's my process:
- Add a free text field to your demo form: Ask "How did you hear about us?" and leave it open. A dropdown with "Social media" teaches you nothing, while answers like "my manager sent me your benchmark in Slack" tell you exactly what worked. This is self-reported attribution.
- Make content worth forwarding: Original data, a clear template or a sharp opinion gets pasted into Slack. A generic overview doesn't. Write the piece a buyer would send to their team with "this is what I meant".
- Make each piece readable in the chat: Put the main point in the title and first lines, so the preview card that shows up in Slack or WhatsApp already makes sense. Add a short URL that's easy to paste.
- Show up where buyers talk: Join the communities your ICP uses, like RevOps or product groups on Slack, and add useful answers there. Don't drop links. Members notice and remember who helps.
- Ask sales to log the story: Have reps note on the first call what the buyer read or heard before booking. Those notes fill the gaps the form leaves.
A dark social example
Here's a hypothetical example. A data warehouse startup publishes a free cost calculator for cloud data spend. In GA4, direct traffic to the calculator page jumps from 300 to 1,100 visits in one month, with no campaign running. Organic search and LinkedIn referrals stay flat.
On the demo form, the free text answers explain it:
| What buyers wrote | What it points to |
|---|---|
| "A friend shared your calculator in a data engineering Slack" | A private community share |
| "My VP sent it to our team" | An internal forward |
| "Saw it in a newsletter, then a colleague sent it again" | A newsletter plus a forward |
| "Someone posted it in our company channel" | An internal Slack share |
The numbers are made up for the example. The pattern is common: a spike in direct traffic to a deep page, like a tool or a report, usually means someone shared it. People rarely type a long URL like that from memory.
We wrote more on reading these signals in our post on AI search attribution, since AI answers create the same blind spot.
How to measure dark social
You won't get one clean number. You'll get a few signals that point the same way. Here's what I track:
- Self-reported answers: Tag the free text answers each month by source, like Slack community, colleague, newsletter or podcast. Count how often each one shows up next to deals in your CRM.
- Direct traffic to deep pages: In GA4, filter direct sessions by landing page. Direct visits to your homepage are normal. Direct visits to a blog post or calculator with a long URL are usually shared links.
- Branded search: Watch searches for your company name in Google Search Console. When private sharing grows, more people look you up by name. See branded search for how to read it.
- Sales call notes: Count how often reps hear "someone sent me" or "I saw it in a group". The sample is small, and it's often the most detailed one you have.
Common dark social mistakes
- Treating direct traffic as people who typed your URL: A lot of it is shared links. Check the landing pages before you decide what direct traffic means.
- Using a dropdown for "How did you hear about us?": A fixed list forces buyers to pick the closest match, which is usually "Google search". Leave the field open and tag the answers yourself.
- Cutting content because attribution can't see it: If a post never gets last click credit but buyers name it on calls, it's working. Judge it by what buyers say.
- Spamming communities: Dropping your links into every Slack group gets you banned and remembered for the wrong reasons. Answer questions and let others share your work.
Here's a first step you can take today. Add one open text field to your demo form and read the first 30 answers. How many of them name a place your analytics never showed you?
Read more
FAQ
Who coined the term dark social?
Alexis Madrigal, then a senior editor at The Atlantic, named it in a 2012 article. He noticed that a large share of the site's visits came from links people shared in email and chat, which showed up with no referrer.
Is dark social the same as direct traffic?
Not exactly. Dark social visits often land in GA4 as direct traffic because the app didn't pass a referrer, but direct traffic also includes people typing your URL or using a bookmark.
Can you track dark social at all?
You can't track the conversations themselves, and you shouldn't try. You can see their effect through a free text "How did you hear about us?" field, branded search trends and what buyers say on sales calls.
Does dark social include AI chat tools?
Some teams count it in the same bucket, since a buyer who reads about you in ChatGPT and then types your URL also shows up as direct. AI answers are their own channel, and the fix is the same one: ask buyers where they heard about you.


