How to Track Dark Social and Word-of-Mouth Without Guessing
Dark social, the sharing that happens in a DM, a group chat, or a private Slack channel instead of a public share button, doesn't show up in any analytics platform, which leaves most marketers assuming it doesn't matter much. It usually matters quite a lot more than the reporting suggests, especially in categories where buyers trust a peer recommendation more than any ad.
This guide covers practical ways to make dark social visible enough to act on, even though you'll never fully close the attribution gap.
Vendors Covered in this Article
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Can You Ever Fully Measure Dark Social Attribution?
Perfect attribution for dark social isn't realistic, and chasing it usually wastes more time than the insight is worth. The goal is a reasonable estimate you can act on, not a precise number you can defend to the decimal point. Set that expectation with leadership up front so a rough estimate doesn't get dismissed for lacking false precision.
Combine several imperfect signals rather than relying on one. Direct-traffic spikes that don't correlate with any known campaign, branded search increases after an event or press mention, and self-reported source data all point toward the same underlying activity even though none of them measures it directly.
Combine several imperfect signals to estimate dark social activity:
- Direct-traffic spikes that do not correlate with any campaign you know you ran, which may reflect private sharing.
- Increases in branded search after an event, a podcast mention, or press coverage.
- Self-reported source data collected through an optional, easy-to-skip question at signup and reviewed over time.
- Calls that would otherwise show up as untracked direct traffic, once they are attributed with call tracking.
Should You Ask Customers How They Heard About You?
A simple "how did you hear about us" field at signup, kept optional and easy to skip, still captures a meaningful share of dark social activity that no tracking pixel ever would. Keep the answer options broad enough to include "a friend or colleague recommended it" or "saw it shared in a group chat," since a narrow list of only official channels will push respondents to guess the closest official-sounding option instead.
Follow up with a short optional question asking where that recommendation happened, if they're willing to say. Even a low response rate on this second question, aggregated over enough signups, tells you which private channels are actually driving referrals.
Call Tracking Catches What Web Analytics Can't
A prospect who saw your brand mentioned in a private group often doesn't click a link at all, they call. Dynamic call tracking through a platform like WhatConverts assigns a tracked number to different campaigns and sources, so calls that would otherwise show up as untracked direct traffic get attributed to something closer to their real origin, even when that origin is a word-of-mouth referral rather than a specific ad.
Pair call tracking with a system like KrispCall for the calls themselves, so the same platform recording the source can also capture call outcomes, giving you a fuller picture of which untrackable-by-default leads actually convert into revenue, not just which ones dial in.
Reading Direct Traffic and Branded Search as Proxies
A spike in direct traffic or branded search that doesn't line up with any campaign you ran is a reasonable proxy for dark social or word-of-mouth activity, especially right after a podcast mention, a conference talk, or a piece of press coverage. It's not proof, but a consistent pattern over several such events is a real signal worth building into how you evaluate those activities' actual return.
If your overall email open rate sits meaningfully above the roughly 35 percent industry average during a period with no obvious campaign explanation1, word-of-mouth-driven signups arriving more engaged than your average subscriber is a plausible explanation worth checking against your source data.
Building a Referral Path That's Easier to Track
Give happy customers an easy, specific way to share, a referral link, a discount code tied to their name, rather than leaving word-of-mouth entirely informal. This doesn't eliminate dark social sharing you'll never see, but it converts some of what would have been untrackable into a channel you can actually measure and reward.
Review these proxy signals quarterly alongside your known-channel data, so dark social activity gets weighed in strategic decisions instead of being treated as if it doesn't exist just because it's hard to measure precisely.
For example, suppose a company gives customers a referral link and a discount code tied to their name, and after a quarter compares those signups with its direct-traffic and branded-search trends. Signups arriving through referral links are a measurable slice of word of mouth, while the rest of the movement in direct traffic points to sharing you cannot see. A common mistake is treating that unseen portion as zero. The fix is to review both sets of numbers together each quarter and give the proxy signals a stated weight in planning, even if the weight is only a rough estimate.
What Good Looks Like
A working approach to dark social treats it as a real, if imprecise, channel: it combines self-reported source data, call tracking, and direct-traffic or branded-search proxies into a reasonable estimate that gets reviewed regularly, rather than ignoring the activity because it can't be measured exactly.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Assigns tracked numbers to different sources so calls that would otherwise show as untracked direct traffic get attributed closer to their real origin.
Captures call outcomes alongside the tracked source, giving a fuller picture of which hard-to-attribute leads actually convert.
Frequently Asked Questions
Can we ever get exact attribution for dark social referrals?
Not with full precision, and trying to force exact numbers usually wastes more effort than it's worth. A reasonable estimate built from several proxy signals, direct traffic spikes, self-reported source data, call tracking, is the realistic goal rather than a precise figure.
Is call tracking worth setting up if most of our leads come through forms?
It depends on your category, but even a modest share of calls that would otherwise show as untracked direct traffic is worth attributing correctly if those calls tend to be higher-intent. Test it for a quarter before deciding it's not worth the ongoing cost.
Should we ask customers directly how they heard about us?
Yes, and keep it optional and easy to answer with broad options, including something like a friend's recommendation or a private group. Even a modest response rate on this question captures signal that no analytics tool can see on its own.
Sources
Where we quote a benchmark, we show its source. Other figures in this guide are estimates or general guidance, so check them against your own numbers.
- Email Open Rate (All Industries). Mailchimp Email Marketing Benchmarks (data updated Dec 2023), 2023.
Related Guides
Building a Case Study Pipeline That Doesn't Depend on One Person
A worksheet-style walkthrough for turning happy customers into finished case studies on a repeatable cadence, instead of whenever someone remembers to ask.
First-Touch vs. Multi-Touch Attribution: Picking the Less-Wrong Model
How first-touch and multi-touch attribution each distort your numbers in a different way, and how to pick which distortion you can actually live with.
Organic Content, Technical SEO, or Backlinks: Where to Spend First
A comparison of three ways to grow organic search visibility for a small company, with the tradeoffs and the order that usually pays off fastest.
What a Healthy Landing Page Conversion Rate Looks Like
How to read landing page conversion benchmarks correctly, with a worked example showing when a rate under the median is fine and when it isn't.
Building a Category Narrative Your Market Will Repeat
How to build a category narrative that gives buyers new language for a problem they already have, instead of a rebrand nobody outside the company notices.
A Content Syndication Checklist for Reaching Readers You Don't Already Have
A checklist for syndicating your best content to other publishers and newsletters, plus the mistakes that make syndication a waste of time.