Campaign attribution tells you which channel or campaign brought a visit. Conversation attribution tells you which specific reply, post, DM, call, or community message created it. The unit of work in founder-led distribution is a conversation, not a campaign, and standard attribution stops one layer above it.You answer a question in a Reddit thread. You send a cold email to one prospect. You reply to someone on X. You DM a lead on LinkedIn. You follow up after a call with a link.
Then a visit shows up in your analytics, labeled “social,” “email,” or “referral.”
What you usually want to know is narrower:
Which specific reply, post, DM, or call brought this person here?
That is the question standard attribution is not built to answer. It is the question conversation attribution is built to answer.
What campaign attribution actually answers
Google Analytics defines attribution as “assigning credit for important user actions to different ads, clicks, and factors along the user’s path to completing the action.” An attribution model is the rule or algorithm that decides how credit is split across the touchpoints on that path.
GA4 currently ships three models: data-driven attribution, paid and organic last click, and Google paid channels last click. The older rule-based models (first-click, linear, time-decay, and position-based) were deprecated in November 2023.
Notice the vocabulary. The unit is a touchpoint: an ad click, an organic click, a channel engagement. The grain is campaign, channel, and creative. That vocabulary is excellent for deciding which ad budget earned a conversion. It is the wrong grain for deciding which human conversation earned a click.
Where campaign attribution stops
UTM parameters are the standard way to add campaign context to a destination URL. Google documents utm_source, utm_medium, utm_campaign, utm_term, and utm_content, plus newer fields like utm_id and utm_source_platform. The recommendation is to always set source, medium, and campaign.
The interesting field for our purposes is utm_content. GA documents it as a way to “differentiate creatives”, for example, two call-to-action links inside the same email, so you can tell which version performed better. That is the closest standard UTM field to “which specific message.”
And that is exactly where the limit shows up.
utm_content=reply_04 only means something if somebody remembers what “reply 04” was six weeks later. At conversation scale, the operator has to mint a unique value for every reply, DM, and email, keep an external ledger mapping each value to the human context, and keep the naming consistent (UTM values are case-sensitive; Spring and spring are different rows). The link becomes a tiny database record you are managing by hand.
If you need a conventional campaign URL, HeyZinc ships a UTM link builder. UTMs solve a real problem: they give analytics tools a standard vocabulary for campaign attribution. They describe a campaign, a medium, and a creative. They do not describe the conversation that produced the click.
We wrote a full mechanics deep-dive on this in tracked links vs. UTM parameters. The short version: a UTM answers “which campaign?” A tracked link answers “which conversation?”
The unit of work is a conversation, not a campaign
If you run paid ads, campaign attribution is the right grain. You have a budget, you have creatives, you want to know which combination converted.
If you are doing founder-led or community-driven distribution, the unit of work is different. It is:
- a reply in a Reddit thread,
- a DM on LinkedIn or X,
- a cold email to one prospect,
- an answer in a community Slack or Discord,
- a follow-up link sent after a sales call.
You are not running one campaign. You are having dozens of conversations a week. The wins are usually one specific message in one specific thread, not “social” in aggregate.
For that workflow, the attribution question is per-conversation, not per-campaign. And the message-creation side has its own tools: a cold-email generator for the send, a cold-DM generator for the DM. What is missing is the layer that connects each individual message to what happened after the click.
What conversation attribution means
Conversation attribution is attributing a visit (and downstream, a conversion) to a specific conversation: a specific reply, post, DM, email, or call follow-up, rather than to a channel or campaign.
One honest scoping note: “conversation attribution” is a HeyZinc name for this layer. GA4 does not ship a “conversation attribution” report, and the term is not a standardized analytics dimension. The legitimate claim is narrower: standard campaign attribution is campaign-, channel-, and creative-grained, and there is an operational layer beneath that, the individual conversation, that campaign attribution does not natively capture.
Conversation attribution answers two questions campaign attribution cannot:
- Which reply, post, DM, or call brought this visitor?
- Is that visitor on the site right now?
The first is a labeling problem. The second is a timing problem. Both matter.
How HeyZinc tracked links close the gap
The mechanic is simple, and we cover it in depth in the tracked links vs. UTM parameters post. The short version here:
Each HeyZinc tracked link carries a short first-party token in the URL. The per-conversation context (source page URL, page title, platform, and for Reddit the subreddit) is kept server-side with the link record, not jammed into the URL as a hand-typed label. The browser extension builds that context from the page you are already on, so a link made while replying to a specific thread carries that thread’s context without you typing it.
The resulting visit shows up in the dashboard attached to that conversation’s context: visits, unique visitors, sessions, engaged visitors, and an active-now state when a visitor is currently present.
The difference from a UTM is structural. With a UTM, you maintain the mapping between utm_content values and conversations yourself, in a spreadsheet. With a tracked link, the mapping is captured when the link is made and stays with the link record. You stop bookkeeping a URL and start reading a dashboard.
