Dark social attribution is the job of working out how much of your “direct” traffic really came from people sharing links privately. The visits are real, the sharing is real, but the analytics record carries no source. If you treat Direct as people typing your URL, you will undervalue whatever content gets passed around and overvalue everything else.
This post covers where the term came from, why the visits land where they do, what the original numbers showed (and why they are dated), and four practical methods. None of them produces an exact count. Together they get you closer.
What is dark social?
Alexis C. Madrigal named it in a 2012 Atlantic article, Dark Social: We Have the Whole History of the Web Wrong. Describing visits that arrive without referrer data, he wrote: “This means that this vast trove of social traffic is essentially invisible to most analytics programs. I call it DARK SOCIAL.”
His explanation of when referrer data goes missing still holds as a starting point: “The main situations in which this happens are email programs, instant messages, some mobile applications”, plus, in his 2012 example, moving from a secure site to a non-secure one. In plain terms, dark social channels are anywhere a person copies a link and sends it to someone else outside a platform that passes a referrer.
Why dark social traffic lands in Direct
Madrigal described how it appears: “It shows up variously in programs as “direct” or “typed/bookmarked” traffic, which implies to many site owners that you actually have a bookmark or typed in www.theatlantic.com into your browser. But that’s not actually what’s happening a lot of the time.” Instead, “someone Gchatted someone a link, or it came in on a big email distribution list, or your dad sent it to you.”
Google Analytics works the same way today. Its Help Center page “Default channel group” defines Direct as “the channel by which users arrive at your site/app via a saved link or by entering your URL”, and the rule is that the source “exactly matches “(direct)”” and the medium is “(not set)” or “(none)”. A visit with no referrer and no campaign tags meets that rule, whatever actually prompted it. Our post on direct traffic covers the other things that end up in the same bucket.
This matters for anyone reporting on content. A guide that earns organic visits and then gets forwarded around a team shows the first part as Organic Search and the second as Direct, so its real reach is split across two channels and partly hidden. If you only judge content by its search numbers, as described in our post on organic traffic in GA4, you will miss the sharing side entirely. Search Console won’t help either, because it only reports clicks from Google results; see Search Console vs Google Analytics for where each tool stops.
What the 2012 numbers showed
The article’s measurement trick came from Chartbeat. Visitors without referrer data were split into those landing on a homepage or section page and everyone landing on an article. Nobody types a long article URL, the reasoning went, so the second group must have been “following some sort of link”. Chartbeat counted them as “direct social”. The figures below are from October 2012 and describe that moment only.
| Measure (2012) | Figure | Scope |
|---|---|---|
| Dark social share of social traffic to stories | 56.5% | The Atlantic, one day |
| Known social networks’ share | 43.5% | The Atlantic, one day |
| Dark share of social referrals | Almost 69% | Chartbeat media sites |
| Facebook share of social referrals | 20% | Chartbeat media sites |
| Twitter share of social referrals | 6% | Chartbeat media sites |
| Direct/dark social share of all referrals | 17.5% (search: 21.5%) | Chartbeat media sites |
Read the 69% carefully: it is a share of social referrals, not of all traffic. Across all referrals, direct/dark social was 17.5%, second only to search. And the numbers belong to a different web. The article’s footnote quotes Chartbeat’s Josh Schwartz: “Only about four percent of total traffic is on mobile at all.” With browsing now very different, the 2012 shares can’t be applied to your site today. Nobody has a reliable current figure for dark social traffic, and any tool claiming one should be asked how.
Dark social attribution: four practical methods
You can’t make the dark visible, but you can shrink it and estimate the rest.
1. Tag every link you control
Newsletters, share buttons and links you post yourself can carry UTM parameters. GA’s Email channel rule matches when “Source = email|e-mail|e_mail|e mail OR Medium = email|e-mail|e_mail|e mail”, so a tagged newsletter link moves out of Direct. Our guide to UTM parameters covers naming conventions. This handles your own sends, not your readers’ forwards, though a forwarded tagged link keeps its tags.
2. Watch Direct traffic to deep pages
This is the Chartbeat logic, still usable. Direct visits to your homepage may be typed or bookmarked; Direct visits to a long article URL are much more likely to be shared links. Segment Direct by landing page and watch the deep URLs. Check for bots first, as covered in bot traffic in Google Analytics, because automated hits can inflate the same report.
3. Ask people
A “how did you hear about us?” field on sign-up or contact forms catches what analytics can’t. Answers such as “a friend sent it” or “a colleague’s email” are dark social by definition. Self-reported data is imperfect, but it is the only source that sees inside private messages.
4. Match spikes to known sends
When Direct traffic to a page jumps, compare the timing with things you know happened: a newsletter, a mention in a community, a post shared by a partner. A spike that lines up with an untagged send is a fair attribution; one that lines up with nothing is worth investigating.
What dark social marketing can and can’t do
Madrigal’s conclusion still stands: “The only real way to optimize for social spread is in the nature of the content itself. There’s no way to game email or people’s instant messages.” Make pages worth forwarding, add easy copy-link and share options that carry tags, and measure using the methods above.
The limitation is fundamental. None of these methods gives an exact count, and they overlap: a tagged link forwarded privately shows as Email, an untagged one as Direct. Use them to see direction, not to report a precise dark social share. For the wider picture, compare with your website traffic sources and referral traffic.
One more check before acting on any of this: rule out your own setup. An untagged newsletter, pages missing the analytics tag or a recent site change can all distort what Direct shows. Fix those first, then read what’s left as likely shared links.
Frequently asked questions
What is dark social?
Sharing that arrives without referrer data, such as links sent by email or instant message. Alexis C. Madrigal named it in a 2012 Atlantic article, writing “I call it DARK SOCIAL.”
Why does dark social show up as Direct traffic?
Google Analytics puts visits in Direct when the source is “(direct)” and the medium is “(not set)” or “(none)”. A shared link with no referrer and no tags matches that rule.
How much traffic is dark social?
There is no reliable current figure. The well-known 69% is from 2012 and covers social referrals on Chartbeat’s media sites, when only about four percent of traffic was on mobile.
Can dark social be measured exactly?
No. Tagging, deep-page Direct analysis, surveys and timing comparisons narrow it down, but none produces an exact count.
The takeaway Direct traffic to deep pages is mostly people sharing links. Tag what you send, ask how people found you, and report dark social as an estimate.
