Tracking links shared in Slack and Discord: a six-step setup that ignores preview bots

To track links shared in Slack and Discord, give each channel or server its own short link on a domain you control and count clicks at that redirect.

A click from a chat app often reaches your site with no referring page, which Google Analytics 4 reports as Direct, and both apps fetch the URL themselves within moments of the post, so your log shows a bot request before the first person clicks.

This works for links you control: ones you post, or hand to a customer or community manager to post.

Why Slack and Discord clicks arrive without a source

Alexis Madrigal named the problem in The Atlantic on October 12, 2012. He called it dark social: visits with no referrer data, which he traced to email programs, instant messages, some mobile apps and secure-to-non-secure links.

Analytics tools file them as direct. Google Analytics 4 does so by rule.

Its default channel group defines Direct as a source of exactly (direct) with a medium of (none) or (not set), and Google's direct traffic help adds that redirects and URL shorteners can strip the details that identify a source.

Web clients can drop the referrer too: rel="noreferrer" makes the browser omit the Referer header, MDN says. Whether each Slack and Discord client sends one today is worth measuring in step 2, and an app update can change the answer.

What the preview crawlers do to your click count

Both apps request a link when someone posts it, to draw a preview card, and those requests reach your server before any reader does.

Slack ignores robots.txt, saying it acts for a human and does not follow links like a crawler.

Because the cache is service-wide, one link posted in 3 workspaces within half an hour can produce a single Slackbot request that no log can assign to a workspace.

Version numbers drift: Slack's page shows Slack-ImgProxy 0.19, while the crawler-user-agents list on GitHub records 0.59, 0.66, 1.106, 1.138 and 149. Discord's help article, updated October 3, 2026, says to match the word Discordbot anywhere in the string.

How to track links shared in Slack and Discord

The examples below follow one page, example.com/pricing, posted in 1 Discord server, 2 Slack workspaces and a set of founder DMs, with a $39 monthly plan.

Step 1: Give every place its own short link

Create one link per place, each pointing at the same destination, and name it after the place: Discord #launches, Community Slack #tools, Acme customer Slack. Every click then belongs to that place whether or not the app sends a referrer.

In TrackRev's New campaign form, pick "A link per channel" and add a custom short link.

The channel picker offers 9 channels and neither Slack nor Discord is one, but bulk CSV import (up to 500 rows) and the REST API accept both.

A slack or discord row stamps utm_source and utm_medium=social on the destination, unless the destination already has its own tags.

url,name,channel,utm_campaign,utm_term,utm_content,tags
https://example.com/pricing,Discord #launches,discord,oct-launch,,launches,community
https://example.com/pricing,Community Slack #tools,slack,oct-launch,,tools,community
https://example.com/pricing,Acme customer Slack,slack,oct-launch,,acme,customer

Skip Smart Links for these two apps. A Smart Link guesses the channel from the referrer. Slack proxies or strips the referrer, so TrackRev has no Slack mapping and those clicks land in Other or Direct.

Discord is on the recognized list, but only a click carrying a discord.com referrer gets the label.

Put the links on your own domain with one CNAME record (custom tracking domains). If you tag destination URLs by hand, the free UTM builder writes the tags, and medium social lands in Organic Social under Google's channel rules.

A redirect that forgets to forward the query string drops them, and the 5-step stripped-UTM test finds the hop that does it.

Step 2: Post a test link in each app

Post a fresh link in a private channel or test server on each client: desktop, web and phone. That is 6 test posts across the 2 apps.

Use a new link per test because of the 1-hour rule above, and one link per message because Slack expands none in a message with more than 5, then click each one from each client and read the report.

A client that sends no referrer costs you nothing, because the link's name carries the source.

Step 3: Flag preview fetches and still redirect them

Match crawlers on a user-agent substring (Slackbot, Discordbot), flag the request as a bot, store it, and leave it out of the counts. I would not block them, because a refused fetch leaves the post bare.

Discord publishes its public IP ranges in 2 formats, JSON and TXT, for anything that needs more than a spoofable user agent. The bot filtering guide covers prefetch headers and empty user agents.

TrackRev does this at the redirect with the isbot 5 user-agent list, which matches all 4 crawlers in the first table, plus empty user agents and prefetch headers.

A flagged request is still redirected and stored with is_bot set, and the link page, attribution and partner reports leave it out.

Step 4: Carry the click to the signup and the payment

A click becomes revenue when the visitor ID reaches the payment.

TrackRev's first-party pixel reads the ID on your first pageview, an email address binds the visitor to the customer at signup or checkout, and Stripe is read with a restricted key every hour, or by webhook for updates within seconds.

Built by hand, the shape is the same:

The Payment Links walkthrough has the setup.

Two gaps stay open: a click in the phone app followed by a laptop visit to your address arrives as Direct, and a signup email that differs from the payment email can slip past the join.

Step 5: Rank links by revenue per 100 clicks

Raw clicks reward the biggest audience. Divide by counted clicks to compare places of different sizes.

Discord #launches brings 53.7% of the 2,218 clicks and 21.4% of the $546 MRR. The founder DMs bring 3.9% of the clicks and 28.6% of the MRR.

Acme converts 6 of 11 signups to paid (54.5%) against 3 of 14 for Discord (21.4%).

Four paying customers from 86 clicks is too small a sample to repeat on purpose, so test the lead with a second batch of DMs.

Step 6: Wait out the trial before you judge a link

A link's revenue arrives after the click by the length of your trial plus the payment's confirmation time, and some payment methods, such as SEPA Direct Debit, confirm later.

For a 14-day trial and a click on October 1 (a 30-day trial moves the charge to October 31):

What these links can't tell you

A redirect logs a device, a place and a time, and a name appears only when the visitor later signs up.

A Slack link copied into a DM keeps crediting its original place, and a teammate who types your domain without the link sends Direct traffic.

A signup field asking how people heard about you catches those, and the post on self-reported versus tracked attribution covers weighing the two.

What TrackRev does here, and when you don't need it

TrackRev is SaaS affiliate software with link tracking built in.

It logs each click on the server with time, country, city, device, browser, operating system and referrer, adds the link's saved UTM tags to the destination, and joins clicks to payments from Stripe, Paddle Billing, Polar, Lemon Squeezy, Creem and Dodo Payments.

Revenue per link needs a paid plan (pricing).

If you only need to know whether Slack sends sessions, a UTM link and Google Analytics 4 are free and enough, and any shortener with bot filtering covers plain click counts.

The Links and Analytics page lists what a link records.