Dark Social Attribution: See What Analytics Misses

Your analytics says a third of your traffic is "direct". Almost none of those people typed your URL. This guide is about where they actually came from, and how to find out per share.

What dark social is

Dark social is any sharing that happens through channels your analytics cannot observe. The classic examples: someone pastes your link into a WhatsApp chat, a Slack workspace, a Discord server, an email or a DM. It also covers a lot of public sharing that behaves the same way, like links in Reddit comments and X replies, where the referrer gets lost on the way to your site.

The person clicked a link a real human shared. Your analytics just has no idea where that link lived. The share itself is invisible, only the click arrives.

Why it shows up as "direct traffic"

Analytics tools attribute a visit using the HTTP referrer, the address of the page the visitor came from. When there is no referrer, the visit is filed under "direct," a bucket originally meant for people typing your URL by hand.

Referrers go missing constantly. Native mobile apps (WhatsApp, Telegram, the Reddit and X apps) usually pass none. Email clients pass none. HTTPS-to-HTTP transitions drop them. Link wrappers like t.co and various privacy settings mangle or strip what is left.

So "direct" is not people typing your URL. It is mostly dark social: shares in places your analytics cannot see. If you make decisions from the channel report while most word-of-mouth lands in "direct," you are optimizing the visible tail and ignoring the body.

How much of your traffic is dark

You can estimate your own share of dark social without any new tooling. Take your "direct" sessions and subtract the visits that plausibly are typed-in traffic: visits landing exactly on your homepage from returning visitors. Direct visits landing on deep URLs, like a specific guide or a pricing page, were almost certainly shared somewhere, because nobody types those by hand.

Industry studies have claimed that well over half of all outbound sharing happens through dark channels. Treat any global number as directional: the honest answer is that the share varies a lot by audience, so measure your own instead of quoting someone else's.

Deep-link direct traffic is your quick proxy: the higher it is, the more of your growth is happening in channels you cannot currently see.

Why UTMs do not fix it on social

UTM parameters attribute campaigns you planned, not sharing other people do. On social and chat channels they fail in three ways.

First, you can only add UTMs to links you place yourself. The moment a reader reshapes or re-pastes your link, the tag describes the wrong hop, and most dark social is other people re-sharing.

Second, platforms interfere. X rewraps every URL through t.co, Reddit users routinely strip tracking junk before posting, and some apps and privacy tools remove utm_ parameters outright. There is a whole guide on the UTM alternative that works on Reddit and X if this is your main pain.

Third, long parameter-stuffed links look like spam, which costs you clicks in exactly the communities, like Reddit, where trust decides whether anyone clicks at all.

How to attribute dark social (tracked links at share time)

The reliable approach flips the problem: instead of hoping the destination can reconstruct the source, record the source at the moment of sharing. Every share gets its own short redirect link that already knows where it is about to live.

Manually, that means creating a fresh short link per share and keeping a log: this link went into this Reddit thread, that one into a Slack community, that one into an X reply. Clicks on each link are attributable no matter what the platform does to referrers, because the link itself is the identifier.

The manual version breaks down at scale, since nobody keeps a spreadsheet for every reply. This is the part PasteTrack automates: when you paste a link into Reddit or X, it is wrapped automatically and the context, like subreddit or thread, is captured at paste time, with zero extra steps in the moment.

Reading which share converted

Clicks are only half of attribution. The question that matters is which share produced signups, not which produced traffic.

To close the loop, connect each tracked link to a conversion event on your site. Then the report you actually want becomes readable: this Reddit comment produced three signups, that X reply produced one, the launch post produced none. That is a channel decision you can act on, double down where conversions came from, stop investing where they did not.

Once dark social is attributed this way, "direct" stops being a black hole and starts ranking your invisible channels against the visible ones.