On privacy: a custom first-party token is less likely to be caught by rules aimed at known advertising and analytics parameters: browsers and engines do strip known tracking query parameters, and MDN’s web privacy documentation notes that several browsers remove known tracking parameters. But this is a reduction in exposure, not immunity. A determined privacy tool can still remove, rewrite, block, or refuse any tracking. The accurate claim is narrower: a custom first-party token is less exposed to the rules that target the best-known parameter names, while the richer context lives server-side.
Across channels: email, social, DMs, calls, community replies
The pattern is one tracked link per conversation, across every channel you actually use:
- Email and cold email: one link per send. You see which outreach message produced the click, not just “traffic from email.”
- Social replies (X, LinkedIn): one link per reply, carrying the source-page context. You see which specific reply earned the visit.
- Direct messages: one link per DM. LinkedIn DMs are a distinct enough workflow to deserve their own treatment; we cover tracking conversions from LinkedIn DMs separately.
- Calls: the call is itself one touchpoint. A post-call follow-up link attributes the visit that came after the call, so you can tell which call produced the next site visit.
- Community threads (Reddit, Slack, Discord): one link per reply, with subreddit context preserved for Reddit. “Traffic from Reddit” is a platform answer. “Traffic from the answer I posted in this specific discussion” is a sales-context answer.
Every channel gets the same treatment: the link remembers the conversation that created it.
Real-time attribution is an action window
Attribution is most useful when the visit is still live, not in tomorrow’s report.
The active-now state, when a visitor is present, plus companion and mobile notifications turn a reporting line into a response window. If someone clicks a link from a cold email and is browsing your site right now, you do not have to discover the visit in next week’s analytics. You can decide whether to follow up, call, start a live chat, or do nothing.
Two qualifications matter here. First, the active-now state and the notification surfaces depend on what the visitor has shared and how your workspace is configured; mobile alerts are a configured workflow, not a universal delivery guarantee. Second, the point is not to interrupt every visitor. The point is to make the moment visible when the context is still useful, so you can start a conversation with a website visitor at the right time rather than after the fact.
This is where conversation attribution meets proactive outreach. Attribution tells you which conversation brought the visitor. Real-time engagement lets you act on it while intent is fresh.
From visit to conversion: closing the loop
The end state of attribution is connecting a conversation to a conversion, not just to a visit.
Here the multi-touch reality matters. A prospect is often touched by several conversations across channels before converting: a community reply, then a DM, then a cold email, then a call. Conversation attribution does not replace multi-touch or data-driven attribution for aggregate budget decisions; GA4’s models are the right tool for “which channel mix earned the conversion.” Conversation attribution gives each touchpoint richer, per-conversation context, and it makes the earliest conversation visible, the one that last-click attribution typically erases.
One limit worth stating plainly: seeing a visit does not prove a conversion, and a click does not prove causal lift. Conversation-level conversion claims need attributable internal data or a controlled result. What tracked links give you is the connection: this conversation produced this visit, and this visit (with the right downstream tracking) became this customer. For the full path from a single message to a customer, see tracking which message generated a customer.
What conversation attribution cannot do
It is worth being explicit about the limits, because the category attracts overclaim.
- It is not identity resolution. A tracked link attributes a visit to a conversation. It does not name the visitor unless the visitor identifies themselves. If you want the identity side, that is a separate problem; see how to identify website visitors.
- It is not immune to privacy tooling. A custom first-party token reduces exposure to known-parameter rules. It does not guarantee survival against a determined privacy tool.
- It is not a replacement for campaign or multi-touch attribution. Use UTMs and GA4’s models for aggregate budget allocation. Use conversation attribution for the operational layer beneath: which message, which reply, which call.
- Live and mobile signals are configuration-dependent. Active-now and notifications depend on your workspace and device setup, not a universal delivery guarantee.
How to operationalize it
The workflow is simple enough to do during a normal day of outreach:
- One link per conversation. When you are about to share a URL in a reply, DM, email, or post-call follow-up, make a tracked link for that destination.
- Let it capture context. The link picks up the source page, title, platform, and subreddit from the page you are on.
- Paste the share URL into the conversation.
- Watch the dashboard. See visits, unique visitors, sessions, engaged visitors, and active-now state for that link.
- Decide the next action while intent is fresh. Follow up, call, start a live chat, or do nothing. The point is visibility at the right moment, not interruption.
- Optionally connect it to an auto-engage rule, with a human takeover path. A link from a social reply can inform a tailored opening message; it should not force an automatic message on every click.
- Over time, compare. Which conversations produced visits? Which produced follow-ups? That is the conversation-level signal campaign attribution cannot give you.
Final thoughts
Campaign attribution answers “did the campaign work?” Conversation attribution answers “which conversation started the next one?”
For teams that distribute by having conversations (across email, social, DMs, calls, and community threads), the second question is the one that actually changes what you do tomorrow. You stop guessing which reply earned the visit. You stop bookkeeping a spreadsheet of utm_content codes. And you get a chance to act while the visitor is still on your site, instead of reading about them in next week’s report.
That is the difference between knowing a channel worked and knowing which message earned the next customer.
If you want the full mechanics (the token, the dashboard, the privacy posture), the tracked links vs. UTM parameters post is the deep dive. If you want to try it on your own outreach, HeyZinc is here.
